Document 5b0JrXgBMzEQ00qpywJXYqYMD

REFERENCES (FDA Docket No. 94P-0420/CP 1) SECTION II. D thru SECTION IV. SECTION I I . D REGULATORY TOXICOLOGY AND PHARMACOLOCY 2 1 , 2 3 3 -2 4 1 (1 9 9 5 ) The NTP Talc Inhalation Study: A Critical Appraisal Focused on Lung Particle Overload1 Gunter O berdorster Department of Environmental Medicine, The University of Rochester, Rochester, New York 14642 Received October 1, 1994 Recently published results in a N TP report of a 2year inhalation study w ith talc in rats and mice seem to fit the category o f being associated w ith particle o ver load quite well: Exposure concentrations of 6 and 18 m g/m 3 induced pulm onary inflam m ation and fibrosis in m ale and fem ale rats and induction o f lung tum ors (in fem ale rats only) of the high exposure group; m ice of either sex show ed an inflam m atory response but did not show pulm onary fibrosis or lung tumors. A nalysis of the particle accum ulation kinetics in lungs of both rats and m ice indeed show s that lung overload had been reached at both exposure concentrations in both species result ing in increased talc accum ulation o f high lung burdens. This and the chronic inflam m atory response indicate that the m axim um tolerated dose (MTD) had been e x ceeded at both exposure levels. This result w as predict able based on the outcome o f a 4-w eek range-finding study prior to initiation o f the chronic talc study; how ever, the short duration of the range-finding study may have been inadequate to g ive great confidence in the prediction and therefore may have accounted for the failure to include a concentration below the MTD in the chronic study. Further analysis of the results of the chronic talc study show that talc particles behave like other low -toxicity particles such as T i0 2and toner with respect to effects on lung clearance and chronic pulmo nary inflam m ation. The conclusion of the NTP report that there is clear evidence of pulm onary carcinogenic ity o f talc in fem ale rats should, therefore, be qualified by a statem ent that this is a secondary effect due to the high pulm onary p article load and its associated chronic toxicity. Accordingly, the relevancy o f the tumors ob served in the fem ale high-exposure group for occupa tio n a l hum an ex p o su r e s m ay be q u estio n ab le, e issa A cadem ic P ress, Inc. there was clear evidence of carcinogenic activity of talc in female F344/N rats based on increased incidences of alveolar/bronchiolar adenomas and carcinomas of the lung, and that there was no evidence of carcinogenic ac tivity of talc in male or female B6C3F! mice.2 In this study, male and female F344/N rats were ex posed to aerosols of 0, 6, or 18 mg nonfibrous talc/m3, free of S i0 2 and asbestiform minerals, for 113 and 122 weeks, respectively; male and female B6C3F, mice were exposed to the same talc concentrations for up to 104 weeks. This exposure resulted in concentration-related chronic inflammation, cell proliferation, and fibrosis in the lungs of both male and female rats, and 26% of the female rats of the high-exposure group (13 out of 50) de veloped lung tumors. The mice showed only limited chronic inflammation and no increased cell prolifera tive, fibrotic, or tumorigenic responses in their lungs. According to the report (NTP, 1993), the exposure con centrations for the chronic study had been selected based on results of a 4-week talc inhalation study which led to the following conclusion: Exposure concentrations greater than 18 mg/m3were expected to overwhelm lung clearance mechanisms and impair lung function. It is precisely this aspect of overwhelming lung clear ance mechanisms that is of critical importance for the interpretation of results from chronic inhalation studies with highly insoluble particles of low cytotoxicity. These particles were formerly categorized as "nuisance" dusts, but are now categorized as particles not otherwise clas sified (PNOC; ACGIH, 1994). A number of chronic in halation studies in rats involving such types of particles have resulted in pulmonary fibrosis and even lung tu mors at high-exposure concentrations which affect lung Uses and Health Perspectives, NIH, Bethesda, MD, .January 31-Feb- INTRODUCTION ruary 1, 1994. 1There was also some evidence of increased incidence of benign or Conclusions in a recently published NTP report malignant pheochromocytomas of the adrenal glands in rats. Yet, this (1993) of a chronic inhalation study with talc were that aspect will not be considered further in this paper since these tumors are not likely to be a direct effect of talc particles on the adrenals inas much as talc particles are highly unlikely to reach these glands and 1 Presented, in part, at the International Society of Regulatory Toxbackground incidences of pheochromocytomas in control rats were al icology and Pharmacology/U.S. FDA Workshop on Talc: Consumer ready high. 233 027:1-2300/95 $5.00 Copyright : 1995 by Academic Press. Inc. All rights o f reproduction in any form reserved. r e g u l a t o r y t o x ic o l o g y a n d p h a r m a c o l o g y 2 1 , 244-249 (1995) An Analysis of the National Toxicology Program's (NTP) Technical Report (NTP TR 421) on the Toxicology and Carcinogenesis Studies of Talc1 J ay I. G o o d m a n 2 Department of Pharmacology and Toxicology, Institute for Environmental Toxicology, B-440 Life Sciences Building, Michigan State University, East Lansing, Michigan 48824 Received October 1, 1994 The N TP toxicology and carcinogenicity studies of nonasbestiform , cosm etic-grade talc (the N TP Talc Re port) w ere conducted by exposing m ale and fem ale F 344/N rats and B6C3F1 mice to target aerosol concen trations of 0, 6, and 18 m g/m3 talc for 6 hr daily, 5 days per w eek. Based on results o f the high dose, the Report concluded that talc caused lung tum ors in fem ale rats and pheochrom ocytom as in m ale and fem ale rats, and there w as no evidence o f carcinogenic activity in mice. A thorough evaluation of lung toxicity revealed that talc-induced lung tum ors occurred only in the group of anim als that exhibited the most profound degree of chronic toxicity. H ow ever, these data w ere presented as em pirical observations rather than discussed in a manner that would relate them to the risk assessm ent im plications of the bioassay, i.e., relevant data were collected but not " used." In addition, the evaluation of the pheochrom ocytom as w as inadequate because it failed to place sufficient em phasis on the spontaneous incidence o f this tum or in rats. These deficiencies caused the author to vote against the conclusions pre sented in the Talc Report w hen it w as review ed by the N T P Board of Scientific Counselors. The appropriate conclusions are (1) the data do not indicate that the pheochrom ocytom as w ere treatm ent-related; (2) the m axim um tolerated dose (MTD) w as exceeded in the fe m ale rats exposed to the high dose; and (3) talc is not expected to cause lung tum ors under conditions of exposure that fail to result in m arked chronic lung tox icity . 199 5 A cadem ic Preee, Inc. INTRODUCTION The NTP toxicology and carcinogenicity studies of nonasbestiform, cosmetic-grade talc (the NTP Talc Re- port) were conducted by exposing male and female F344/N rats and B6C3F1 mice to target aerosol concen trations of 0, 6, and 18 mg/m3 talc for 6 hr daily, 5 days per week (NTP, 1993). A unique aspect of these studies was the inclusion of a very thorough evaluation of lung toxicity, in addition to the standard gross pathology and histopathology assessments. The NTP Talc Report was reviewed by the NTP Board of Scientific Counselors (The Board), Technical Reports Review Subcommittee on June 23-24,1992. Jay I. Goodman participated in this review as a member of the Board and a member of the Board's Technical Reports Review Subcommittee. The Board concurred with the recommendations of the NTP staff and voted to approve (by a vote of seven yes votes to one no vote (J. I. Goodman)) the following conclusions: "Under conditions of these inhalation studies there was some evidence of carcinogenic activity of talc in male F344/N rats based on an increased incidence of benign and malignant pheochromocytomas of the adrenal gland. There was clear evidence of carcinogenic activity of talc in female F344/N rats based on increased inci dences of alveolar/bronchiolar adenomas and carcino mas of the lung and benign and malignant pheochromo cytomas of the adrenal gland. There was no evidence of carcinogenic activity of talc in male or female B6C3F1 mice exposed to 6 or 18 mg/m3." The specific aims of this paper are to focus on the studies involving rats, the species in which there was deemed to be a carcinogenic effect, and (1) to provide an overview of the talc bioassay (NTP, 1993) with an emphasis on the rationale for the author's disagreement with the conclusions reached by the other members of the Board regarding the carcinogenicity of talc and (2) to provide a realistic perspective regarding the signifi cance of the talc bioassay. 1Presented, in part, at the International Society of Regulatory Tox icology and Pharmacology/U.S. FDA Workshop on Talc: Consumer Uses and Health Perspectives, NIH, Bethesda, MD, January 31-February 1, 1994. 2To whom correspondence should be addressed. Fax: (517) 353 8915. SYNOPSIS OF THE CARCINOGENICITY STUDY OF TALC Groups of 50 male and 49 or 50 female F344/N rats were exposed to aerosols containing 0, 6, or 18 mg/m3 talc, 6 hr a day 5 days per week, until mortality in any 0273-2300/95 $6.00 Copyright '<'> 1995 by Academic Press, Inc. All rights o f reproduction in any form reserved. 244 ANALYSIS--NTP TECHNICAL REPORT 421 245 TABLE 1 Summary of the Lifetim e and 2 -Year Carcinogenicity Studies of Talc" Male F344/N rats Female F344/N rats Exposure levels Body weights Survival rates Neoplastic effects Level of evidence of carcinogenic activity 0, 6, or 18 mg/m3 High-dose group slightly lower than controls 9/50, 14/50, 16/50 Adrenal medulla: benign or malignant pheochromocytoma (26/49, 32/48, 37/47) Some evidence 0, 6, or 18 mg/m3 High-dose group slightly lower than controls 11/50,13/49,9/50 ' Lung: alveolar/bronchiolar adenoma or carcinma (1/50,0/48, 13/50) Adrenal medulla: benign or malignant pheochromocytoma (13/48, 14/47, 23/49) Clear evidence National Toxicology Program, 1993. exposure group reached 80% (113 weeks for males and 122 weeks for females). The survival of male and female rats exposed to talc was similar to that of the controls, and the mean body weights of the animals in the high dose group were only slightly lower (not more than ap proximately 10% lower) than those of controls after Week 65 (NTP, 1993). The results were interpreted as indicating some evidence of carcinogenicity based on be nign and malignant pheochromocytomas in male and fe male rats and clear evidence of carcinogenicity based on alveolar/bronchiolar adenomas and carcinomas of the lung in female rats. The study is summarized in Table 1. Groups of 47 to 49 male and 48 to 50 female B6C3F1 mice were exposed to aerosols containing 0, 6, or 18 mg/ m3talc, 6 hr a day 5 days per week, for up to 103 or 104 weeks. Final mean body weights and survival of the mice exposed to talc were similar to those of the controls (NTP, 1993). The results were interpreted as indicating no evidence of carcinogenic activity in mice. INTERIM EVALUATIONS IN RATS In a parallel series of ancillary studies, additional groups of animals were exposed to talc, as described above, and examined for interim pathology evaluations or pulmonary function tests after 6, 11, 18, and 24 months and lung biochemistry and cytology studies after 24 months. These are outlined below. PARAMETERS ASSESSED IN THE ANCILLARY STUDIES Lung talc burden Pulmonary function Lung biochemistry; bronchoalveolar lavage fluid was analyzed to determine: Cell injury--lactate dehydrogenase. Chronic inflammation--polymorphonuclear leu kocytes, pulmonary macrophages, alkaline phos phatase, and protein. Lysosomal activation--/3-glucuronidase and acid phosphatase. Response to oxidant injury--increased glutathi one reductase. Lung collagen metabolism, protein synthesis, and proteinase activity. LUNG TALC BURDEN, LUNG WEIGHTS, AND RESPIRATORY FUNCTION The lung talc burden at the 6-, 11-, 18-, and 24month interim sacrifice times was similar in male and TABLE 2 Lung Talc Burden (Norm alized to Exposure Concentration) of Rats Male Female 6 mg/m3 18 mg/m3 6 mg/m3 18 mg/m3 6-Month interim 12-Month interim 18-Month interim 24-Month interim 0.439 0.040* 0.731 0.098 1.22 0.12 1.74 0.21 0.602 0.013* 1.165 0.113* 1.53 0.05 1.34 0.19 0.406 0.032 0.785 0.043 1.28 0.06 1.52 0.15 0.464 0.007* 0.787 0.187 1.35 0.04 1.63 0.13 "Table F3, National Toxicology Program, 1993. *Units are presented as mg talc/g control lung/mg/m3. * Significantly different (P ss 0.05) from the 6 mg/m3group by Dunn's or Shirley's test. 246 JAY I. GOODMAN TABLE 3 T otal L ung C apacity o f Rats" 0 mg/m3 6 mg/m3 18 mg/m3 showed statistically significant increases in the females and, in general, the magnitude of the increases was higher in the females. 6-Month interim 11-Month interim 18-Month interim 24-Month interim Male 19.86 0.544 20.06 0.32 20.30 0.45 20.50 0.83 19.48 0.46 18.44 0.39* 18.87 0.41** 20.20 0.28 19.25 0.39 17.67 0.45* 16.34 0.52* 16.47 1.53 6-Month interim 11-Month interim 18-Month interim 24-Month interim Female 14.20 0.25 13.29 0.21 13.94 0.26 14.85 0.31 14.56 0.27 12.91 0.17 12.68 0.28* 13.73 0.34** 13.80 0.27 12.06 0.26* 11.43 0.31* 11.50 1.07* "Table F10, National Toxicology Program, 1993. 6Units are presented as ml; ratio is (dosed group mean/control group mean) X 100. *P=S0.01. ** Significantly different (P 0.05) from the control by Dunn's and Shirley's test. female rats exposed to 6 or 18 mg/m3talc, respectively (Table 2). The absolute and relative lung weights of male rats exposed to 18 mg/m3 were significantly greater than those of controls at the 6-, 11-, and 18-month interim evaluations and at the end of the lifetime study, while those of female rats exposed to 18 mg/m3 were signifi cantly greater at the 11-, 18-, and 24-month interim evaluations and at the end of the lifetime study. In both males and females, there was a concentration-related impairment of respiratory function which increased in severity with increasing duration of exposure (NTP, 1993). While decreases in total lung capacity were noted in males and females, statistically significant reduc tions at the 24-month interim were only seen in females and this occurred at both the 6 and 18 mg/m3 doses (Table 3). BRONCHOALVEOLAR LAVAGE AND LUNG BIOCHEMISTRY Following the completion of the pulmonary function tests at the 24-month interim evaluation a thorough as sessment of biochemical and cellular parameters that could provide insight regarding lung toxicity was made. Bronchoalveolar lavage was performed on the remaining rats in these groups and the lavage fluid was evaluated for enzymes and protein (Table 4) and cell content (Ta ble 5). In addition, lung collagen metabolism and protein synthesis were assessed (Table 6) and proteinase activ ity in both lavage fluid and lung homogenate superna tant fluid was evaluated (Table 7, males; and Table 8, females). While dose-related increases in these parame ters (indicative of toxicity) were noted in both males and females, there was a larger number of parameters that ASSESSMENT OF THE NTP TALC REPORT The NTP Talc Report noted that both in vitro and in vivo studies indicate that talc is not genotoxic and it included a thorough evaluation of lung toxicity (NTP, 1993). However, these data were presented as empirical observations rather than discussed in a manner that would relate them to the risk assessment implications of the bioassay. In other words, relevant data were col lected but not used. In addition, the evaluation of the pheochromocytomas was inadequate because it failed to place sufficient emphasis on the spontaneous incidence of this tumor in rats. These deficiencies caused the au thor to vote against the conclusions presented in the Talc Report when it was reviewed by the NTP Board of Scientific Counselors. The rationale for this decision in the context of a more realistic perspective on the Report is presented below. The carcinogen bioassay is a qualitative test; it is not a risk assessment. The use of biological information is required in order to place results of the bioassay into proper perspective and to take a rational approach to ward risk assessment (Goodman, 1990, 1994). This no tion was endorsed in a recent review of the NTP by the NTP Board of Scientific Counselors where it was stated that "Mechanistic studies should pervade the activities of the N T P .. . . With this emphasis the NTP should be able to strengthen its interpretation of test results in the context of human health. Such interpretation should be come a part of reports issued by the NTP" (NTP, 1992; Goodman, 1994). It is appropriate to consider the NTP Talc Report in this context. Talc-induced lung tumors were not detected in male rats, female mice, or male mice. In rats, the principal toxic lesions associated with inhalation exposure to talc included chronic granulomatous inflammation, alveolar epithelial hyperplasia, squamous metaplasia and squa mous cysts, and interstitial fibrosis of the lung. These lesions were accompanied by impaired pulmonary func tion. While the talc burden in the lungs of males and females was similar (Table 2), the degree of chronic tox icity and inflammation was substantially higher in the females (Tables 3-8; NTP, 1993). In mice, inhalation exposure to talc produced some chronic inflammation. In contrast to rats, alveolar epithelial hyperplasia, squa mous metaplasia, and interstitial fibrosis were not ob served. Overall, markedly less talc-induced lung toxicity was produced in mice than in rats (NTP, 1993). These observations need to be juxtaposed with the findings re garding lung tumors. It is apparent that an increase in lung tumors was seen only in the test animals that clearly exhibited the highest degree of chronic lung tox icity, the female rats exposed to the high dose (18 mg/ m3) of talc. The amount of toxicity leads the author to ANALYSIS--NTP TECHNICAL REPORT 421 TABLE 4 Bronchoalveolar Lavage Fluid Enzym es of Rats at the 24-M onth Interim Evaluation" 0 mg/m3 6 mg/m3 18 mg/m3 Male 0-G lucuron idase6 Lactate dehydrogenase Alkaline phosphatase Glutathione reductase Total protein' Female 0-Glucuronidase6 Lactate dehydrogenase Alkaline phosphatase Glutathione reductase Total protein' 1.09 + 0.40 1634 545 364.7 147 103.03 16.43 1.78 0.40 3.33 0.97 1655 266 427.8 + 30.9 100.6 1.7 1.20 + 0.22 18.86 + 3.20* 3193 + 606 572.8 86.8 99.35 19.79 3.12 + 0.64 41.05 + 4.39** 3906 + 444* 853.6 + 79.7** 135.2 22.4 4.30 0.36** 89.24 14.24** 8262 380* 1604.7 + 143* 110.99 0.55* 5.79 0.55* 154.16+ 17.21** 14E3 1E3** 2594.7 221*** 460.0 44.8* 12.96 0.28** Table F4, National Toxicology Program, 1993. 6Units presented as m lU/g control lung. ' Units presented as mg/g control lung. * Significantly different (P 0.05) from the control group by Dunn's and Shirley's test. ** P i 0.01. 247 conclude that the MTD was exceeded in the female rats exposed to the 18 mg/m3dose. The overall incidences of pheochromocytomas (be nign, malignant, or complex) of the adrenal medulla that occurred in the 18 mg/m3exposed rats were significantly greater than those of the controls (males, P = 0.006; fe males, P = 0.024) (NTP, 1993). However, the spontane ous incidence of these tumors was 53% in males and 27% in females. In the author's view, the criterion for statis tical significance should be P < 0.01. This is consistent with the "guideline" indicating that a compound should be considered carcinogenic if the highest dose produces an increase in a common tumor that is significant at the 1% (P < 0.01) level or an increase in a rare tumor that is significant at the 5% (P < 0.05) level; otherwise, the compound is regarded as a noncarcinogen (Haseman, 1983; Haseman et al., 1986). A rare neoplasm was de fined as a neoplasm occurring with a frequency of less than 1%. Therefore, there is no statistically significant increase in pheochromocytomas in talc-treated female rats and talc should not be deemed carcinogenic at this site in these animals. There has been a pronounced increase in the sponta neous occurrence of pheochromocytomas in male F344 rats, in studies conducted by the NTP, over the past 10 years (Rao et al., 1990; NTP, 1993). In addition, the high rate of pheochromocytomas in the control female rats in the talc bioassay (13/48, 27%) is in sharp contrast to reports indicating that the spontaneous incidence in control female F344 rats, in NTP studies, has been less than 10% (Rao et al., 1990), e.g., in the range of 5.2-7.1% between 1980 and 1983 (Haseman and Rao, 1992). Un fortunately, the NTP failed to make a comparison of the spontaneous pheochromocytomas observed in the talc TABLE 5 B ronchoalveolar Lavage Fluid Cell Populations o f Rats at the 24-M onth Interim Evaluation" 0 mg/m3 6 mg/m3 18 mg/m3 Male Polymorphonuclear cells6 Lymphocytes Macrophages Epithelial cells Female Polymorphonuclear cells Lymphocytes Macrophages Epithelial cells 0.333 0.167 0.000 0.000 93.67 3.3.72 6.00 3.61 0.625 0.315 0.000 0.000 91.38 1.75 8.00 2.01 24.417 2.557* 0.500 0.258 70.25 2.53* 4.83 1.41 25.778 2.673** 0.722 0.188* 71.22 2.95** 2.28 0.50* 32.50 3.000* 0.500 0.500 62.75 1.75* 4.25 1.75 37.000+ 1.528** 1.333 + 0.667* 57.33 4.67** 4.33 2.60 "Table F5, National Toxicology Program, 1993. 6 Units presented as percentage of total cells. * Significantly different (P i 0.05) from the control group by Dunn's and Shirley's test. ** P i 0.01, 248' JAY I. GOODMAN TABLE 6 Lung C ollagen M etabolism and Protein Synthesis in R ats at the 24-M onth Interim Evaluation" 0 mg/m3 6 mg/m3 18 mg/m3 Male Lavage fluid collagenous peptides3 Total lung collagen3 Collagen production13 Collagen degradation* Noncollagenoua protein synthesis3 Female Lavage fluid collagenous peptides Total lung collagen Collagen production Collagen degradation Noncollagenous protein synthesis 39.79 5.07 13.87 0.60 1.58 + 0.17 31.67+ 1.72 142.1 + 14.5 78.27+ 11.64 14.32 0.66 0.982 + 0.185 14.41 +2.44 173.9 + 34.5 46.99 6.51 15.98 0.39* 1.60 + 0.17 27.74 1.42 199.8 22.1* 115.36 8.61* 19.95 1.58* 1.804 0.144* 21.59 4.99 325.8 90.9 79.21 13.73 18.88-3.35* 1.63 0.22 9.18 2.38* 312.2 10.6** 174.71 13.56** 36.47 3.39** 2.264 0.347** 9.38 1.63 554.3 107* Table F7, National Toxicology Program, 1993. bUnits are presented as ng/g control lung. c Units are presented as mg/g control lung. d Units are presented as percentage new protein. *Units are presented as percentage new collagen. 3 Units are presented as dpm X 103/g control lung. * Significantly different (P 0.05) from the control group by Dunn's and Shirley's test. **P;0.01. " bioassay with the historical spontaneous incidence of pheochromocytomas in rats. This would have been par ticularly appropriate (as the author indicated during the Board's review of the NTP Talc Report, June 23-24, 1992). The marked increase in the spontaneous incidence of pheochromocytomas in female rats coupled with the lack of a statistically significant increase of this tumor in the treated animals (when the appropriate P value is employed) and the trend toward an increasing sponta neous incidence in males indicate that the data pre sented in the NTP Talc Report do not warrant a conclu sion that talc causes pheochromocytomas in rats. CONCLUSIONS The appropriate conclusion to be drawn from the NTP Talc Report is that the MTD was exceeded in the female rats exposed to the high dose and talc is not ex pected to cause lung tumors under conditions of exposure that fail to result in marked chronic lung tox icity. Clear evidence of carcinogenic activity of talc was TABLE 7 P roteinase A ctivity in L avage Fluid and Lung Hom ogenate Supernatant Fluid o f Male Rats at the 24-M onth Interim Evaluation" 0 mg/m3 6 mg/m1 18 mg/m3 Lavage fluid Acid proteinase Cathepsin D Cathepsin B Homogenate supernatant fluid Acid proteinase Cathepsin D Cathepsin B Neutral proteinase PMN elastase cathepsin G Macrophage elastase collagenase 0.994 0.329 0.147 0.147 0.924 0.415 10.92 0.64 8.53 0.91 2.39 0.41 0.715 0.168 0.490 0.218 0.225 0.099 1.866 0.174 0.599 0.150 1.267 0.094 17.51 0.90* 14.04 0.62* 3.48 0.37 2.417 0.304* 1.936 0.242* 0.482 0.077 4.307 0.218* 2.420 0.147** 1.887 0.365 25.13 1.50** 21.03 1.56** 4.10 0.06* 4.505e 4.457 0.377** 0.000" *Table F8, National Toxicology Program, 1993. 3Units are presented as mg/hr/mg control lung. 3n = 1; no statistic calculated. * Significantly different (P 0.05) from the control group by Dunn's and Shirley's test. ** P < 0.01. ANALYSIS--NTP TECHNICAL REPORT 421 249 TABLE 8 Proteinase A ctivity in Lavage Fluid and Lung Hom ogenate Supernatant Fluid of Fem ale Rats at the 24-M onth Interim Evaluation" 0 mg/m3 6 mg/m3 .18 mg/m3 Lavage fluid Acid proteinase Cathepsin D Cathepsin B Homogenate supernatant fluid Acid proteinase Cathepsin D Cathepsin B Neutral proteinase PMN elastase cathepsin G Macrophage elastase collagenase 1.52 0.126 0.015 0.015 1.61 + 0.26 14.04 0.95 10.05 0.68 3.99 0.58 0.648 0.087 0.785 + 0.142 0.054 0.037 3.46 + 0.33* 1.310 0.292* 2.15 + 0.22 29.43 + 1.18** 22.97 + 1.07** 6.46 + 0.60* 5.040 0.418** 4.351 0.261** 0.068 0.175* 6.05 0.73** . 4.043 0.578** 2.01 0.17 38.61 1.81** 30.25 1.60** 8.37 0.42** 12.293 1.598** 10.313 2.694** 2.012 1.126* " Table F8, National Toxicology Program, 1993. bUnits are presented as mg/hr/mg control lung. * Significantly different (P < 0.05) from the control group by Dunn's and Shirley's test. **P<0.01. seen only in female rats exposed to the high dose of talc and only under circumstances in which there was evi dence of marked chronic lung toxicity. These remarks are not intended to connote that the talc bioassay is in valid. The good news is that under the conditions of this study mice did not develop tumors, male rats did not de velop lung tumors, and female rats exposed to the 6 mg/ m3dose did not develop lung tumors. Furthermore, it is doubtful that the pheochromocytomas observed in the talc-treated rats were treatment related. Thus, at the high dose, talc exhibited equivocal evidence of carcino genic activity toward the adrenal medulla of male rats. There is a considerable debate regarding the proper criteria for a MTD (Kociba, 1987). However, in a practical sense the chief concern with the MTD is how the resulting data are used in the risk assessment pro cess (McConnell, 1989). Any high dose, no matter how high, that permits animals to live long enough to de velop tumors is not necessarily an appropriate high dose. The recent review of the NTP addressed this is sue and concluded that "the implicit assumptions un derlying extrapolation from the MTD . . . do not ap pear to be valid. Therefore, both the criteria for selec tion of the high dose used and the default criteria that are employed for extrapolation from high dose to low dose must be reevaluated in a critical manner" (NTP, 1992; Goodman, 1994). A consideration of the data presented in the NTP Talc Report leads the author to conclude that it would not be appropriate to use the lung tumor end point in female rats as the basis for a risk assessment involving a linear extrapolation to estimate a theoretical risk to humans that might be exposed to relatively low doses of talc. REFERENCES Goodman, J, I. (1990). The carcinogen bioassay in perspective: A re quirement for incorporation of biological information. In Vitro Tox icol. 3, 1-7. Goodman, J. I. (1994). A rational approach to risk assessment requires the use of biological information: An analysis of the National Toxi cology Program (NTP), final report of the Advisory Review by the NTP Board of Scientific Counselors. ReguL Toxicol. Pharmacol. 19, 51-59. Haseman, J. K. (1983). A reexamination of false-positive rates for car cinogenicity studies. Fundam. Appl. Toxicol. 3, 334-339. Haseman, J. K,, Winbush, J. S., and O'Donnell, M. W. (1986). Use of dual control groups to estimate false positive rates in laboratory animal carcinogenicity studies. Fundam. Appl. Toxicol. 7, 573-584. Haseman, J. K., and Rao, G. N. (1992). Effects of corn oil, time-related changes, and inter-laboratory variability on tumor occurrence in control Fischer 344 (F344/N) rats. Toxicol. Pathol. 2 0 , 52-60. Kociba, R. J. (1987). Issues in biochemical applications to risk assess ment: How should the MTD be selected for chronic bioassays? E n viron. Health Perspect. 7 6, 169-174. McConnell, E. (1989). The maximum tolerated dose: The debate. J. Am. Coll. Toxicol. 8, 1115-1120. National Toxicology Program (NTP) (1992). Final report of the advi sory review by the NTP Board of Scientific Counselors. Fed. Regist. 67,31721-31730. National Toxicology Program (NTP) (1993). NTP technical report on the toxicology and carcinogenesis studies of talc in F344/N and B6C3F1 mice. NTP TR 421. U.S. Department of Health and H u man Services, National Institutes of Health. Rao, G. N., Haseman, J. K., Grumbein, S., Crawford, D. D., and Eustis, S. L. (1990). Growth, body weight, survival, and tumor trends in F344/N rats during an eleven-year period. Toxicolog. Pa thol. 18,61-70. Sontag, J. M., Page, N. P., and Safiotti, U. (1976). Guidelines for car cinogen bioassay in small rodents. National Institutes of Health (NIH), DHEW Publ. No. 76-801, Washington, DC. 16 The Aetiology of Ovarian Cancer M.S. Baylis, W .J. H enderson, C.G . Pierrepoint and K. Griffiths Introduction The estimated number of new notifications of ovarian cancer in 1983 in the United States will be 18200 with 11500 women dying from the disease in the same period [lj. New registrations of ovarian cancer in England and Wales in 1974 numbered 4178 and 5 years later, 3784 women had died of the disease. Such data illustrate the poor prognosis associated with this condition, with less than 25% of women expecting to survive 5 years (Fig. 16.1). The risk of developing ovarian cancer is similar to that for carcinoma of the cervix, 1 in 80, yet as many women die from ovarian cancer as from the other female genital tract malignancies combined. A factor which gives rise for concern is the increasing incidence of the disease. From 1931 to 1970 in England and Wales the mortality rates per million women rose by 15% each decennium [2], During the period 1971-1980, however, the rate of increase had plateaued to 6.5% (Fig. 16.2). The diversity of the structure and functions of the ovary results in a great variety of tumours of complex histological origin (3,4). Tumours of epithelial origin constitute 58% of all ovarian neoplasms and 80-90% of all ovarian carcinomas [5,6]. There is an age specificity for the various histological sub-groups of ovarian malignancy. Sex cord tumours are very rare before the third decade but increase in frequency to the fifth, when their incidence remains constant. Germ cell tumours occur in the 10-50 year old age group, while epithelial tumours rarely appear before the age of 10. They then increase expo nentially until 40 years of age and have their highest incidence in women in their early sixties. Varying nomenclature and different histological classifications have made comparative histopathological and epidemiological studies virtually impos sible. The fact that more than 70 different types of neoplasms have been described I^N- -Gynaecological Oncology ** Death rate/106 women Time since registrohon (y) Fig. 16.1. The percentage survival rate over a 5-year period of women registering with ovarian cancer in England and Wales Fig. 16.2. The death rate of women from ovarian cancer from 1931 to I9W) in England and Wales. highlights the difficulties. The World Health Organisation classification of ovarian tumours 7) should, in future, permit more meaningful comparisons to be undertaken. Possible Causal Factors Environmental Analysis of international mortality rates shows that Northern Europe has four times the incidence rate of Asia [8|. Studies of Japanese migrants living in the The Aetiology of Ovarian Cancer 159 United States [9j show that these Asiatic people exhibit rates of ovarian cancer incidence intermediate between those of their country of origin and their country of adoption. A number of reports have appeared in the literature indicating that certain families show a susceptibility to ovarian malignancy (10J with serous cystadenocarcinomas the most common histological type observed Prophylactic bilateral oophorectormy has been advocated for these women as soon as they have completed their families, although such an aggressive approach does not guarantee immunity to the development of pelvic cancer, which may exhibit a histological picture similar to that of the ovary and therefore supports the concept of `generalised instability' of coelomic epithelium. It has been reported that 5%~14% of women with the Peutz-Jeghers syndrome develop ovarian tumours [11,12,13]. Similarly, the basal cell naevus syndrome is associated with the development of ovarian fibromas, although other cell types also occur (14). The high incidence of ovarian malignancy in the industrialised countries of Northern Europe and North America points strongly to environmental factors playing a significant role in the initiation of the malignant process. A number of studies have shown an association between smoking and the early, spontaneous onset of the menopause [15,16], There was also a relationship with the level of smoking [17] . Studies of the rodent ovary demonstrated the presence of a non-specific microsomal mono-oxygenase, which metabolises polycyclic aromatic hydrocar bons, present in cigarette smoke and industrial pollutant, to active carcinogenic compounds. Benzo(a)pyrene, also a constituent of cigarette smoke, was found to be toxic to the primordial follicles of rodents and to initial ovarian tumours in mice [18] . Mattison and Thorgeirsson found evidence of a similar microsomal mono oxygenase system in the human ovary and proposed an association between polycyclic aromatic hydrocarbons and the occurrence of an early menopause in smokers and also a possible carcinogenic role for the induction of ovarian cancer [19] . It is said that the menopause in women occurs when the oocytes are depleted, which normally occurs by the process of atresia [20]. Early ovarian failure may therefore occur if the rate of oocyte loss is accelerated. There is no evidence at present, however, to suggest those women who smoke are more susceptible to ovarian malignancy than non-smokers. Infectious Diseases Infectious diseases have been considered as a possible aetiological factor in the causation of ovarian cancer. A higher incidence of rubella infection between the ages of 12 and 18 years in ovarian cancer patients than in controls has been reported [21], whilst there was a lower frequency of mumps [22] and of pneu monia and influenza [23], Hormonal It is considered that endogenous hormones play a role in the aetiology of ovarian cancer as, in all international reviews of cancer rates, a strong relationship is known to exist between the incidence of ovarian and breast cancer. Women who develop primary carcinoma of the breast have twice the expected risk of develop ing a primary ovarian cancer and, vice versa, women with primary ovarian cancer IW) (ivniiecological Oncology run 3 10 4 limes the risk of developing a primary breast malignancy. One hypothesis lo explain ovarian epithelial cancer relates growth regulating hormone excess involving the gonadotrophins which act indirectly on the epithelial cells, causing them to replicate following ovulation and cover the denuded surface of the ovary. The epithelium may he incorporated into the stroma of the ovary during the formation of the corpus luteum [24|, resulting in inclusion cysts which, because of the multipotcntiality of the surface epithelium, may give rise to malignant tumours exhibiting Mullerian cell types |25). This hypothesis also suggests that the interruption of the normal cyclical control of the ovary by gonadotrophin would he protective and reduce (he number of inclusion cysts |26|. This would also account for the rarity of ovarian epithelial neoplasia prior to puberty. A number of case control studies support this concept, showing that pregnancy and lactation have a protective role [23,27,28,29). Furthermore, in women under 50. incomplete pregnancies were found to be protective, as was the use of oral contraceptives. Bcral inversely correlated death from ovarian cancer with the number of children born [30). It would seem that nulliparous Caucasian women of northern European descent, who live in an industrialised country and who also have a family history of cancer, are at greatest risk of developing epithelial neoplasia of (he ovary. The incidence of breast, endometrial and ovarian cancer was noted by Stadel [311to be inversely correlated with the dietary intake of iodine, a deficiency which may lead to primary hypothyroidism. Primary hyperthyroidism is associated with elevated prolactin levels [32J and also a 'potentiation' of the effects of gonadotrophins [33J. a situation which might lead to excessive oestrogen synthesis and secretion and a possible overall increased risk of ovarian, endometrial and breast malignancy. Biskind and Biskind [34) demonstrated that ovarian tumours of granulosa cell origin could be induced in the rat, probably as a result of excessive gonadotrophin release brought about by castration of the animal and transplantation of part or whole of one ovary into the spleen. As the blood leaving the spleen drains directly to the liver the steroids produced by the transplanted ovary are immediately metabolised and rendered ineffective in the normal feed back of hypothalamus and pituitary. Studies from this Institute [351confirmed the elevated levels of follicle-stimulating hormone (FSH) and luteinising hormone (l.l I) following this transplantation procedure. This is an interesting model but is of limited value as granulosa cell tumours are rare in women Excessive gonadotrophin stimulation in the rat may also be achieved by immu nising with testosterone 3-fO-carboxymethyl imino-bovine serum albumin)-T-3HSA) [36|. This leads to a grossly elevated, total serum testosterone con centration. with a decrease in free circulating, biologically-active testosterone which is bound by the anti-testosterone antibodies. Exogenous hormones have also been implicated in the aetiology of ovarian cancer [37) in a follow-up survey of 908 women who had received conjugated equine oestrogen postmenopausally. It was also stated that there was a greater risk of developing ovarian cancer in those women who had. in addition, taken stilboestrol. Talc and Asbestos It is of interest that those women employed in the rubber, electrical and textile industries are at greater risk of developing ovarian malignancy than those The Aetiology o f Ovarian Cancer 161 involved in other manufacturing industries [38,39]. The association between asbestos exposure and lung cancer in man was first suggested by Lynch and Smith in 1935 [40]. Further studies confirmed that certain types of asbestos, in addition to their fibrogenic properties, also possessed carcinogenic potential [41]. In 1960, 32 cases of pleural mesothelioma in persons exposed to crocidolite in South Africa were described [42|. Eleven cases of peritoneal tumours associated with asbestos exposure were also reported by Real [43] and a further nine by Entiknap and Smither [44). Moreover, Graham and Graham [45] reported the presence of birefringent crystals in human ovarian cancers and related them to asbestos but they were unable to characterise them. Certain ovarian tumours show histological similarity to pleural and peritoneal mesotheliomas. This may be explained by the fact that the ovarian surface epithelium is of mesothelial origin, which confers upon it its unique ability to produce ovarian tumours of multiple ceil types including combinations of serous, mucinous, endometrioid, mesonephroid and mesothelial cell types [46). An extraction replication technique was developed in the Tenovus Institute to investigate foreign particulate material present in biological tissues [47]. This technique was used to investigate and identify various types of contaminating particles in sections of tissue from, for example, the lung and ovaries [48]. It failed to demonstrate the presence of asbestos in human ovaries, as suggested by Graham and Graham, but did establish the presence of talc particles in both normal and neoplastic tissue [49|. Unequivocal identification of individual par ticles was achieved by using an electron microscope microanalyser (EMMA). There is good evidence that talc may reach the ovary by traversing the female genital tract. Egli and Newton [50] instilled carbon black particles suspended in a 30% dextran solution into the posterior fornix of vaginas of three women under going hysterectomy. At laparotomy, in two of the three patients, carbon particules could be found at the fimbrial ostia of the fallopian tube within 40 min of the vaginal instillation. Further confirmation of the migration of particulate material from the vagina to the peritoneal cavity, with uptake into the ovary, was provided by studies using technetium-99 labelled human albumin spheres instilled into the vagina for the purpose of hysterosalpingography by radionuclide scin tigraphy [51,52]. Work from this laboratory [53] has shown that when a suspen sion of talc is introduced either into the posterior aspect of the uterus or even into the vagina of rats, the material can be located in the ovaries within 4 days of the instillation. Talc is widely used in the pharmaceutical and cosmetic industries and has been used in association with certain types of contraceptive. Many women dust their perinea with talc and some apply it to their sanitary wear [54], Furthermore, talc particles have been shown to embed in both cultured human lung fibroblasts and in rat ovarian epithelial cells, probably by the process of pinocytosis. Oxygen incineration studies of human ovarian tissue indicated that the amount of talc present could be of the order of 2 x 105 particles per mm3 [55], Reports that commercial talcs are contaminated with particles of asbestos [10] have not been substantiated in this Institute, where nearly all forms of natural talc produced from various countries around the world have been analysed. It is well established that talc can induce granulomata in the peritoneal cavity [56] and contamination of the ovary may result in disturbances in the ovarian stroma leading to altered steroid biosynthesis. In this laboratory, a suspension of asbestos-free Italian 00000 talc in phosphate- r & A** f I (ivnaccological Oncology buffered saline w;is injected below the rat ovarian bursa |57|. Within 6 weeks of injection many animals demonstrated cystic changes which, by I year, had developed into ovarian bursal cysts. 5 cm or more in diameter (Fig. 16.3). Histo- Fig. 16.3. Bilateral ovarian bursal cysts from a rat 12 months after the sub-bursal injection of a MiNpciiMoii of talc. Fig. 16.4. Rat ovarian surface germinal epithelium subsequent to the sub-bursal injection of a suspension of talc. Note papillary formation and the multilayering of the surface epithelium. Stained with II & E. (x 200) The Aetiology of Ovarian Cancer 163 logical examination of these talc-treated ovaries showed multilayering of the surface germinal epithelium (Fig. 16.4) with papillary outgrowths, a pathology not inconsistent with preneoplasia and which resembled the histological findings in the ovaries of guinea pigs treated with intraperitoneal asbestos [4SJ. The pos sibility that contaminating talc particles may play a role in initiating ovarian dysfunction, which could lead to neoplasia, requires careful consideration. The possible adverse affects of presence of talc in the ovary are being intensively investigated in this Institute. Several different animal species are being used, including the chicken, which is known to have a high incidence of ovarian adenocarcinoma when kept in artificial light |2J. The possibility that talc may act as a co-carcinogen is being examined. Its adsorptive properties are well known and, once located in the ovary, may retain circulating carcinogens for longer periods than normal. Equivalent studies are being carried out in vitro since it has been shown that talc will penetrate epithelial cells in culture and, although transformation of the cells has not been shown, the addition of a chemical such as dimethylbenzanthracene, which does not usually induce ovarian cancer, may well induce such changes. It certainly cannot be said at present that talc causes ovarian cancer. Its presence in such tumours was unexpected and unwarranted. It is because the cause of ovarian carcinoma is unknown and that talc has a chemical similarity with asbestos that these investigations were instigated. References 1. Silverbcrg E (1983) Cancer statistic 1983. CA -- A Cancer J Clinicians 33:9 2. Campbell H (1975) In: Brush MG. Taylor RW (eds) Gynaecological malignancy: Clinical and experimental studies. Williams & Wilkins, Baltimore, pp 111-132 3. Wynder EL. Dodo H, Barber HK. (1969) Epidemiology of cancer of the ovary. Cancer 23:352- 370 4. Russell P( 1979) The pathological assessment of ovarian neoplasms 1. Introduction to the common 'epithelial' tumours and analysis of benign epithelial tumours. Pathology 11:5-26 5. Doll R.MuirC, Waterhouse J( 1970) Cancer incidence in five continents. Vol. II Springer. Berlin, pp 29-37 6. Katsube Y, Berg JW, Silverbcrg SG (1982) Epidemiologic pathology of ovarian tumours; a histopathologic review of primary ovarian neoplasms diagnosed in the Denver Standard Metro politan Statistical Area, 1 July-31 December 1969 and 1 July-31 December 1979 Int J Gyncecol Pathol 1:3-16 7. Serov SF, Scully RE, Sobin LH (1973) Histological typing of ovarian tumours, No 9 World Health Organisation, Geneva 8. Waterhouse J, Muir C, Correa P, Powell J (1976) Cancer incidence in five continents, Vol 111, IARC Scientific Publication No 15. Lyon, pp 453-485 9. Haenszel W, Kurihara M (1968) Studies of Japanese migrants 1. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43-68 10. Lingeman C(1974) Etiology of cancer of the human ovary: A review. J Natl Cancer Inst 53:16031618 11. Dozois RR. Kemper RD, Dahlin DC (1970) Ovarian tumours associated with the Peutz-Jeghers syndrome. Ann Surg 172:233-238 12. Scully RE (1970) Sex cord tumour with annular tubules. A distinctive ovarian tumour of the PeutzJeghers syndrome. Cancer 25:1107-1121 13. Christian CD (1971) Ovarian tumours: An extension of the Peutz-Jeghers syndrome. Am JObstct Gynecoll 11:529-534 IM Gynaecological Oncology 14. Berlin NI (1966) Basal cell naevus syndrome. In: Combined clinicial staff conference at the National Institute of Health Ann Intern Med 64:403-421 I.V Daniel IIW (1976) Osteoporosis of the slender smoker. Arch Intern Med 136:298-304 16. Baily A. Robinson D. Vessey M (1977) Smoking and the age of natural menopause. Lancet 11:722 17 Jiek H. Porter J. Norrison AS (1977) Relation between smoking and age of natural menopause. Lancet 1:1354-1355 18. Jull JW( 1973) In: Busch J (cd) Methods in Cancer Research. Academic. New York, Vol VII.pp 131-186 19. Maltison DR. Thorgcirsson L (I97K) Smoking and industrial pollution and their effects on the menopause and ovarian cancer. Lancet 1:187-188 20. Ingram 1)1 (1962) In: /nckerman S (ed) The ovary. Academic. New York. Vol I. pp 247-273 21 McGowan !.. Parent I.. I.cdnar W. Norris 1(J (1979) The women at risk of developing ovarian cancer. Gynecol Oncol 7:325-344 22. West RO (1966) Epidemiologic study of malignancies of the ovaries Cancer 19:1001-1007 23. Joly DJ. Lillienfeld AM, Diamond EL. Brass IDJ ( 1974) An epidemiologic study of the relation ship of reproductive experience to cancer of the ovary. Am J Epidemiol 99:190-209 24. Radisavljcvic SA (1977) 'ITic pathogenesis of ovarian inclusion cysts and cystomas. Obstet Gynecol 49.424-429 25. Mertig AT, Gore II ( 1961 ) Tumours of the female sex organs. Part 3 Armed Forces institute of Pathology. Washington. DC, p 76 26 Fathalla MF (1971) Incessant ovulation -- A factor in ovarian neoplasia. Lancet 1:163 27 New house ML. Pearson RM, Fullerton JM. Boresfen FAM, Shannon MS (1977) A case control study of carcinoma of the ovary. Br J Prev Soc Med 31:148-153 28 Annegcrs JF. Strom H. Decker DG, Dockerty MD. O'Fallon WM (19^9) Ovarian Cancer: Incidence and case-control study. Cancer 43:72.3-729 29. Casagrande JT. Louie EW. Pike MC, Roy S. Ross RK, Henderson Be, (1979) incessant ovulation' and ovarian cancer. Lancet 11:170-172 30. lierai V. Fraser P. Chivers C (1978) Does pregnancy protect against ovarian cancer? Lancet 1:1083-1086 31. Siadcl BV (1976) Dietary iodine and the risk of breast, endometrial and ovarian cancer. Lancet 1:890-891 32. IngbarSJ, Woebar KA (1974) In Williams RH (ed) Textbook of endocrinology 5th ed. Saunders, Philadelphia, pp 95-232 33. I.ongeopc C ( 1975) Breast cancer and cstriol dynamics: Program of the Breast Cancer Task Force- National Cancer Institute. National Institute of Health. Washington. 34. Biskind MS. Biskind GR (1944) Development of tumours in the rat ovary after transplantation into the spleen. Proc Soc Exp Biol Med 55:176-179 35. Scagcr SE. Boyns AR. Blumgart LM (1974) Histological and hormonal studies in rats bearing ovarian implants in the spleen. Effects of portocaval anastamosis. Eur J Cancer 10:35-40 36. Millier SG, Groom GV. Boyns AR. Cameron EHD (1974) Development of polycystic ovaries in rats actively immunised against T-3-BSA. Nature 250:433-434 37 Hoover R, Fraumcni JF. Gray LA ( 1977) Stilbeostrol (Diethylstiboestrol) and the risk of ovarian cancer. Lancet 11:533-5.34 38. Manmcn TF (1975) In; Levinson C (cd) The new multinational health hazards. International Chemical Foundation. Geneva 80 39. Campbell Jti (1969) Tumours of the fowl. Heinemann Medical. London pp 171-200 40. Lynch KM. Smith WA ( 1935) Carcinoma of the lung in asbestos silicosis. Am J Cancer 24:56-61 41. Doll R (1955) Mortality from lung cancer in asbestos workers. Br J Ind Med 12:81-86 42 Wagner JC. Sleggs CA. Marchand P ( I960) Diffuse pleural mesothelioma and asbestos exposure in Norih-Wcstcrn Cape Province. Br J Ind Med 17:260-271 43. Keal EE (I960) Ashcstosis and abdominal neoplasms. Lancet 11:1211-1216 44 Entiknap JB. Smithcr J ( 1964) Peritoneal tumours in asbesiosis Br J Ind Med 21:20-31 45. Graham J. Graham R (1967) Ovarian cancer and asbestos. Environ Res 1:115-128 46. Parmley TH, Woodruff JD ( 1974) The ovarian mesothelioma. Am J Obstet Gynecol 120:234-241 47 Ilentlcrson WJ (1969) A simple replication technique for the study of biological tissues by electron microscopy. J Microsc 89:369-372 48 Henderson WJ. Gough J. Ilarse J (1970) Identification of mineral particles in pneumoconiotic lungs. J Clin Pathol 23:104-109 49 Henderson WJ. Joslin CAF, Turnbull AC, Griffiths K (1971)Talcand carcinoma of the ovary and cervix J Obstct Gynecol Hr Commonw 78:266-272 511 Hgli GE. Newton M (1961) The transport of carbon particles in the human female reproductive tract. Fertil Steril 12:151-155  flfc o * The Aetiology of Ovarian Cancer -- .165- 51. Venter PF. Iturralde M ( 1979) Migration of a paniculate radioactive tracer from the vagina to the peritoneal cavity and ovaries. S Afr Med J 55:917-919 52. Iturralde M, Venter PF (1981) Hysterosalpingo-radionuclide scintigraphy (HERS) Semin Nucl Med 11:301-314 53. Baylis MS. Hamilton TC, Henderson WJ, Pierrepoint CG. Griffiths K (to be published) The demonstration of the migration of talc from the vagina and posterior uterus to the ovary in the rat. Environ Res 54. Cramer DW, Welch WR, Scully RE, Wojciechowski CA (1982) Ovarian cancer and talc. Cancer 50:372-376 55. Henderson WJ. Melville-Jones C, Wilson DW. Griffiths K (1978) Oxygen incineration and electron microscope X-ray microanalysis of mineral panicles in biological tissues. J Histochem Cytochem 26:1087-1093 56. Lichiman AL. McDonald R. Dixon CF. Mann FC (1946) Talc granuloma. Surg Gynecol Obstet 83:531-558 57. Hamilton TC, Fox H. Buckley H, Henderson WJ, Griffiths K (1984) The effect of talc on the rat ovary. Br J Exp Pathol 65:101-106 .< I G ynaecological O ncology Edited by C. Paul Morrow and George E. Smart With 75 Figures and 110 Tables Springer-Verlag Berlin Heidelberg New York Tokyo C. Paul Morrow, MD Department of Gynecologic Oncology UCLA School of Medicine Los Angeles California, USA George E. Smart, FRCOG, FRCSE Department of Obstetrics and Gynaecology University of Edinburgh Edinburgh, Scotland, UK ISBN 3-540-15775-1 Springer-Verlag Berlin Heidelberg New York Tokyo ISBN 0-387-15775-1 Springer-Verlag New York Heidelberg Berlin Tokyo Library of Congress Cataloging-in-Publication Data Main entry under title: Gynaecological oncology Includes bibliographies and index. 1. Generative organs. Female - Cancer. I. Morrow, C. Paul, 1935- . H. Smart, George E., 1935- . (DNLM: 1. Genital Neoplasms, Female. WP 145 G997] RC280.G5G86 1985 616.99'265 85-17253 ISBN 0-387-15775-1 (U.S.) The work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically those of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photocopying, machine or similar means, and storage in data banks. Under 54 of the German Copyright Law where copies are made for other than private use, a fee is payable to 'Verwer tungsgesellschaft Wort', Munich. by Springer-Verlag Berlin Heidelberg 1986 Printed in Great Britain The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Product Liability: The publisher can give no guarantee for information about drug dosage and application thereof contained in this book. In every individual case the respective user must check its accuracy by consulting other pharmaceutical literature. Phototypesetting by Computerized Typesetting Services Ltd, North Finchley, London. Printed by Henry Ling Limited. The Dorset Press. Dorchester 2128/3916-543210 m <nfrIM .iMui mimw.h .ir v in ' -, W t 'MiM nM i -VH ^ h t r r i avnW ATiO k s i i / mi x pfll m as - Brand ol c o n iu g an o m o g a n i IAfilli USP Pfl(MARK Blind ol coniugaled estrogens Vaginal Cu m , In a nontiqutlylng b in I I^W tliN M iA V l H H N Ilin m illM lH h i IIIA.M lli| W.SkW IM KIM I IHlAi CARCINOMA I I i.r jt vein ill in > id .id wiHi m i l.ikniy i-sliiKgriis I uit|K-ljiil Adeguale Ov-gnostii r t .iw I Vm n 'htni.Mii.il \.ini|Mnuj aIm i ruihi.il,-a s'- iiM he uihteiliken ip lob* Oui njiig. e I <ii Ali i I m1' ! ondtaquoMnl in/i Irnl oi ii'tinii"v .ibi--mal v.lqiual lifc-rthity l* i U- r. I j iik iIy '*1 1 *nMkt- iim i iMHit.il rsliugi-m * mnr is Irsi hazardous lh.nt 'nihrt rshour"* j l ixjiMlvItiigrtiH' tkv.r'; . r.inm ;iiiSSMiiuiiNiHl u i hi guano I h,,) iu t .hu duiuiy ih iiyoan>i is u-u mV>I *W i - mneasrjiis uli.oqeniiai deled* m IN- if|iH*i ii*i' itujJ"' Ml H i '".ilf -UO lr "ale M u*, an mi leased nsk ol agmji *Vrv> .|ii I"- h- 1ri. (ViiU -.-j I Hu- iiii-imc %cimi .tnd vaginal i t " ri m the binale latem kle Ihr )-tn`. I-I N task l n . t (imi litilnl llul mi ncn fui used IMS ihwniq lln-n lurgnanCKN may vili etiwiMlf fi|H"ifli. f > i tmli asini lis* ni Ui-asl i antri llu -n ,l a ( JuSjI u-t-liO'ship rs i.ii iUKUiMini ml lln- iHiMKil) ll'.rl id list, is Ululai 10 Hi.ll lui a ilumlnn ni olhd Infisi , I., fi iisk Ij. Has lu .i- mu naia .iimii Im fMingi'n fln-upy rt,inig t-cgium y I suwyms e meMi-clive lui I l f m i ftii' M lit-jinirni irf llin-.iliTii-i1 m haUlual jlaalBm l>ns-<|ri.isla i lump-S Aihlilhin.il risks li a aitvrise rllin I un LaibUliyOiale ind Hud melabO liMU in jf In- .inmx ijtuil mill lln' in>iumii ul piogeslm *n esliixgr'i irplaceinenl legnnen* I* i.lwtc ui pmgcsim ami ilosayc m.* be iiupuHani m imouiii/inQ ihrsr adverse ellecis lu rm a i i >.imiiMlifin an a fnni|ilrlt medical anil lamily Nsiwy should Or liken yum to Ha miliahun 0> joy r*tioi|c" Iheiapy Adn spy tal iplnera r lo bkiort presser breads aMomeo and lid. Oigans .uni :.nouM msludr a rai'aiui ulami smeai A a yem-i al iu*e esiiugen should noi be pievnbt-d lui longei man om- yrai wdliuui aixiWei pnysu\H nanwMlion being peilo/med Com> lions uiaurnrsrd by "mo lelmiwn n * n amma. epiieti) ungiaine. and caidiac m lend dysliMC lion irqiiiie Caiclul Obst-niiUMi Crilam palienl* may develop minifrslAlionS of CiCfSSne eiuooe'iii siunyialion. h h as juinmai a i-iressive uiiiaie Needing and masiodyina hr ensimg ulciinr lemo-imnaii may m o risr m i/r duiuig rslmgen use Isiiogeos should Or usd min raie m pjhenis m>iii mguneo lu n lunfixm im d iiisu ik*m y m meiihoiiL bone ebseises isooiied with 'irDfH ih r n..j Ihr loikming d'iMi/iiboiaimy irsi mieiaciious nave been nixled Son* only *Hh esnogen iwoqeMin conMxnalmti m al Loniiau-.pli.es) I m cim cd pndhiombm and lacims W nil a and k detleased aimmionibMi 3. imieased mxepineP'U'or induced ptalrlci agyieijabuu*. fttftejsed honnuiysi ? h" leased lliyionl fxodmy glOOu(ui ilB l.i leading lo increased ciNuliliny lolii lliyiod hoimon*. iS measuipd by , KmcIs iWrli-immed by column v by iadwimniunoasa< l i r e 1, iesm intake s deceased reneclmg Ihr elevated IBO her I . urn mnaUm is unalienrd DESCRIPTION I'MlMAI' ifiai|-ig.ilrd r-.lunjri'S l<S*`l .unljiii .i nuiliHr ill t-sbuyens uhlauird 3 Impaued glucose iNfiancr ru i-.i.rh In ..... muh. I -. bVtUh'd In lifnrM-lil Ihf avi t.nh i tli|aiMh.ui ul (ii.iIi.tmI ilffivirl 4 KiThk rd irspnn.sj- lo iiic1vi.i|iiii' li-sl mu i-i-iriW males iiiim- ll i . -Mirti esimi-<.| i'liti Wit li'ii rltliyiliiN giillui l.igi'llu-. aim s i a>nisails i i I / . , esludi.il it|uiI<-ii>h .Uhl I .l.l.yiliih d'i'ifin n a s sa ils ul N'" sudale esi.-.s lai-, (s a.aitaOir I H I >iuj 0 mg i n I . m j .imi ! ` i mg sili ii.j|ii. ni <iii|u g.ili esii.-ui-iis i learn is a>a<' H- as I' It.** my . uigjleil eshiigecS K> gum MDCAllCWS ANO USAGE MOdrule Id se.erf t- --*-1- iipniii-N a-iMk ijii-d Aiin ihr memi p -l illirti is no e>"h-in r inal iNlrogrns aie elle loi nei<OuS .mploms u< rtfpu-.suin Al"il> niighl oct ut duii'ig nirmpause and Mie/ d-.*uhi noi be used lo lira! Ht -t i ni-d>i7i i Pie PREMARIN (conjugated estiogens tablets, USP) S Kedm ed sc-mni (N ile umcanKilion MUTAGENESIS ANO CARCINOGENESIS: long m m . continuous adniim stiiiion ol n ilm il ind syiiliii-nc csln^ens cm ain inmiai speivs m cieases Ih r iiequem y ol caii.inom* oi if*e tucavi. cevn vaymi. ami k m PGEGNANCT CATEGORY I t shoguns should ON lie used (k3 picgnmcy See CIMTRAMOCAnrais ind Boted Mvikng NUASMG MOTHERS: As i geneial pnncipto Ihc adminithalioii ul any drug lo nuiunj moineis should be done only hen deiify necessity since many drugs i i r eacitlcd in human milk TU!' and ina".iyr(iii'nl 1 u.i* (4klbi-.iv d.ikmally lom bone mJM Avutiti* .tipmli. ll*|.d rsntvjrntsm due io no.'" mir li Cj-vIi jlk.n ni PNinaiy i>va>ai> laiiu'r Pi<| MAHiN iCun|uQalril esiiiiyr-n>i V.i juui I'cam is mdK alni mlhe In-alitn n ! . i i iag AOVfRSC REACTORS: H loRowng hc been itpckied -ran esnogenc hetapy changes m vagi n il bleeding pallein and abnoimal anlhdiaal bleeding ot do NeakihiOugn oleedmg sporting, mcieas* w e oi oieni Mxomyomaii. vagmal candidiasis cnaoge m amounl ol (meal sea lion, ienowness o< r>eaigemmi ol breasts, oau PWMAiWi nAS Nll! W IN SWNTS Ul III 11 M Clivi lOH AN 1it*i'i ist inh in ; n it i .mani i Aid) I l i USI MAi (.Any stvt III f.AI.-W lu 1"1 1EIlei (Sii HUMU WAHNIN..I 11* j[i|TifaitCi' n( Uns uMcl is a l/ad n n i/k oi W>elh Ayrrsl Ijbcualorirs sea. vommng abdommai damps, bloaimg. r.hoirsiahc liondice. chloasma or melasma mai may prisisi *h m Aug is Asconnnudd erythema niuliiloime eiyinema nodosum, hemonhagic CONTRAINDICATIONS tsu .-in 1- Nk.ld uni br urd wiiiiten un , I.- 1*4 IrtNuwiI.Q laWMA-N 1 * " i* " ni gal 1il pmpiaiH t i-i r laurei Waihingl * knuA" - mi (*s M I . aiu i>> ld me tarasi n i e|>l "1 Ip|ui4*i"li-I|!M Vi ti ll lutil iid\ bring lirjlp.l h nU.l.it ih J-J- l kl'iM'l ni Sir>l* In i i. 'ICVi-'-h '" .i k - i K - a - 4 i m d gih sr.1 abiiiii.na> grulla* Uhrjing S Ai hIVI- OSTEOPOROSIS. The only cure is prevention. eiuphon. loss ol scalp nan. hnsulism. steepening ol corneal curvaiuie mioieiante loconlaci tenses, headache migiame. A/imrss menial depeu<on chorea, morase m decrease In ne^jhl. reduced calHinyAle lolniance. aggiav.ilion ol porphyiia. edema changes m wtooo ACUTE OVEROOSAGE: May Liuse nausea and voimimg IhiiunbOphlrlliliN ni Ilauniti"-aihian dr.M'nVis ' DOSAGE ANOAOMNSTRATION. 6 I lliigri. n-pl rOiiHl Ifn-i ||>| lu Ii.rf N-I-II if-|iMI.Tl III in. li j%r lln- nsk ill iniomliopliUitHh and' PREMARM* Brand ol con|ugatad a Diogeni labials. USR H*1 (I'vrtai'. i|rtiv*..THi Ik a...... .. nilWAIliN I.iW* I-. .11-1 biiyii.il I I.-.I.*. .m.im Ik- u-.i-d in yi.iheul lir|ii-isi'n-.ilivv In IbrH WARNINGS *w In.In-. .. (vrl J ler-.silik- nil irdspiluii iO*t v ol Ckr-isl I alu.nl woinrn taking h>.yiiiT ibr-i's >d i~.|iieTi > |ii*ao"giTl un* ('iKid-. Ilie uuip'iiiy * \imlie navi- ijl snowo an V . 'ainin Adi' >iSu.ii fsliinjrii id.* enu'nl ibea-> i i.lwiieiiial i jiilT irsk jiikiiiii --.lirgeii uswi a al..i ii i Iidd in Iju-.ilr.-i i l m ............. . .imi m-dii ilepiTNlciil or>lieJlmenr JWdlKjn and f Ini^i li dor hi pahriil on i oA'teoiTl (-lingen pnigfslin Ihfiapy Hus n>k appear lo be Oe I tr.i-.cn IV r till l.Alili-,IN., (irluni I ishogii. 11`rijpy Oui'iig liM-g i.in y l i a-isKiuied aiIIi an mcieasnl usk O' hrlkl congenital ii-|j.iVji N e Irai 1J-vinMi A /* T>talli im i- m ' i" llur n-vk id ungi My i iHibimnl y.ill blaikk-i iliv .i-.f bi amihe" irCfivilig I--.( l upai:. il jbijivi\ h Ia-i -i ii-pi-li il I g* dose ni e .iim jrn s n I as IIiw- uswl lu heal pnnialc and turasi l in te r have bee" slown in n -it j r IN 11- k i*l ni U l.il uivif M(VI ud or liivi y e tiilxi -.in ai1Iniunibiipnlcbi'i tn ITH-11 lh-s ' ai,**'l I1' > * ( br 1 ill fih i Uri! (( ' f i nikwrvri III (mil thrurll* al I WtbilvaM'U la' nsk (k.vrd In l-vJC r s lh v i, .k..rs I ' r ibises M .-*.Mivyrn ir|ilasf'iie"l Htetapy \h'n*M nul rtireiJ Ihr im uineitdr-l rt-iW HI ivi pii-. s.'r sNc.iiJ tn- MuMl'i'i'il wild rjlingrii u-< es(w\ rally il Ittgli ibra-s air nsml I Ungi-ns m.ir h-.nl In ..i-M-ir h|a-n at i-uua (Im-.I-. aiIIi lut-.tsl i amr< .livlliivw nirlaslasrs PAfCAUinNS IN- aibkiii h* a |mnj-.iiii b- - ia non- days ol a i yell* id rs!UB|e .idiiuni.slial--i lr|--Ifilly l|MU-IS Ibi " I ll III rlt.lmii'h..|l !<(Iflla'.M SllWki's* I Inll.Hielluall Sllligi'Sl IlS-fl III 1 U U,- I .. -g-tpii ir di-. ih-.l In utiliid' ilia"" -i 'iikMIfMial "inluialnvi and 'liimnaltiin ul I An cm'hfAHr loi x/vxi in m usa iHf Ha iiouhnanl ol modaula lo ovaie visornotor symp lom and/oi aimpim v.^pmiij a.MOcuiod anlh Urn menopause |fl 3 nig lo I ?S mg or m x i daUyt In kmesl dar mal ai" cimbri symptoms simuld be Chosen and merbealien ihoutd be Oscon imuwl as promptly as possible Admimsnabon should be cyclic teg irec weeks on and one ween oil) Aiiempis io discontinue or lapn meduaiion should be made al i n w lo sn month Inierviis V G>e*i rrrocany Myimesnogniisin Osimgiornsis Hypoosuogenism dor In <mntle lirpugonaoism - ? b mg lo 7 i mg daily m durided doses lot ?1) days loliowed by <0day rest period HNoeOng does noi occur by me end t i Ihn panod. the same dosage schedule is icpealed Temale casiiabon ut pnmary ovanan laiima I ?5 mg daily, cy clically Ai|us< u|M>aid a oownwaid according lo respoma ol lhe panent Tor mamianance. adpii' dusage (o hiaasi livel lhal ai" proviOr ollrilivt conliol Osieoporusis U 6?T> mg ditiy Admiinsiraiioii himio on cyclic leg ih>eC Aeeks on and one week nil) PREMARW* Brand or conlugaiad atlrogant Vagmal Craam Giteli i'filn a iif lot i fia t lerni use <uny I o He-imeni ol altophic vaguutis or kramOSri vulvae Ine kmesi dose mai will r.uriliot ymptuitis sbOuNl be rhosen and medicanun shouid be dtscon imut-d as ptompily a.%poss>(iie Ademp* io disi i-iimur m laprr iiicd-.aiiui sijio be iiiitOe al lince lo tu ntuoili mleivais Usuai dnsagr i auge ? u lu 4 g da<l> iniiaagialiy ilependtng un itu> seveuty ol Ihe condtlwn l'aueMS Adh jn uiiai I itlcms Alm jtc biaii-ii wun edti IRIMAKM U bid o< VaQinM Deam slutukl be Iiuanlmi-il li suytis il ciakainbi.il fin i ri .atd appropriate measincs Unno lo mie Dui li.tl<Uliai<cv m tli- ceni ul pviMvIriil fwtt-t ijiiiii) tbin-tual v-gnial lii-filnig CI li"IB b May 13 IS'J1 (.1 IRIS 6 AngusI .11 y'c* Reference: I. l'renufin* l'eckjizr Iny-il World l emlership in Woman's Ileniti) Cio,,,m \ WYETH AYERST LABORATORIES MntlluttlliiA. l 'A l `M01 Molecular and Biologic Factors in the Pathogenesis of Ovarian Cancer Jonathan S. Berek, M,D. Otoniel Martinez-Maza, Ph.D. The deiH'lopmetif amI progression of epithelial ovarian earn er can he correlated with iwarns biologic and mole cular factors. Tumor growth has been associated with aberrant and dpsfunctional expression and mutation of various genes. These genetic defects include oncogene overexpression, amplification or mutation, aberrant tumor suppressor gene expression or mutation, and the inappropriate expression of cytokines anil growth factors and/or the cellular receptors for these molecules. Dysrg ulation of host immune responses may also play a per missive role in the pathogenesis of the disease. Since ovarian cancer has been associated with the frequency of ovulation, the related proliferation of epithelial cells may increase the chance of a genetic accident that could contribute to the activation of an oncogene or inactiva tion of a suppressor gene. These events, combined with the inherent ability of ovarian epithelial cells to respond to and produce various cytokines and growth factors, could promote oncogenesis. (J Reprod Mod 39:241-246, 1944) Keywords: ovarian neoplasms, ovary. I'roin the IX-parimcnt t*i Obstetrics and Gynecology, Univeriily of t .llilornia at I AugcXs Medical Center. Ia*s Angeles, Caliiomia. Address reprint rr.|iu*ts to: |nnathan S. Berek, M.D., Depart ment o i Obsletri.s and Gynecology, Room 24-127 CHS, Univer sity of California at I.os Angeles Medical Center. 10833 LeConte Avenue, U>s Angel.*, t A `HKI24 Introduction Ovarian cancer, which results in a higher mortality rate than any other gynecologic malignancy, is the fourth leading cause of cancer death in women in the United States.** A major factor in this high mor tality rate is that at the time of diagnosis most patients already have advanced disease. Early detection of ovarian cancer is hampered by a lack of appropriate tumor markers, and clinically, most ovarian cancer patients fail to develop significant symptoms until they reach advanced disease. Ovar ian cancer has a high frequency of metastasis yet generally remains localized within the peritoneal cavity.11 Tumor development has been associated with the aberrant, dysfunctional expression a nd/or muta tion of various genes. This can include oncogene overexpression, amplification or mutation, aberrant tumor suppressor (antioncogene) expression or mutation, and the inappropriate expression of cytokines and growth factors and/or the cellular receptors for these molecules. Also, subversion of host antitumor immune responses may play a role in the pathogenesis of cancer.1*u '> Most ovarian cancers appear to arise from the ovarian epithelium, a single layer of cells found on the surface of the ovary. While several factors have been implicated in the genesis and growth of ovar ian cancer, several studies have shown that the fre quency of ovulation is associated with an increased risk of the development of ovArian cnncer/45V1!W-t`5 Because ovulation is associated with disruption of the surface epithelium of the ovary, the ovarian epithelium must proliferate to heal this ovulationassociated wound. Therefore, ovulation must be associated with the production of growth factors that enhance the growth and/or differentiation of ovarian epithelial cells. Certainly, various cytokines and growth factors, including transforming growth factors a (TCF-a) and IL-6, have been seen in ovari an follicular fluid.20-53-56 Perhaps the repeated pro liferation of ovarian epithelial cells associated with ovulation contributes to the development of ovari an cancer by increasing the chance of a genetic acci dent (error in DNA replication) that could con tribute to activation of an oncogene or inactivation of a tumor suppressor gene. Such an event or events, combined with the inherent ability of ovari an epithelial cells to respond to and produce cytokines and growth factors, could explain, in part, the high frequency of ovarian cancer and its association with ovulation. 0024-7758/44/3404 0241/$! 50/0 Tho journal of Reproductive Medicine. Inc. 242 77r<- Journal of Reproductive Medicine Oncogenes Hu- aberrant expression of various oncogenes in ovarian tam er has been described They include HliR-2/nrn, nr-, nri/c. (ms, un and mi//).'1 7 -,"'*i-75 The many oncogenes that have been characterized can be divided into a lew major groups based on their biologic role and cellular location .4 ,s For instance, several oncogenes, including HER-2/ hcm and /ms. appear to enctnle for transm em brane receptor m olecules involved in ligand binding. Other oncogenes, including ras, represent inner membrane proteins involved in signal transduc tion. Finally, som e oncogenes, including wye. my/1 and inn. represent nut tear transcriptional regulato ry proteins. Therefore, protooncogenes represent various tcllular proteins with normal physiologic tolis in various cellular processes involved in the induction oi tcllular growth and differentiation. These activities are subverted by mutation or over expression. resulting, in part, in generation of the transtormod phenotype. Recently, the role of I IHR-2/ncif in ovarian cancer has received much attention." '''"*I IEK-2/ucu is an oncogene product that was seen to be overex pressed in breast cancer; overexpression o f HER2 / hcu in breast cancer was seen to be associated with a poor prognosis."" Normal ovarian epitheli um expresses low-m oderate levels of H ER-2/m 'i/.'" 11HR-2/m u is overexpressed in about 30% of ovari an malignancies and appears to be indicative of a poor prognosis and survival.8 '""" Therefore, HER2 / licit lias the potential to be clinically useful as a marker tor ovarian cancer Lichtenstein and co-workers have show n that IIHR-2/mii overex pressing ovarian tumor cell lines were more resistant to tumor necrosis factor (TNF) or Ivmphokinc'.ii'tivaied killer cell-m ediated lysis than cells that expressed low er lev els o f HER2/ucit. suggesting that overexpression o f this onco gene might impart a biologic advantage to tumor cells bv enhancing their resistance to cytotoxicity.41 1In IIER-2/ii'i/ oncogene"" fits into both the cat egory of oncogene and that of growth factor recep tor since the I ILR-2/wit gene both is a protoonco gene and appears to code for an epidermal growth factor (F.GF) receptor-like m o le c u le .R e c e n tly , a ligand (growth factor?) that binds to this putative growth factor receptor has been described and par tially ch.iracteri/etl Another oncogene that has been seen to be overex pressed in ovarian cancer is i fm>. which also encodes for a growth factor recep tor. in th ista se lor the receptor for M-CSF.1^O vari an cancer cells also have been seen to produce M- CSIV'"' Since the HER-2/ mcwg en e product appears to be a growth factor receptor, antibodies directed to the extracellular portion of this m olecule might be able to down-regulate its growth-prom oting activity. Bast and co-w orkers have found that anti-HER2/ lieu monoclonal antibodies to cells that overexpress IIER-2/ wm can inhibit their growth." "' Antioncogenes Another genetic lesion that has been implicated in the genesis and developm ent of ovarian cancer involves the p53 tumor suppressor gene: the p53 gene has been seen to be overex pressed (and mutat ed) in 30-50% of ovarian cancers.4* * 7 The p53 gene, on chrom osom e l7p, is a tumor suppressor gene, and mutations or loss of p53 have been seen in many human cancers. The product of the p53 gene is a nuclear phosphoprotein that can be expressed by normal cells and plays a role in the regulation of cellular growth and developm ent but that is over expressed when mutated. ,`-4,-4HThe loss of normal p53 function, due to mutation and overexpression or to deletion of the normal p53 gene, is often asso ciated with a malignant phenotype. Mutation of p53 results in the dom inant transformed phenotype since the expression of a mutant form of this tumor suppressor gene leads to the dysfunction of normal p53 by preventing the association of p53 to form a functional DNA-binding, regulatory complex. In recent studies w e have seen that p53 tumor suppressor gene expression can be induced in an ovarian cancer cell line by TNF- together with the induction of cell death by apop tosis.2wTherefore, a mechanism by which TNF induces tumor cell death may involve the op-regulation of tumor cell p53 gene expression. Another tumor suppressor gene that has been suggested to play a role in the developm ent of ovar ian cancer is the retinoblastoma (Rli) locus: a signif icantly high frequency of allelic deletion of RB has been seen in ovarian cancer tissue.42 Growth Factors and C ytokines Growth factors and cytokines have been seen to play important roles in the d ev elo p m en t and grow th o f can cer,122 in clu d in g ovarian ca n cer.4* 4" 4* ''' In fact, it has been proposed that epithelial ovarian cancer may be a cytokinepropelled disease.4"' *1 A m ong the several growth factors that have been  -------^ t Volume 39, Number 4/April 1994 243 examined in ovarian cancer are transforming growth factor-fJ (TGF-p) and EGF and related fac tors, including TGF-a.7-9'53 TGF-0 is a 24-kd glyco protein that can inhibit the growth of many epithe lial cells.** EGF is a 6-kd glycoprotein, with a struc ture similar to that of TGF-a, that can stimulate the proliferation of normal ovarian epithelial cells; both HGF and TGF-a bind to the same receptor.7*9 The response of ovarian cancer cell lines to EGF and TGF-a appears variable.53 In fact. Bast and coworkers have seen that malignant epithelial ovarian cells are less responsive to EGF than are their nonmalignant counterparts.7 TGF-3. which can act as an autocrine growth-inhibitory factor for normal ovarian epithelial cells, may act in a similar fashion for some ovarian cancer cells.7Therefore, loss of the ability to produce active TGF-p or to respond to this factor may have occurred in some ovarian cancers. Various cytokines, including tumor necrosis factor-a (TNF-a), interleukin-1 (IL-1), macrophage/ monocyte colony-stimulating factor (M-CSF) and interleukin 6 (IL-6), may also play important roles in ovarian cancer pathogenesis.7-14'21'34-45'49 Cytokines, including IL-1 and IL-6, have been seen to enhance the proliferation of some ovarian cancer cell lines,7-73 and various cytokines have been seen to be produced by ovarian cancer cells, including M-CSF, CM-CSF, IL-1, TNF-a and IL-6.14'21'34 Many ovarian cancer cells express both M-CSF and fms, the M-CSF receptor.721 34-35-61 M-CSF is a homodimeric, 90-kd glycoprotein. Levels of fms transcripts were seen to correlate strongly with ovarian tumors of high histologic grade and advanced clinical stage and were associated with a poor outcome.34 Also, elevated plasma levels of MCSF were seen in 70-80% of patients with ovarian cancer.7-34-35 It has been postulated that MCSF-stimulated macrophages might produce other cytokines, such as IL-1 and IL-6, that can further stimulate tumor cell growth.34 Therefore, M-CSF could potentially act as both an autocrine-panacrine tumor stimulatory factor and a factor that can mod ify the tumor cell environment, resulting in enhanced tumor cell growth. IL-6 IL-6 is a growl hand-differentiation-inducing fac tor that is potentially relevant to the development and/or progression of ovarian cancer. IL-6 has been shown to act as an autocrine-paracrine growth fac tor for several types of human tumors, including multiple myeloma cells3038and renal cancer cells.31 In our own recent work we found IL-6 to act as an autocrine growth factor for acquired immunodefi ciency syndrome-associated Kaposi's sarcoma.52 Also, IL-6 has been shown to act as a programmed cell death-preventing factor in B-cell tumors.64 Our studies indicate that many epithelial ovarian cancer cells produce IL-6.70 IL-6 is a pleiotropic cytokine with a wide range of biologic effects.30-37 The gene for IL-6 has been cloned and sequenced and the structure of the IL-6 gene described. IL-6 is a 26-kd glycoprotein consist ing of 212 amino acids, based on a sequence deduced from cloned IL-6 cDNA, with some sequence homology with human G-CSF as well as several recently described cytokines, including oncostatin M, ciliary neurotrophic growth factor and leukemia inhibitory factor. The gene for IL-6 in humans is on chromosome 7 and consists of 5 exons and 4 introns. The IL-6 biologic activities that have been described are induction of differentiation of activated B cells, support of plasmacytoma and myeloma growth, induction of acute phase reac tants and stimulation of hepatocytes, nerve growth factor-like activity, induction of cytotoxic T cells and support of hematopoietic differentiation.30 Clearly, IL-6 is a pleiotropic factor, with both growth- and differentiation-inducing properties. IL-6 can be produced by several types of cells, including T lymphocytes, monocyte/macrophages, fibroblasts, certain tumor cells, epithelial cells and endometrial cells. The receptor for IL-6 (IL-6R) has been cloned.74 It is composed of two polypeptides: an 80-kd IL6-binding protein, which is associated on IL-6 bind ing with a dimer composed of a second, 130-kd, signal-transducing molecule. We have detected IL6R (80 kd) expression in ovarian cancer cells.71 Recently the gene for NF-IL6, a nuclear transcrip tion factor that binds to a cytokine-responsive (IL-1) element in the IL-6 regulatory region, was cloned.2 The NF-IL6 protein has some homology with the fos and myc oncogene products and interacts with the NF-IL6-binding motif within the IL-6 gene regula tory region, inducing IL-6 gene expression. Nor mally, NF-IL6 is not expressed, but its expression is induced by exposure of IL-6-produdng cells to var ious inducers of IL-6 gene expression, including LPS and cytokines. We have seen that ovarian cancer cells of epithe lial origin (which constitute 90% of ovarian cancers) \ produce and secrete IL-6.70Several epithelial ovari an cancer cell tines, including SKOV-3, OVCAR-3 1. 4 2-U 77ir lounutl o f Kr/muiui'dri' M rJmne .ml l. \l 3. priHluc* Mibskmli.il am ounts of biologiiulU active II -6." Ov.iri.m cancer .oil lin ts of nonepithi'lial origin (I'A-I and 222) do no! secrete . 11.-r*. although cytoplasmic II -6 w as detected in t'AI a*. well as in all the other epithelial ovarian cancer voil lines tested.'0 Primary ovarian tumor cell cul tures also produce substantial levels of 11.-6/" with primal v isolates oi ovarian tumor cells displaying intracellular ll-r> bv itnm unoporoxidase stain ing. I he II. o prinluivd by ovarian cancer cells was examined bv rndioimimmoprecipitation and found to correspond in molecular weight to mono- to / lvm phoid-derivcd 1 1-6'" Also, II.-6 mKNA expression indicated that most epithelial ovarian tumor ell lines display delectable levels of II.-6 gene expression. " Several cytokines (inlerferon-y, II IS.'-Y H I amt IN l) were seen to up-regulate IL-6 production bv ovarian tumor cell lines.7" Primary i niliir.*s> ol normal human ovarian epithelium also were seen to produce II.6.'*1 Ihese results indicate that ovarian cancer cells-- both established ovarian i am or cell lines, primary tumor isolates and nor mal ovarian epithelial ells-- proiluce biological ly a live 11.-6 that is iiulisliuguish.iblc from the II -o prodiueil bv peripheral blood m ononuclear ells Ovarian earner cells also express the IL-6 recep tor. In a nvent study wo delected expression of the NU-kd Mibunil r the II.-6 receptor in most lvarian cancer ell lines teste! bv Northern blot analysis.71 Naturally, it II.-6 is an aiilcrme positive >r negalive growth (actor, ovarian cancer cells nl only should be proilucingiletectnblc |uantitiesot s*cret ed 11-6 anil II.-6 mKNA but also should express detectable ijuantitiles ol IL-6 roieptiir. Therefore, the observation that ovarian cancer cells express the II 6 reieptor allow s the possibility that IL-6 might play a role as a grow th /d iiloren iiation /viab ility lador tor Ihese 'ells. I le\ ati-il li'vels ol IL-6 wen* ili,ti,cli,l in asilic flui1-"and serum 11Irom wonu*n with ovarian caner. Ascitic llniit isolated Irom patients with ovari an cancer had very high levels >1 IL-6 (3.2-42 ng/m l.). while those from controls w ithout cancer were iiunh hover (Lb 2.6 n g /m l.) .7" Elevated serum II -6 levels were s o i l to correlate with the pri'seiue >1 more exU-nsive disease: elevated levels ( 0 .2 U /m l ) 1 IL-6 were seen in 16/21 (76%) of ovarian an*r patients with macros*pic tisease hut m only I3'<{ ol patii'iits with microscopic d is ease or in 17'.; ol healthy control d o n o rs.1-1 The mean serum 11.-6 Hentration in the group of ovarian caiuvr patients with macroscopic disease (n = 21) was 0.26 i 0.04 (SEM); serum levels ol IL-6 in healthy adult donors were 0.12 t 0.03 U /m L IL-6 (m aximum < 0.020).11 Serum IL-b reached very ele vated levels (> l U /m l.) in som e patients. Recently w e examined the in nilhbiologic effects of IL-6 by correlating levels of IL-6, acute phase pro teins (C -reactive protein |( Rl'|, haptoglobin and a2macroglobulin) and IgM /A /C i in ovarian cancer patients' ascitic fluid or serum (unpublisheil obser vations). Significantly elevated levels of IL-6, CKI\ haptoglobin and IgA and decreased levels f IgM were siv n in ovarian cancer ascites (n - 36) when compared to peritonea) fluids from a control group (donors with nonmalignant gynecologic condi tions, n = 20). Again, in this study elevated serum IL-6 levels (6.5 r 5.2 vs. 0.2 r 0.02 U /m L ) were seen in ovarian cancer (n - 7) as w as increase! serum C KP w hen compared to controls (n 7) A scites CRI* Ciirrelated with IL-6 in both ovarian cancer (P < .01) and controls (P <: .001). These results su g gest that the increased levels of II.-6 m vivo in patients with variait cancer are associated with increased lev4s of acute phase proteins and IgA. This is particularly interesting since IL-6 can act both as a 0-cell stim ulating factor, inducing the d if ferentiation of activated B cells to Ig-secreting cells, and as a potent indun-r for the production of acutephase rptants.'"47,,7-,'M Another recent observation suggests that IL-6 can play a role as a grow th factor for ovarian cancer cells: activate! m onocytes produce factors that enhance ovarian tumor cell growth Working with Robert Bast and colleagues at Duke University, w e have seen that culture supernatants from activated human monocytes, which support ovarian tumor grow th, contained significant am ounts if IL-6.7-1 Furthermore, inti--IL 6 serum inhibited part of this m onocyte-produced growth-supporting activity. Therefore, IL-6 has tlu* potential to act as a paracrine grow th tactor fr ovarian cancer cells. Also, wo have seen that ovarian cancer cells cul tured in serum -free com b lion s, w hich result in greatly Kvreased e n d ogen ou s IL-6 production, reijuire exogen ous II.-6 for optim al proliferation (unpublished observation). W hile it is clear that IL-6 can act as an a u to crin e/p .iracriiie grow th factor for hum an tum ors, the I'ompletc role of IL-6 in the host response to cancer or as a tumor growth-prom oting or -su p p ressin g factor is not clear.4**Certainly, IL-6 has the potential to modulate antitumor immune Volume 39, Number 4/April 1994 245 responses and may thereby act as an inhibitory fac tor for tumor cell growth. Recently we examined the possibility that IL6 acts as an autocrine growth factor for ovarian can cer cells by inhibiting IL-6 gene expression by expo sure to IL-6 antisense oligonucleotides. This was seen to result in greatly decreased cellular prolifer ation: exposure of ovarian cancer cell lines (CAOV3, OVCAR-3 and OC-436) to various concentrations of a single-stranded antisense 11.-6 oligodeoxynucleotide, specific to a sequence in the second coding exon of the IL-6 gene, resulted in decreased IL-6 production and a >80-85% inhibition of cellular proliferation.71 However, the addition of exoge nous IL-6 failed to restore the proliferation of the antisense-treated cells. Also, antibodies to IL-6 did not consistently inhibit cell growth, nor did rIL-6 enhance growth at low cell concentrations. These results suggest that exogenous IL-6 does not direct ly induce the proliferation of ovarian cancer cells, although endogenous IL-6 production is needed for optimal cell growth. Since the majority of epithelial ovarian cancers produce IL-6, the direct specific inhibition of IL-6 gene expression may be of poten tial therapeutic value. Preliminary studies indicate that various tu mor cells can be protected from natural killer (NK) cell-mediated killing by IL-6 (unpublished observations).*-1 Cells (CAOV-3 ovarian cancer cells) treated with dexamethasone, which results in greatly decreased IL-6 production,62 showed greatly increased sensitivity to NK cell-mediated lysis. Addition of exogenous recombinant IL-6 to dexamethasone-treated cells restored resistance to NK lysis. These preliminary results suggest that IL-6 (endogenously produced or exogenous) can protect tumor cells from NK-mediated kill ing II. 10 Epithelial malignancies of the ovary remain con fined to the peritoneal cavity even in advanced stages of the disease. It has been suggested that the growth of ovarian cancer intraperitoneally could be related to the local deficiency of antitumor immune effector mechanisms, allowing tumor growth with in the peritoneal cavity.12The presence of immuno suppressive factors in the ascites from ovarian epithelial tumor-bearing patients has been de scribed previously.3,32 In recent studies we observed that ascitic fluid from patients with ovari an cancer contains significantly elevated levels of IL-10,28 a cytokine with various immunosuppres sive activities. IL-10 was originally identified as cytokine syn thesis inhibitory factor because of its activity as an inhibitor of cytokine production.25The gene for IL10 has been cloned.54 This 35- to 40-kd cytokine is produced by the "type 2" CD4-positive helper T cell clones (TH2) and inhibits cytokine production of the "type 1" subset (TH1) in murine systems TH1 and TH2 are two subpopulations of T helper cells, initially defined in mice, that control the nature of an immune response by secreting characteristic and mutually antagonistic sets of cytokines:55 TH1 clones specifically produce IL-2 and IFN-y, whereas TH2 clones produce IL-4, -5, -6 and -10. IL-10 inhibits the production of IL-2 and IFN-y by TH1 cells.27 While T cells corresponding strictly to TH1 and TH2 subsets have not been found in humans, increasing evidence points to the existence of TH1and TH2-type responses in humans; a similar dichotomy between TH1- and TH2-type responses has recently been reported for the human.23,60 Human IL-10 also inhibits the synthesis of IFN-y and other cytokines by human peripheral blood mononuclear cells.76 Furthermore, IL-10 is exceed ingly potent at suppressing the release of such cytokines as IL-1, IL-6, IL-8 and TNF-a by activated monocytes.16,24,26 IL-10 further down-regulates class II major histocompatibility complex molecule expression on monocytes, resulting in a strong reduction in antigen-presenting capacity of these cell9.24 Together, these observations support the concept that IL-10 plays an important role not only in the regulation of T cell responses but also as an antiinflammatory mediator. IL-10 is produced by T cells, Bcells and monocytes.76 Among T cells, CD4positive cells appear to be the main producers of IL10, whereas CD8-positive cells appear to be poor producers of this cytokine. Very recently we measured levels of various cytokines, in collaboration with John Abrams, DNAX, Inc., in ascites obtained from women with ovarian cancer and found that levels of various cytokines, measured by enzyme-linked immuno sorbent assay, were significantly elevated when compared to levels in peritoneal fluid from women who did not have ovarian cancer. In particular, lev els of IL-10 and, as expected, IL6 were particularly elevated in ovarian cancer ascites.28 Our prelimix nary results indicate that ovarian cancer cells do not produce IL-10; ovarian cancer cell lines and prima ry tumor cells cultured in vitro did not produce I . tu tmii 246 77ir journal o f Reproductive Medicine detectable levels of II - 10, w hile several of these lines did produce II.-tv F.levnted levels of IL-10 were seen in nearly all ascites samples from w om en with ovarian cancer. Because peritoneal im m unosuppression is characteristically seen in this disease, we are currently examining the role of IL-10 as a poten tial peritoneal im m unosuppressive factor in ovari an cancer. I\ ttoin'iiJCytokine* in P i'iiihin Cumer Since the develop m en t o! ovarian cancer may involve cytokine dysrgulation, w e recently deter mined. in collaboration with John Abrams, the lev els ol various cytokines in ascites collected Irom wom en with ovarian cancer. As m entioned above, high levels ol ascites IL-h and -10 are seen in the great majority of ovarian cancer cases. Of those cases, S0-h0'i. were seen to also contain significant ascites levels of IL 2, TNI n, IFN->, C-CSF and CMCSF.7- Levels of varii>us cytokines, including 11.-6, IL-10, TNF-u. C-CSF and CM-CSF, were signifi cantly elevated as compared to levels of these sam e cytokines seen in peritoneal flu ids from normal women. A .similar pattern was seen in sera from women with ovarian cancer, with II.-b and IL-10 the serum cytokines detected most frequently. A signif icant correlation wa> seen between Il.-f> (and IL-10) and acute phase proteins and im m unoglobulins as well as the immunosuppressive activity of as citic lluids and tumor histologic stage. Studies to determine whether a specific cvlokine/tum or marker/acutc phase reactant pattern could be u se ful in monitoring the progress ot patients w ith ovar ian cancer are under way. Certainly it is possible that certain cytokine pat terns. in particular high levels of IL-10, could result in a peritoneal environment characterized by immune un responsiveness and promotion of tumor growth. Conclusion The dysrgulation of various biologic and molecu lar lactors may play important roles in the d ev elo p ment and progression of ovarian cancer. These include oncogene overexpression, mutation or loss of antioncogenes, growth factor and cytokine stim ulation, and modification o f the host environment and host antitumor immune response. Understand ing the specific mechanisms involved in ovarian cancer development and growth will allow oppor tunities tor the design of effective antitumor treat ments. Acknowledgments The authors express their gratitude tor the laboratory work and assistance of C athy C asey, Walter Cotlieb, Michael Johnson, Klara Kaldi, Reba Knox, Robert Stenson and Joanna Watson and for the gen erous provision of reagents by John Abrams and Tadamitsu Kishimoto. References I. A.iron**n SA. Crowlli factors .ml lancer. Science 254:1146, 1441 2 Akira S, Kstiiki II, Sugil.i I, el ,il. A nuclear factor for If.-6 expression (Nl -II.o) is .1 mvmlvr of a C/lilll* family. 11:lW. IW) 3. Badger A, Oh S, M oollen F: Differential eflei'ls of an immunosuppressive fraction from asiites fluid of patients with ovarian cancer on spontaneous and antibxnJy depen dent cytotoxicity. Cancer Kes 41:1133, I<#81 4 Baker VV, Borst Ml'. Dixon D, ef at: C'-niyc amplification in ovarian cancer. Gynecol Oncol 38:340, 1440 5. Uarletia C. I a/./_aro I). Prosperi Porta R. el nl. C-MYBactiva tion and the pathogenesis of ovarian canter Fair ) Gynaecol Oncol 13.S3. 1442 6. Bast KC. Xu KJ, Rodrigue/. GG, et al: Inhibition tif breast and ovarian tumour cell growth by antibodies and imnuinotoxins reactive with distinct epitopes on the extracellular domain of IIHR-2/neu U-erbB2) In Ovarian Cancer 2: Biolo gy. Diagnosis. and M.inagemeiit. Edited by FSharp, WP Mason, VVCrcasman l.ondon. Chapman & Hall. 1*442, pp h7 7 1 / 7 Bast RCf Xu FJ. Rodrigue/ GC. et at: Factors regulating the growth of normal and malignant ovarian epithelium. In Ovarian Cancer 2 Biology. Diagnosis, ami Management. Fdited by FSharp, WP Mason, VV( reasman I ondon. Chap man & Itall. 1442. pp 61 -6t> 8 Berchuck A. Kamel A, Whitaker R, el al: Overexpression of HER-2/neii is associated ivith poor survival in .idvanred epithelial ovarian canter. Cam er Res 50:4087. 1440 ** Berclmck A. Kohler Ml'. Bast RC Jr: Oncogenes in ovarian cancer. Ilemalol Oncol Clin North Am 6-.H13, 1442 10 Berchuck A. Marks |R. Hast RC )r Expression of Ihe epider mal growth factor receptor, I IF-K-2/neu and p53 in tivarian cancer In Ocarian Cancer 2 Biology. Diagnosis, and Man agement. Edited by F Sharp. WP Mason, VVCreasman. Lon don. Chapman & Ifall. 1442. pp 53-54 11. Berek JS. Epithelial ovarian cancer In Practical GynecologicO ncology Edited by |S Berek. NF Hacker B.illimore. Williams & Wilkins. 1484, pp 327-3M 12. Berek JS, Bast KC', Lichtenstein A. et al: Lymphocyte cytotoxicitv in the peritt>nealcavitv and bl-xkl of patients with ovar ian cancer. Obslet Gynecol r>4 7tw. 1484 13 B erek JS, C h o n g L . K aldi K, et nF S e ru m II.-6 lev e ls co rrela te w ith disease Status in epithelial ovarian cancer patients Am | O b stet <iynecol 164 1038, 1441 14. Berek JS, VV.itson JM, Martimv-Ma/a (>: Role of interleukin6 in ovarian cancer In Ovarian I ancer 2. Biology, Diagnosis, Volum e 39, Number 4/April 1994 247 and Management. Edited by F Sharp. WP Mason, W Creasnian. London, Chapman 6r Hall, 1992. pp 101--113 15. Bishop JM: Molecular themes in oncogenesis. Cell 64:235, 1991 16. Bogdan (.'. V ojovotz Y. Nathan C: Macrophage deactivation by IL 10. | Exp Med 174:1549. 1991 17. Boltz EM, Kefford KF, Leary JA.etal: Amplification o f c-raaKi oncogene in human ovarian tumours Int | Cancer 43:428, 1989 18. Bookman MA, Bast RC. Jr: Immunobiology and immunotherapy of ovarian cancer. Semin Oncol 18:270,1991 19. Bcxikm.m MA. Bcrek JS: Biologic and immunologic therapy of ovarian cancer. Hematol Oncol Clin North Am 6:94!, 1W2 2. Buyalos Rl', Watson JM, Martinez-Maza O: Detection of interleukin-6 in human follicular fluid. Fertil Steril 57:1230, 1992 21. Carson LF, Moradl MM, U B-Y, et al: Characterization of cytokines produced by ovarian cancer cells, in Ovarian Can cer 2: Biology, Diagnosis, and Management. Edited by F Sharp, WP Mason, W Creasman. London, Chapman Ac Hall, 1992. pp 93-99 22. Cross M, Dexter TM: Growth factors in development, trans formation, and tumorigenesis. Cell 64:271, 1991 23. Del I'rete CF, De Carli M, Ricci M, et al: Helper activity for immunoglobulin synthesis of T helper type I (Thl) and Th2 human T cell clones: The help of Thl clones is limited by their cytolytic capacity. J Exp Med 174:809, 1991 24. de Waal, Malclyt R, Abrams J, et al: IL-10 inhibits cytokine synthesis by human monocytes: An autoregulalory role of IL-10 produced by monocytes. J Exp Med 174:1209,1991 25. Fiorentino l)F, Bond MW, Mosmann TR: Two types of mouse helper T cell. IV. Th2 clones secrete a factor that inhibits cytokine production by Thl clones. J Exp Med 170*2081. 1989 26. Fiorentino DF. Ziotnik A, Mosmann TK et al: IL-10 Inhibits cytokine production by activated macrophages. J Immunol 147:3815, 1991 27. Fiorentino DF, Ziotnik A,, Vieira P, et al: IL-10 acts on the antigen-presenting cell to inhibit cylokine production by Thl cells. J Immunol 146:3444, 1991 28. Gotlieb WH. Abrams JS, Watson JM, et al: Presence of IL-10 in the ascites of patients with ovarian and other intraabdonv inal cancers. Cytokine (in press) 29 Gotlieb WH, Watson JM, Rezai A, et al: Upregulation of p53 tumor suppressor gene expression by tumor necrosis factoralpha: Relation In the production of apoptosis in an ovarian cancer cell line (submitted for publication) 30. Hirano T. Akira S, Kishimoto T. Biological and cllnlcsl aspects of interleukin 6. Immunol Today 11:443,1990 31. Hollslein M, Sidransky D, Vogelslein B. el si: p53 Mutations in human cancers. Science 253:49, 1991 32. Johnson M l , Scnsintaffar JL, Watson JM. et al: Detection and preliminary characterization of an immune suppressive fac tor in ascites from women with ovarian cancer (abstr). Proc of Meeting of the Society for Gynecologic Investigation, San Antonio. |99| 33. Johnson MT, Watson JM. Berek JS, et al: lnterleukin-6 induces resistance of ovarian tumors to natural killer cellmediated cytotoxicity (submitted for publication) 34. Kacinsky BM:CSF-1 and its receptor on ovarian and other gynaecological neoplasms. In Ovarian Cancer 2: Biology, Diagnosis, and Management. Edited by F Sharp, WP Mason, W Creasman. London, Chapman & Hall, 1992, pp 115-125 35. Kacinsky BM, Carter D. Mittal K. et al: Ovarian adenocarci nomas express fms-complementary transcripts and fms anti gen, often with co-expression of CSF-1. Am J Pathol 137:135. 1990 36. Kawano M, Hirano T, Matsuda T, et al: Autocrine generation and requirement of BSF-2/1L-6 for human multiple myelo mas Nature 33283,1988 37. Kishimoto T, Hirano T: Molecular regulation of B lympho cyte response. Ann Rev Immunol 6:485, 1988 38. Klein B, Zhang X-G, Jourdan M, et al: Paracrine rather than autocrine regulation of myeloma-cell growth and differenti ation by interleukin-6. Blood 73:517,1989 39. Kokal Y, Do bash i K, Weiner DB et ah Phosphorylation process induced by epidermal growth factor alters the onco genic and cellular neu (NGL) gene products. PNAS 85:5389, 1988 40. Kommoss F, Bauknecht T, Birmelin G, et al: Oncogene and growth factor expression in ovarian cancer. Acta Obstet Gynecol Scand (suppl) 155:19. 1992 41. Levine AJ. Momand J, Finlay CA: The p53 tumour suppres sor gene. Nature 351:453, 1991 42. LI SB, Schwartz PE. Lee WH, et ah A llele loss at the retinoblastoma locus in human ovarian cancer. J Natl Cancer Inst 83:637, 1991 43. Lichtenstein A, Berenson J, Gera JF, et al: Resistance of human ovarian cancer cells to tumor necrosis factor and lymphokine-activated killer cells: Correlation with expression of HER-2/neu oncogenes. Cancer Rea 50:7364, 1990 44. Lidor YJ, Xu FJ, Martinez-Maza O, et al: Constitutive pro duction of macrophage colony stimulating factor and IL-6 by human ovarian surface epithelial cells. Am J Obstet Gynecol (in press) 45. Malik S, Balkwill F: Epithelial ovarian cancer: A cytokine propelled disease? Br J Cancer 64:617,1991 46. Malik STA, Naylor MS, Balkwill FR: Cytokines and ovarian cancer. J Ovarian Cancer 2 Biology, Diagnosis, and Man agement. Edited by F Sharp, WP Mason, W Creasman. Lon don, Chapman Ac Hall, 1992, pp 87-92 47. Marinkovic S, Jahrels GP, Wong GG, et ah IL-6 modulates the synthesis of s specific set of acute phase plasma proteins in vivo. J Immunol 142:808, >989 48. Marks JR, Davldoff AM, Kern* BJ. et al: O vevexpmsion and mutation of p53 in epithelial ovarian cancer. Cancer Res 51:2979, 1991 49. Martinez-Maza O, Berek JS: Interleukin 6 and cancer thera py. In Vivo 5 583,1991 50. Mazars R, Pujol P, Maudelonde T, et al: p53 Mutations in ovarian cancer A late event? Oncogene 6:1685, 1991 i l . Mlki S, Iwano M, Miki Y, et al: lnterieukin-6 (IL-6) functions as an in vitro autocrine growth factor in renal cell varc'mo- .. A. 248 Thr /pum al of Rcftroduilire Medicine m.iv 11 Hs I eli* 2*Uon?. i w Nili,- W K e /a i AK. s .il.i/.ir-( i o n / a l e / 11-. e t .il1 A llS K.ij *i *v| v v.iii <>m.i J i i i w J i i*l|v p ro d u c e lim i re s p o n d lo Mila'iUiikin I ' l 'io N .ill A ..n i V i l'S A 87 4<lhH. 19*40 Miil.v (.11. I la s h iu in lo S. I lu ru -a ii |, i'i .il. K i'^ u la h n n ot R itm ili o l lu m i.m o v a ita n >.iih v i vlls I h O v a ria n ( '.iiu v i 2: Hm lo^v. |) i . i p u M \ .uni M an a g em en t I .l i u ti lv I- s h a r p . V\ | ' M .ison. \Y i iiMsin.ii) I n in i.m ( Jtapm .ni I lo11 1*^2, i r i-7 'I M m .i.' kVV V ie u a I*. I m r e m in o DI-, f i al I io m o lo ^ y tl tv io k u n - v\ niln-sis in h ib iio rv la to t (Il I(n io lite I |*s(-in H. m virus vile IK, RII S ifinv 248:1230. l***<0 ^ M ovin.inn IU. M ooie KW Un* r o lf o t II. IH in cross reputalio n ot l l l l .im i 1112 r.`vjH.iis*'N Im iin n io l lodai* I2.A 49. i+n V< M o li.. II. lini A. Ni. Iis I.. ci al. I \iti.tu ot m te rlfu k m 6 ^cinex pi essim i m m i o M i ^ f s i s .i ro lf l.ir ib is cy lo k u u - in .irtelo^ Iifs is l'rts N .ill A ,.n l S .I IN A M7 3092. 1990 17 t'k .n iio io A N m ifsliiin.i V V>kovai ita S. cl .il: I revjuciil alli'ln Inssfs ,inil iiiiii.ilions ol lite |^ 1 m 'iic in h u m an of.iri.m t .lin c i < .n itf i Kos s | ''IT I. I**9| JVIcs t K.n iis sS . k o sk i KA, e t .il Isolatim i ol ll)c lie ti'U liR 2 sim u il.iio rv li^.uul: A 44 k.l ^ h t o p ro li-m lim i n u lin c s d iiI. ifiiti.ilioii ol iii.niiin.irv Unnol vclls t eli t*9 20S 1*492 1> l 'i u i M s r p n l f i i i i o l o ^ of o x .iri.in .lin e r In O v .in .in M alignane ics Dm filo s i n .uni I h e i a |v u li c A vlvam es. Evlilcd In M *l'i\ci I Im b u rri. I lu m lo il I iv in ^ s to n c . IUK7. pp I IO mi Komi.lam in i s Ifu u i.in l l l l .u n i f l l ? s u h s e ts : l> o u b l in* iin u c liuiniiiiol lo tl.iv 12 2`V*. I**4I ni K.tiii.ikfisluntn S. \ u H . b ra n d i S |. -I .il ( o iislilu liv c prtlm lio iio l m .u ( o p t ia t e i o Io iiv siiin u l.ilin ^ l.ivtor by lum m n o v .iii.in .u n i brc.ist t . u n i r t e l i lines. 1 1 liti Io v est 8V 92I, 1989 2 .iv V l.iU er *>H. N inlli.in.ini l . ! .il ^ul.ilioii <*l expresSion t>| in lfrli'iik iii n. M olci u l.ir.u n l i lu iit.il s lm lif s A n n N I A.ni S i 5*7 153, 1*489 ni KtsIfnlnir^C |. km-lin.i IA, N.ip M. fi al. Ininiimulustoclieinit .il tlfltv lio n ol ih- ras o r n i ^ f i i f p e n d in i p2 I in .n iia m c il m a n a li i .u n c r 'l.n k ol torrflaliou w illi iluiu.il ouUonn* A i. li l'.ilb ..11a b M ei! 112.1*1. 98 <4 N il-o u rin I A l l a w l e i Re. N u p p ie s s n n t o t p ro g r a m n i n i ilcalli and (i, arrest in ll-cell hyhrtdoma* bv inierleukm-h is noi accompanied by altered expression of immediate early rfs|*oiise m*nes. ) ( ell Physiol 145:*'*4. 1990 t>5 SilverlHT F.. Borin ( C. Scpiires IS. Cainef statistics IWI ( A Cancer | Clin 40.9. |9 **i lift Sl.itilon l)|. (anlolplim VY. Ituies I A. el al' Slinlies ol 111! 2/n cii prolo-oino^eue in liiiiiian breast ami o fan.ui earner. Science 244:707. |9H*j h7. Stensoii K. Watson JM, kmvetuowski PA. el al: ( .encration, ehar.u lfri/alion and growth in nude nine ol lvo ovarian luinor veil lines (abslr). Piov of Meeting of Svn'iely lor G yne cologic Investigation, San Aniomo. 1991 rH TosatoC, Seamon KH. ( '.oldman NIX el al: Idv-nlifiealion of a nion.x yle-dfriw-d limn.in cell grim III laclor as mlv-rlerorib, IHSI -2. II -M Scieinv 23*4:502. 1*^ 6*4. Wakeliv-ld I M, Sjhirn Mil Suppression of carcinogenesis: A role lor T(iF-p.iiul relalvil nioUvules in prevention ol earner. tn Tumor Suppressor Genes tdiled bv C Klein. New York. Marcel IVkker. 1*4*40. pp 217-243 70. Watson I. Sensiiualfar |l., v-rekJS. el ai Constitutive produv lion of interleukin by ovarian c.mwr cell lines and by primary ovarian tumor vultures. Cancer Kv* ^0:h**5*i, 1*490 71. Watsvm )M. Derek JS. Marline/-Ma/a (H ir n n lli inhibilion ol ovarian cancer cells iuvlvned hv antisense II -n oligonu cleotides CyneCol O iuol (in press) 72. Walson JM. ( iollivb WM. Abrams JII. el al- Immune analysis ol ascitic fluid imm patients with ovarian uincer: Relation ship between cvlokmes. acute phase pion-ins. minuinoglobulins. immunosuppression, and tumor vlassiticalion (sub miiievploi publication) 71 Wu 5 . ^ixtabau^h K. Walson |M. el al: Stimulation of ovari an tui^or cell proliferation with monocyte prixlucls includ ing II.-1-alpha. II.-hand tumoi nectosis lacior-alpha. Am J O i 's l e l ( . y i u v o l l h h * ^ 7 . 1*492 74. Yamasaki k. Taga T, Ilirala Y. el al: Cloning anvi expression of human inlerleukm-h lHSI -2/IFN b,) receptor Svienco 241:825, I9KK 75 /I n hi I>|. ( a m /a if/ CadavivI N, Ahuyi 11. el al. A unu|ue pat tern ol prolo-oncoj*enc .ibnorm.ilities 111 0 v.1r1.1n avIv-niKareinomas Cancer h2:1573. 198 7n. /loiiuk A. Misire kW. Interleukin 10 ( viokine 3 .4r>h. 1991 * Insulinlike Growth Factors Their Regulation of Glucose and Am ino Acid Transport in Placental Trophoblasts Isolated from First-Trimester Chorionic Villi Douglas A. Kniss, Ph.D. Phillip J. Shubert, M.D. Peter D. Zimmerman, B.S. Mark B. Landon, M.D. Steven G. Gabbe, M.D. cells were maintained in serum-supplemented medium until confluent, at which time they uwe shifted to serumfree medium for one day. Experiments were initiated by transferring the cells to glucose-free assay buffer and incubating them with IGF-I, IGF-II or insulin. Glucose uptake was measured by the transport of2-deoxy-D-t 1.23Hglucose (2I3HIDG) in the presence or absence of cytochalasin B, which has been shoum to competitively inhibit glucose uptake. IGF-I. IGF-II and insulin each enhanced 2I3HIDG transport in a dose dqvndent fash ion. Amino acid transport was measured by incubation of the cells with IGF-I for 60 minutes, followed by a 5minute challenge with a-lmethyl-3Hlaminoisobutyric acid. IGF-I caused a dose-dependent increase in uptake of the amino acid analog. Radioreceptor assays using l ,25llinsulinlike growth factor I (I12511IGF-1) demon strated that the trophoblast-derii>ed cells contained highaffinity, saturable receptors for IGF-I that also bound IGF-II. Insulin bound to this binding site with very low affinity These data indicate that in contrast to previous reports that glucose and amino acid transport in the pla centa are unregulateil processes, IGFs may exert hxal modulation of acute metabolic actions in trophoblastderived cells via IGF-I receptors. (J Reprod Med 39:249-256,1994) Keywords: growth substances, chorionic villi, tro phoblast. The transjwrt ofglucose and amino acids from the mater nal to fetid circulation through the placenta is critical to the delivery of fuel for normal fetal growth and develop ment l ittle information indicates that transplacental glucose or amino acid transport is influenced by hor mones or folypeptide growth factors. We developed a continuous cell line of cytotrophoblastlike cells derived fromfirst-trimester human chorionic villi as a model sys tem to stiuly the regulation of glucose and amino acid transport by insulinlike growth factors (IGFs). Using immunocy/ochcmical and biochemical criteria, the cells were shown to manifest a troplioblastlikc f>henotype. The From Ilu* Division of Material-Fetal Mi-divine. Department of Obstetrics and Gynecology. Ohio State University College of Medicine. Columbus, Ohio. This work was sup|H>rled in part by grants from the Diabetes Research and Education Foundation, Inc. the American Dia betes Association Ohio Affiliate and the March of Dime* Birth Defects Foundation- Address cones|Hndence to: Douglas A. Kniss, Ph.D., Depart ment of Obstetrics and Gynecology, Ohio Stale University Col lege of Medteiin'. 1654 Upham Drive, Means Hall, Columbua, O il 412111 I22H. Introduction The human placenta must meet the metabolic demands of the developing fetus through the deliv ery of important fuels, such as glucose, amino acids and fatty acids. Glucose is the major carbohydrate source for oxidative metabolism in the fetus and is transferred across the placenta via an energyindependent, carrier-mediated process that appears to be unregulated and determined solely by mater nal blood glucose levels ' 1Derangements in mater nal carbohydrate utilization (i.e., diabetes mellitus) that result in maternal hyperglycemia are known to have profound effects on fetal development, well being, or both, including structural malformations and fetal macrosomia.4 According to the Pedersen hypothesis, fetal macrosomia in the diabetic preg nancy derives from the delivery of excess glucose to the fetus by the placenta and a compensatory increase in insulin release by the fetal pancreas.9 Insulin serves as a growth factor to drive cellular hyperplasia and hypertrophy, resulting in an Increased risk of diabetic fetopathy. Controversy exists, however, as to whether 0024-7758/^1 /l04-0244>/$l 50/0 2Tin- lournal of Reproductive Medicine Inc H t 11*>) M. .* sax 0 Mie M.iiiinll.in I*ION I h i. IW I Risk factors for ovarian cancer: a case-control study M H o o ili'. V Hi t .i l ' & P S m iilr 'hpi*irnnc4T*tit-ril--U*`mi4M4ti)i-Lju i.*uul Ii.ajfiujLLpHUmuiloj' v l ini Dtjhinnir/U ni F.pitlrmiolt'fH d l'opuLition Siu nu \ Lotnlon .s, / //u'oiii ,./ In'pinil UcJhi/w. kcpprl Slrci't Him*rr .V/rwf/. I.ontlttn H'CIF. 7UT. J'K. miti 'itrtpt'ntiWtinccr K,\,,ut h f-unti l.piJi fUt/c < ( finn al Iritil\ l uti Rthlt lif*- hitirnuirt . Oxfortl OX* f*HF. l'k . N uniiniri A liospil.ilhascij case . n t M i n i s >1 n v .in jn i.inecr wav cntkliiticil in l.om lon and O xlotd hciuceii (K lniict l*#7K and l ch n i.if) I`IRI M enstrual character isiics. reproductive an d contraceptive history ami ItiMurs >>l exposure to various environm ental factors were com pared between J35 w om en with his tologically diagnosed epithelial ovarian cancer and 451 controls High gravidity. hysterectom y, female sterilisa tion ami oral contraceptive use were associated with a reduced risk o f ov aria n cancer. Infertility and late age at m enopause were associated with an increase in risk While these factors were related, they were each found to be uidcpciiJcnily associated with ovarian cancer risk alter adjusting lor the cllccl o f the other factors. Whilc restili* Imiti icccni case control sim hcs Nave c o n sistenti) shown ih.n tiuiliip.ini) and orai contracepttve use are associateti wnli a rcducctl risk o f ovaii.m cancer. th .issi>si.iitoti ol th e.ilicer wnh odici rcprodiiciixc. liormon.il aiti! telateci la ciois sudi as age al menopause, hisiory ol' liystcectoinv oi use ol' oestrogen rcpiaccinctii therapy is Ics* clc.tr We h .i\c contlncicil a hospil.il-hasetl case control stud\ in I ond,ut .ind O xloid w hkh was desiglieli lo inves tii:.ite ilu- tndcpeiiileni cotitrthiiiions ol rcproductivc history .uiil to n ii.icep m e use io ov.trtan cancer risk In parttcuhir. il was pl.iiincil io adem p in segregate ini th elicci un risk ol llllcltlllts 11a>111 dl.ll ol volutila!) lllllll.llMMI ol l.ltntl) S|/C. lite .issoti.iitoit K iweeti ovari.ut cancer and tnIter po**ible acliologik.il .igeilis was al.so ctatiiiiied S u b je c ts and n u t hi ids Between October lv7s and Icbniarv 19*3 live interviewer* identified and i|ticstioned women with a diagnosis ol'ovarian cancer and women selected as controls al H hospitals in London ami two m Oxiord A standard questionnaire was used to obtain tnioiin.ilton on rcpriHlucmc and menstrual history and on exposure to various substances such as exogenous ivsirogcn*. cigarette* and talc A month by month record was made o f the specific contraceptive m ethods used bx each woman between the age* of 16 and 45 years, or. if under 4 ' years, up to the tune o f diagnosis teases I or inter view icoiK iols) I lie m ethods were elassilied .is sheaths, diaphragms, iiiifanictinc devices, oral contraceptives or 'other methods' (spcnmcklcs rhythm and coitus mierriipius). Women who reported using a contraceptive diaphragm were asked if they had stored n in talc A lso recorded were months duiing which a woman was not using contraception due to sexual abstinence, pregnancy, menopause or because she or her partner haJ been sterilised. I'hc other months when a woman reported using no method o f contraception although sexually active have been classified as m onths o f unprotected intercourse I'hc total duration o f use o f each contraceptive method, o f any contraceptive m ethod, o f unp rotected intercourse and o f pregnancy were com puted tor each woman flic study w.is lonlincd to women aged less than 65 years whose diagnosis o f ovarian cancer had been made within iwo years o f interview A total o f 2M) eases were interviewed and palhi'logical specimens were histologically elassilied by Pro fessor C Hudson and Dr M Curling from St Bartholomews Hospital A total o f 235 women with epithelial ovarian cancer were included in ihe analyses. For these wom en, the tumour type was described as serous in 101 (43% ) cases, mucinous in 3H i l 5 "o) cases, endom etrioid in 52 (22% ) cases ( oricsponJence M IWih Received 3 Ichru.uy Wrv. and m icwx'J fonn v Mjy 11HV and clear cell in 12 (5% ) cases. Mixed and undifferentiated types o f epithelial tumours accounted for the remaining 32 (14% ) cases. Fxcludcd from the analyses were nine women with a non-epilhcliul ovarian neoplasm . 11 with a primary tumour in an unknown site outside the ovary. 21 with a primary tumour in an unknown site although one consistent with an ovarian origin, one with a benign tumour and three for whom pathology material could not be obtained. I or each case it was planned to select two age-matched controls from women being treated m the same hospital W om en with bilateral oophorectom y were excluded from the c o n lio l group as were women adm itted with conditions that have been related lo reproductive history or oral contracep tive use tall circulatory and gynaecological diseases, gallblad der and thyroid diseases, rheumatoid arthritis, malignant disease o f the breast, ulcrus and bladder, and melanoma) It proved logistical!) im possible to select iw o uge-maichcd controls for each case from the sam e hospital and it was decided merely to ensure (hat the age distribution o f the controls was approximately the same as that o f the eases. For 63 eases recruited from a London hospital where only cancer patient* arc treated, control* were selected from other London hospitals. For these reasons, the data were analysed using an unmatched approach with adjustments being made to relative risk estimates for age and socio-econom ic status. A total o f 451 controls have been included in the analyses. Ihe admission diagnose* for these patients were gastrointes tinal disease (105). bone or joint disease (70). respiratory disease (39). renal or other urinary disease (35). neurological disease (30). fracture* or other ni|unc* (2K). skin or sub cutaneous tissue disease (1 7). malignant neoplasm s o f the digestive organs (15) and bone or skin (2). benign neoplasms o f the digestive organs (4) respiratory system (4) and other sites (K) and various other conditions and sym ptom s (94). 1 his linul category included patients with haem orrhoids (15) and those with symptoms relating to the respiratory system (10). gastrointestinal tract |2U) and urinary system (10) Maximum likelihood estim ates o f relative risk (HR) together with their 95". confidence interval (95% C l) and lest* for trend where appropriate were com puted by multiple logistic regression technique* (Brcslow & D ay. I9gt)) using the (iL IM statistical package (Baker & Nelder. I97H). All relative risks have been ad ju red for age in 5-year strata (20 24. 25 29. . . . 60 64) and for social class m six categories (I.II.Ill non-inanuai. Ill m anual. IV and V). Age o f (lie cases was taken as age ai diagnosis o f ovarian cancer and o f (he controls as age at interview Social class was based on occupation (Olticc o f Population Censuses and Surveys. 19 70) using husband's occupation for ever married women and own occupation for those who had never married Olhcr relative risk adjustments and tests lor trend have been made with the exposures as continuous variables When (he data were examined by place o f interview (London or Oxford), there were no notable differences in the risk estimates RISK I AC K IR S 1 OK O V .\k , \ N ( I K 593 associated with the major variables o f interest. The relative risks have not, therefore, been stratified by place o f interview. The terms nulligravid and gravid have been used to denote, respectively, wom en w ho have never knowingly conceived and wom en w ho have had at least one pregnancy. Parity has been defined as number o f live and still births. 1utile It Relative risks lor o v .m an i am ci a 1IS.M ated n u li p u g n a m i history 1 iiritib/f Ciiaviduy*' (rand Nulhgiavid (! u i (o n tr o l 176 '7 6 S') 74 KH f v y . I I I 1 U* 1 7 (i i 2 6) Results I he age distributions o f the cases and controls are shown in I able I. The average age o f the eases was slightly higher than that o f the controls There was an excess o f eases in social classes I. II. and 111 non-m anual (5K%) as compared to controls (43% ) (H = 0 .0 5 ) and. because o f this, all relative risks have been adjusted for social class as well as age. Table II shows ihc relative risks lor ovarian cancer associated with various aspects o f pregnancy history. Nulligravid women had a higher risk o f ovarian cancer than gravid wom en (KK 1 7 . 95% ( `I I I 2 6). The relative risks were elevated both in nulligravid wom en w ho had been sexually active and m those w ho had not. although significantly so only for the sexually active A m ong those w ho had ever been pregnant, the relative risks decreased as the number of pregnancies mcuMsed. (/' (trend) = 4 3. / ' < .0 05) Similarly, am ong parous wom en, the higher the parity the lower the relative isks (/' (trend) = 3,9. / ' <. 0 05) Alter adjusting for parity, the relative risks associated with successive numbers o f incomplete pregnancies (spontaneous and induced abortions) also decreased although the trend was mu statistically significant <X' (trend) = 0.5) W om en having their first preg nancy after the age o f 35 years had a significantly higher risk ul ovarian cancer limn wom en wuh a first pregnancy before ihc age o f 20 years. I heir risk was also higher than that for nulligravid wom en. There was, however, no marked nor significant trend o f increasing risk the later the age at first pregnancy (X: (trend) = 1 .0 ). Analyses by age j i first livcbirth gave similar findings. A lter adjustment for number o f livebirih\. women w h o had breastfed for more than tw o years in mini had over three lim es the risk o f ovarian cancer co m pared to wom en w ho had never breastfed < 0 . 0 5 ) but overall there was no significant trend the longer the duration of lactation. Analyses o f infertility and subtertility as risk factors for ovarian cancer were restricted to the 2 13 (91% ) cases and 240 (93%) controls w ho reported that they had ever been sexually active. Aumng these wom en. 30 (14..,) with ovarian cancer and 4 (X") controls reported, when so questioned, that they had had problems in becom ing pregnant and. o f these. 16 cases and 12 controls had never conceived Analysis o f the data on contraceptive use suggested that there were otliei wom en who rnighi have been infertile or sublet tile Although sexually active, they had used contraception m frci|uem ly or not at all and had had few or no pregnancies. I'or all wom en w ho had ever been sexually active, the risk o f ovarian cancer increased with increasing duration o f unprotected intercourse after adjustment lor gi.ividity (X: (trend) = 10 2. 1` < 0 01). The effect was most marked am ong nulligravid wom en w ho reported more than 10 vears if unprotected intercourse. Their risk was over six times that o f nulligravid wom en who reported less than three months of unprotected intercourse (Table III). Am ong gravid women, those reporting over 10 years o f unprotected intercourse had a higher risk than other gravid women There was no significant (icnd in risk associated with the duration o f use o f any I able I Age distribution ,nul average age o f eases am i coiU iok (< i u - ir x l 1 U 5 W 4< 54 <5 M 1Mal Vciagc age (years) C Hu- i % ) n |5 5| 27 i l 1 5 87 (.37 01 108 (-16 II) 235 '2 4 C ontrol* r") 33 (7 l) 75 (1 6 6 ) 156 (34 6) 1X7 (-11 5) -51 51 4 sexually active Nulligravid and n o c i sexually aelivc 17 44 22 hi 1 9 (i i t h 1 5 (1) X 2 to N innivi o f pregnaneies 0 1 : 3 4 5 .59 74 ir 41 7| I X . lO 4 1 tl 6 ' 107 li 7 1)4 1 II 37 9X tl 5 (II t 0 X) I t 41 0 4 (0 2 0X1 2o 59 0 4 (O 2 0 8 ) /' lor trend - 4 i r< o u5 (gravid moiiicii only) I mmi a led reduction m relative nsk associated with each piegnaney II X(| |(l 7x U `)4| I 'a r il y " II 1 2 t 4 S 66 X7 1 II' 4X 84 11 7 III 4 1 2l 61 127 0 6 10 4 1 01 40 XK li f. (0 1 1 III 12 to 0 S m 2 1 0) X t4 0 \ 10 1 7l / h u trend -- t ) /' II IIS (parous vc. men only ) I M ini.iled reducinm in lel.m ve iis k associated w ilh cavil b u lli t) X4 p i 7S n 'M i N>i ul iiicoiiipleic picgll.lllCICs"' (1 1 2 ^3 IKS 39 7 4 / ltd 83 IX 19 fui trend 1 ir II 9 tl 7 tl 6 u 5 1 sum m ed fcdnciion in relative 0 92 risk associated w ith Cai h incom plete pregnancy (0 6 14) IO t 1 X) (0 2 1 7) It) 75 I I I ) Age ul tirsi pregnancy tyeais)' IS 19 20 24 25 29 til 34 ri '5 Nulligravid 26 7t 49 17 9 S9 / lot iiend 65 1o 1X2 II 9 (0 s 1 S| 96 1 2 (0 7 2 2) 29 1 2 HI X 2 7) 4 4 1 III |S || 74 2 o I I I ' 7| 1.0 tgiav id VI..1.)cn u tili ) Moniliv o f l.iel.ilion' None ^0 7 12 I t |K 19 24 s? 25 44 (<6 29 13 S 12 / in; 1ir 124 1 i Kll 119 29 1 2 7 21 IS t 4 lor tremi > 8 U K 2 2 (II 5 1 6) |0 5 2 '1 (0 7 6 7) tl 1 IHM All relative risks adjusted for age am i vociail d a w 'R efer c in e category "D ata missing lor 1co n tro l ' Relative risk'. adpisit:d lor parili -'l)a la m u sin g lo r 2 cum and 1 coiitr ol W om en a u h Irvi.h irilis onlv R elative risks adjusted for num ber o f live hifllis contraception (X' (trend) = I 2) although sexually active milligravid women who lu d never used any method o fco n ira ccp n o n had about twice the risk o f ovarian cancer com pared to all other sexually active wom en (Table IV). Of the specific m ethods o f com raception studied, ever having used oral contraception ami having been sterilised were associated with a statistically significantly reduced risk o f ovarian cancer, while no method was associated wuh a significantly elevated risk (Table V). As only three cases had been sterilised it was not possible to assess whether age at sterilisation influenced the risk o f ovarian cancer, fable VI show s detailed analyses o f the relative risks associated with oral 594 M * " > O lH , ; I Me III Relative fisi.' Mi o u r i.tn cancer associated w lh duratio n ol tnipiouM cd intercourse h> gr.tvivJuy ('irv i < nrro/s RR Au//uiram / h/(-' D uration o f un- plutei Id i imcfiOUI'C III:i.>mhsf $3 4 M4 12 2ft 1 0* M 12> 4 ~ 15 :> i m 1 4 07 20 X 6 5 / ' ior trend = 11 2 /' > (HI 1 fV.5% C l) (0.4 6 5) (Ori- 7 Kf (2 1 20 4) D uration ol un- protested micrcouisc miondisi'' s: t 4 Ml 7X 176 M 120 51 M3 > IM 10 27 37 Ml / lor trend -* 2 6 1 1 (0 5 2 4) 1 1 (0 5 2 6) 1 1 (0 4 3 2) 1 6 (0.7 4 0) Sevually active wom en only Relative risk adjusted for age and social class `Reference category. "Time when sexually active arul at risk ol pregnant.> but using no contraception contraceptive use. The rifles decreas'd as duraiton o f use increased allhough, am ong those who had ever used such contraceptives, the trend was not significant (Z: (trend) = 1.2). Whatever their age at first use. women who used oral contracep tives had a lower risk ol ovarian cancer lhan those who had never used them, ihe risk being lowest in those w ho had first used oral contraceptives under the age o f 25 years. The risk o f developing ovarian cancer did not increase as lime since discontinuing use increased W omen who had stopped using or.il contraceptives' morc iha ten-ye;its__prcviou\ly had a statistically significant reduced risk o f 0.3 com pared to wom en who had never used them. Women both under the age and over (he age o f 40 years had a reduced risk o f ovarian cancer associated with oral contraceptive use. but (he reduction was greater in the younger women. Ciravid and nulhgravid wom en who had used oral contraceptives had a reduced risk o f ovarian cancer. Table VII shows the relative risks associated with age at menarche and age at natural m enopause. There was no (rend in risk with age at menarche (X; (trend) * 0.03). In contrast, risk increased the later the age at natural menopause (X: (trend) = 7.1./* < 0.01). Women having their menopause at the age o f 50 years or later had nearly three times the risk o f wom en w ho were menopausal before ihe age o f 45 years. The risks and trend associated with age at m enopause were similar irrespective o f whether they were adjusted for age in live year or one year strata. 1able IV Relative risks lor ovarian cancer associated with d u ra tio n o f use ol contraception by graviduy (.u s ti Controls RR iV 5 r. < 7i W btr'tiiii/ ni!n 1)iir.ili<m ol use >1 > liuicplion Never used s Hi yeais Hi 2n years "> 20 years / 15 14 6 i for (rend 0 9 Hr 21 9 4 1 il* 0 5 (0 1 1 7| 1) 5 (0 1 2 5) 1)4 (0 1 3.2) i ..... D uration ol use ol .'oil- tr.neplion Never used 32 47 < 10 years 10 2d seals > 20 years 25 56 55 147 64 J26 / Mi trend 0 3 0 4 (02 I t) 0 3 (0 1 1.0) 0 4 (0 1 12 0 5 (0 2 15) Sexually active vvomen only Relative risks adjijsicd for age. social class anJ Juration ol unprotected intercourse `Reference category ruble V' Relative risks lor ovarian cancer associated vvith the use o f ditlercnt me!hods of coniraccpnon M fthm l of ouurth.fptton C u \t\ ( u n iro l RR (V5U. C h Sheath Ncvci used F.vcr used I0H 205 105 215 1 0* 1 1 |0X 1 7) Diaphragm Never used hver used 178 329 35 91 1 (r 0.7 (0 4 1 1) liur.iuierine Ncvci used dev u c Hver used 201 3X3 12 37 1 (r 0 8 (0 4 1 7) Oral Never used contraception fcver used 178 306 35 114 1ir 0 5 (0 3 0 9) Parm er with vasectomy No M l 404 1 0` Vex 10 16 : i (0 9 4 9 1 cm alc sienlisaiKdi No 210 375 1 0* Yes l 45 0.2 (0 1 0 6) Other m ethods'- Never 1.\cd 1 ver used 156 292 57 I2X 1 O' t i (0 7 1 7) Sexually active wom en only Relative risks a djusicd for age. social class, gravidity and d u ra tio n o f unproiecicd intercourse ` Reference category 'U se of spermicides, rhythm or conus interruptus. I able VI R elative risks l'or ov.in.in caiwer associa led with oral contraceptive lO O use D uration o f O C use (years) Never used s2 5 5 to II) Age ,i( first O C use ( lis ts ( Illin o is RR r v5".,, C / l 17x lllf. 1 0* 24 70 0 6 (O i 10 29 0 6 ( 2 1 15 0 1 (0 01 / ' lor lrend within users 1.2 \ 0) t 4) 1 0) Never used <25 2 ' 29 3() 34 35 I7X 306 1 i r 6 19 0 1 (0 04 o 5) 6 17 0 6 <0.2 2 0) II 27 0 7 (0 3 1.6) 12 11 0 7 (0 4 1 5) / **' trend wiihm users - 5 9 / ' < 1) 05 ] une since discontinuing O C use (years) Never used I7K 106 1 0* Current users 6 19 0 5 < 5 12 24 0 X 5 10 9 25 0 X 10 X 46 0 1 C lot (rend within users - 2 6 (0 2 1.5) (0 4 V) (0 1 IV) (0 1 0 7) Age (years) < 40 OC use Never used Fvcr used II il 1 0* 9 35 0.2 <0.1 0 9) ^ 40 (K use Never used hver used 167 295 1 IF 26 79 0 7 (0 4 1-2) (iru su liti (iravid women" O C u se Never used Hver used 149 2X0 1 0* 27 96 0 5 (0 3 0.9) Nulhgravid women (X ` use Never used Hver used 29 26 I 0* X IX 0.3 (0 05 2 8) Sexually active women only Relative risks adjusted for age. social class, graviduy and duration o f unproiecicd intercourse `Reference Category ''R elative risks a d |u sicd for graviduy Tabi* VII Relative nsk l'or o s a ra n cancer associated wnli age al m endiche jn d age ai naturai m enopause i 'llM*s ( 'n lllra h R ii < W . - C D Age al m enarche (years)" >14 12 13 < 12 97 K9 46 X' for trend --0 03 197 IK5 66 l 0` 0 9 <0.6 1 3) 1 3 (OK 2.1) Agc.at_nal.ural menopause (year)' < 4 5 ID .14 45 4V 47 77 >5U X4 99 l ' lor irend - 7 1 l ` <0111 1.0* 2 0 (11.9 4.7, 2 5 ( l .l 5.8) Relative risk adjusted for age and social class `Reference category "Data on age at m endiche m issing lor .1cases an d .1 controls. 'D a ta on age at m enopause missing lor 2 eases and I control. Women who reported hysterectomy, with or without unilateral oophorectom y, had a much reduced risk (Table VIII). Since there were only 10 wom en with ovarian cancer who had hud a hysterectomy it was not possible to assess ihe effect o f age at hysterectomy on ovarian cancer risk Total duration o f ovulation was estimated as the months from menarche to diagnosis (cases) or interview (controls), or to menopause, whichever cam e first, m inus the total m onths o f anovulation due to pregnancy and oral contraceptive use. Women who reported a hysterectomy were excluded from these analyses as it was unknown if or when they had stopped ovulating. For all wom en com bined, there was a strong trend o f increasing risk the longer the duration o f ovulation (X' (trend) = 17.X. /'< t).(K )l) lia b le IX). In separate analyses by menopausal status, there was no significant effect o f duration o f ovulation alter adjustment lor the `anovulatory' factors used to estimate that exposure, nam ely, m onths o f pregnancy and oral contraceptive use and age at m enopause for post-m enopausal women and m onths of pregnancy and oral contraceptive use for premenopausal wom en. Duration o f ovulation is very sensitive to age hut the risks and trends were virtually unaffected when adjusted for age in one year rather than live year strata. l ive (2% ) cases and 29 (6% ) controls reported having taken hormone pills as a pregnancy lest and live (2% ) eases and 13 (3%) controls had been given horm ones to prevent miscarriage. For all post-m enopausal w om en, there was a small but non- sigm licanily increased risk o f ovarian c w iciatcd with ever having recicvcd horm one rcplaceinei. .apy (Table X) The excess was confined to women who had reported a hysterectomy who had an 11-fold risk. I he cases did not report more severe menopausal symptoms. A m ong the hormone treated wom en with ovarian cancer. 23' had endom etrioid or clear cell tumours com pared to 3h" in the untreated women. The reproductive and related factors found to be statistically significantly relaied to ovarian cancer risk (gravidity, duration o f unprotected intercourse, use of oral contraception, having been sterilised, age at natural m enopause and having had a hysterectom y) arc not independent and we also compuieCnTie- relalive risks associated with each factor alter adjusting lot the others (Table XI) A s sterilisation is often a consequence o f high parity, the risks associated with gravidity wore not adjusted lor sterilisation as this was considered to be overadjust men t. In this study. 40% o f the sterilised wom en had live or m ore children compared with 9% o f the unsierdised wom en Each o f the variables remained s'tatisiically .significantly related to ovarian cancer risk, suggesting that each may be independently associated with the risk o f developing ovarian cancer. There was no significant difference between the percentage o f cases (53% ) and controls (57%) who had ever smoked cigaret tes. N o cases or controls reported having worked with asbestos. N o cases but three controls reported a radiation-induced menopause W om en who reported using talc more than once a week or daily had higher risks o f Ovarian cancer than wom en who reported less frequent use (Table XII). Although the relative risk o f 2.0 associated with weekly use was statistically significant Table V ili R elative risks lo r m a n a n can cer jsm ci.iIciI w iih ic poi lo l history of hysterectomy uiui/or unilateral oophoiocioi) ( <!'. ( 'onirof* H Ii V i ",. ( I I Reponed womb miaci Reported unilateral oophorectom y hy no h ysterectom y R eponed hysterectomy hut conserved varies Reported hysterectomy and unilateral oophorectomy 2 2 o 170 1 I f 5 9 0 9 (0 4 2 1) x 62 t) 2 (0 1 1)4) 2 III 0 4 (0 1 I D Relative risks adjusted for age am i social class `Reference category Table IX R e la m e risks lo r ov aria n cancer associa led wait J u ra tio n o f ovulation Pllfll/IH/l of oriihui i rw ; <'<<. ( oniroh fa'Ii f u r risk uilittli-,1 fo r une a n il > <in/ Vii.vv fa-lullii* m b tu b im eli f or tier. <*<tu l t An m n l ihiriiiu >n o f anovulation H e l m e n \k m l/iifieil for ajee, m la i Ai. iuro Iio n o f anovulation am i one a l rnenofHiu^e All w om en" <30 10 .14 15 39 >40 59 163 I tr 71 106 20 6K 92 20 19 13 43 X; fr trend = 7 K f> < 0 001 Post-menopausal women < 30 30 34 35 19 >40 14 53 1 r 1 If 1 tr 54 hi 24 21 09 5X 72 24 19 06 14 10 50 4.0 07 X' for trend = 2 3 P < 0 (Mil 7 7 / ' < 0 01 05 Premenopausal women < 30 30 34 35 39 >40 45 19 10 s Women reporting a hysi reeiomy excluded. 'R eference caiegory contraceptive use 110 25 20 3 X' fur trend - 1 0* 15 13 32 4 4 / ' V. 0 05 1 (r ii 09 19 0.6 IXiia missing lo r 6 cases and 4 nirols. 'T otal m onths if pregnancy and fit5^0 v I 11MIII . r I ablr \ Kclalisc inks lor osari.ii) cahier assH.ulcJ ilh the use ol hoim onc icpl.ucmcm therapy lor menopausal symptoms i./* Control* Kfi i.<*- Cil Ail post-m enopausal women I V ol horm one rcpl-ucmcni ihciapy 5 es M 44 1 5 (l) 9 .H i W .'IIICII u p o n mg h>>ictcu.'in> l 'sc o f horm one fcpl.iccn.eni lhciapv No ' t*l 1 O' 5 10 III 9 ( | 7 ft9tl) IV 't mcn.ip.uis.il ..m en other Ih.in il...sc icportuig lis-icicvt>nt) 1 sc .a horm one ic p U c m e in iui.ij>s No yvs 177 187 i tr .'4 >4 1 : (II 7 2 3) l,.*'t-nwii>ip.uis.il .men only Relative n 'k s adjiisled lor age and ..vial JasN R clcferuc .alcgoiy l a h l r M Relative i i 'U associated with i lie 1.ici>>f * lo u n d lo he sigiiilik.illlle Icl.ilC il I Os.Ul.ini ea rn er H K , lest for if end (1 d 1 ) h 1 t 4 ' I M|H..K-.U-,I 1 (' O s l<* 4 1 S) M.s (114 1 4| l i e |ll t 1 | | ll 4 ill 7 1 Ul o ' III 2 1Oj <i S / ' - il5 4 r-i f.l l.'u l.'n 1 IMS .' HI l l Hi 5 2 5) l ' <12 <2) 7 8 /' < ll 05 t >i.il v.mii.uepiivi ,is.- Ncs. i ,,sed 1o* S s seals Ilf. f. 1 seals ' 111 SC Ils n1 HI 3 III HI ' 1 4) U) n2 1 h 4 f. r * onv 1 VCl OCIlt./Cl n ; (II os O0| \g.e .u n.mit.il ineu. pause' <. 45 seals 4> 49 seals > ^ii sears 1 II ] 9 III 8 4S | 2o i l l (> 1) 8 2 r < o ul s e e , h., led ll 2 iisNiefCki.um' m 1 0 5r 1lie lel.iiise fok' .iss.ikiaied skuh eaill laei..r l.ase been adlusted lor age. social ..lass ail..1 all ihe ul her taeiors m Ilie table `Kclerenee category "NcvIi.llls av i n e women only Rela lise risk s not adjusted lor siefili/aii.in sec test !>! details 'Women rep- fling n.iiur.il menopause only / ' v. ii mil I able M l Relative iisks lor os jrun cancer associa led wiih reported Ircqik:cy id tale use in the girmi.il area ( iiwi ContiiJ\ tip iV.<% <7/ Reported IrcqueiKS ol laic use Nesci Karels Mon this WeekK Dads 76 178 1 It- t> 16 U9 (0 3 2 4) 7 24 0 7 (0.7 18) 57 77 2 0 (1 .7 3 4) 71 I 9 1 .7 (0 8 1 9j X" for irend = .780. P = 0 05 Relative risks admsied lor age and social elass `Re-ferenee category 1>jia missing loi 18 Ih* 1eases and 17 (4U0) control s as questions on talk use iiilfosluee.l 'Ihree monihs alief stud) began {P = 0.007). (here was no consistent trend ol increasing risk wiili increasing frequency o f talc use (X: (trend) = 3.HO. P = 0.05). There was no significant difference between ihc percentages o f cases (86% ) and controls (81% ) who had used and kepi their diaphragm in talc Discussion As in most previously reported studies (Booth & Bcrat. 1985) we found that nulligravid wom en had an increased risk o f ovarian cancer and that risk decreased as the number o f pregnancies increased. We also found that the greater ihc number o f incomplete pregnancies the lower the rtsk. although the irend was not significant. Most other studies have not investigated if women o f low gravidity have an increased risk o f ovarian cancer because o f reduced 1erlilily or because o f voluntary limitation o f family size, although July c l ut. (197-4). M cG ow an 7 at. (1979) and Nasca c l at. ( 1*484) found a higher risk in wom en who had tried to conceive but had failed. Our findings also suggest that infer tility is a risk factor for ovarian cancer W om en who had not conceived but had been sexually active for more than It) years without using contraception had about six tunes ihe risk o f all other wom en. Approxim ately hall these women had undergone investigations for infertility, l or only live cases and one control was the cause o f their infertility deter mined. Thus, it is not possible to assess whether this high risk group had normal or impaired ovarian function. Subfcrtility may also be associated with ovarian cancer Gravid wom en icpornng over 10 years o f unprotected intercourse had a 5l)"o higher risk than other gravid w om en, but this increase was not statistically significant Age at first pregnancy was not found to be associated with ovarian cancer risk although wom en having a first pregnancy alter the age o f 35 years had a higher risk compared to wom en having a iirsi pregnancy at earlier ages and to milligravid wom en. Since subfcrtility might he a risk factor lor ovarian cancer, the relative risks associated with age at lirst pregnancy were also ad lusted lor duration ol unprotected intercourse. The raised risk for wom en with a lirst pregnancy alter 35 years persisted (RR = 3.9. 95'% CI I I 14 2) Results from other studies regarding the risk for wom en having a lirst child at relatively older ages are inconclusive, som e Imdmg no association (N cw h ouse a a t . 1977; Casagrande ct ul.. 1979. Cramer <7 at.. 198 3; I esher </ at.. 1985). others an increased risk (July <7 t it . 1974, M cGowan <7 at.. 1979; Hildreth a a t . I9KI, hranccschi a <//.. 1982). O nly Traneeschi <7 ul (1982) found the increased risk to be statistically significant and independent o f parity Overall, there was no association between use o f any con traception and ovarian cancer. O f the specific methods studied, female sterilisation and use o f oral contraception were associated with a significant reduction in risk and no method was associated with a signilicant increase in risk. The associations remained alter adjusting lor gravidity and dur ation o f unprotected intercourse, ihe measure used to indicate infertility. While few studies have examined ihc association between female sterilisation and ovarian cancer, the relation between oral contraceptives and ovarian cancer has been demonstrated in m any studies (Booth & Beral, 1985). l ike others, we dcm onsiiatcd that the longer oral contraceptives had been used, the lower ihe risk Our findings also suggested that the earlier the age at lirst use ihe lower the risk and lhai the protective d i e d o f oral contraceptives persists after their use is stopped It has been suggested that inhibition o f ovulation, as induced by pregnancy and oral contraceptives, is the factor which protects againsi ovarian cancer (Fathalla. 19 7 1>. If so. postpartum anovulation associated with lactation might also be expected to be protective. We found no evidence that ihe longer a woman had breastfed the lower her risk o f ovarian cancer Indeed, the highest risk was found in those who had breastfed longest Results from other studies arc contradic- KISK I A( I OKS I OK OVAK A N t I K $97 lory (Cramer e l a i . 19K.3; Mori et a t , 1984. Risch et u i. 19X3; Cancer and Steroid H orm one Study. 19X7). Our finding that age at incnarchc was not associated with risk is consistent with results from most other studies (Casagrande f t u i, 1979; M cG ow an ct a t , 1979; Hildreth f t at.. 19X1; I'ranceschi e l at., 19X2). Age at natural m enopause, however, was strongly related to risk. While Hildreth et at. (19X1), l-ranccschi / ut (19X2) and T /o n o u et at. (19X4) also demonstrated lTiat the laler the a g c a (m en o p a u se the greater the risk, other studies have found no association (W est. 1966; Newhousc f t a t . 1977; A nnegers c l a t.. 1979; M c(row an e l at.. 1979; Cramer el a t . 19X3). A lower frequency o f hysterectomy, o f unilateral oophorec tomy. or o f both am ong cases compared to controls has also been reported from several other studies (W ynder et a t . 1969; Joly c l at.. 1974; Annegers et ut.. 1979, M cG ow an el at.. 1979; Kranceschi cl at.. 19X2. Cramer et at.. 19X3). As these studies were case control in design, there m ay have been some m isclassilicalion o f controls who. rather than having a hysterectomy with ovarian conservation, actually had a hysterectom y with bilateral oophorectom y. Another explanation lor the findings might be that if at hysterectomy a woman's ovaries look diseased, n is likely ihal they ate removed. If the diseased ovaries were precaiicerous. those women who might otherwise have developed ovarian cancer do not. hollow ing hysterectomy with ovarian conservation, reduced ovarian function or ovarian failure occurs in a pro portion o f w om en, due possibly to the blood supply to the ovaries being com prom ised (Reavis et a t . 1969. H lsw orlh el a t . 19X3). heale sterilisation was also associated with a low risk o f ovarian cancer. Neil el at. (1975) have suggested that the menstrual disturbance that many women experience after sterilisation may reflect changed ovarian function due to damage to the vascular supply to the ovaries. If both hysterectomy and female sterilisation can indirectly a licet ovarian function then both procedures could also influence the risk o f ovarian cancer. Kcicul investigators have shown that the longer a woman ovulates the greater her risk o f ovanati cancer (Cusagr.iiklc </ a t. 1979; Hildreth c l at.. 19X1; f raiiccschi eta !.. 19X2; Wu e t at.. I9XX). We also found that risk increased the longer the duration I ovulation. D uration o f ovulation is. however, highly cor* related wuh (he 'anovulatory' factors used to estim ate the exposure. In an attempt to determine whether duration ol ovulation had an clTcct over and above (hat expressed by its relation with these factors, the risks and trends were adjusted for duration o f anovulation due to pregnancy and oral contracep tive use and. where appropriate, for age at m enopause. The significance o f t he ellcet disappeared. We conclude that it is not possible to determine from these data wheth er u. is the above factors which inhibit ovulation that prevent ovarian cancer, whether repeated ovulations promote it. or whether a co m b in a tion o f both is acting. Our finding o f no overall relationship between horm one replacement therapy and ovarian cancer supports those o f other investigators (N ew housc et at.. 1977; Hildreth e t u i . 19X1. W eiss e t at.. 19X2). An increased risk associated with the therapy was found am ong women vvho reported hysterectom y, but the finding was based on very few eases and may have been due to chance. We did not find an increased risk associated with oestrogen therapy for any particular tumour types as suggested by Cramer c l at. (19X1) and W eiss et at. (19X2). The evidence linking talc with ovarian cancer is controversial (A nonym ous. 1977. Roe. 1979; Longo & Y ou ng 1979. Ci.uiicr c l a t.. 19X2: Hartgc cl a t . I9H3) In ilus study, wom en w ho reported talc use m the genital area more than once a week oi daily had higher risks o f ovarian cancer than wom en w ho used talc less frequently. Tile women were not asked how long they had been using talc, it is possible that because o f their sym ptom s or disease-related pelvic examinations, the frequency o f current talc use by the cases may not have reflected their frequency o f past use. Since these ami other results (Cramer c t a t.. 19x2; lla rtg c ct at.. 19X3) are m sutlieieni to rcicci an association, furl her work is needed on the relation between genital use o f talc and ovarian cancer. We would like lo ih.mk die Im pcn.ii C ancer Kcw.iich I und i. h liiiaiicmg die study. Piolcssor Sir Kicliaid H ull lot Itelplul advice Professor C hristopher H udson and l)i M angold ( inling lor reviewing die histology. I >r I v c W iltshaw. ibe eons nilJills, m uses an d olliel Mali o l th e fi.iriici|u lm g hospitals lo t ilicir co-o|K*r.itioii am i s n p |t,n i o l die Ntnvly. Anil H.iiciuaii. K ale K odriqucs. (iilli.m Saunders and Rosalie I lioinson lo r then skillul mlciv-icwmg. and Nina S am i for typing the n iaunsetipl M argarel flood is lunded hy a grain from die M edical Reseaich Council Urfi-rmees VSM.UKK.S. M MKOM II. DKKIK. IHI. IKXKIKIV. MH A OI Al I ON. w m (1979) Ovarian cancer incidence and case control stud). Cancer. 43. 723 anonv MOllS (1977). ( \ m c n c tale pow der E ditorial. tu rn a . I. I34X HMil R R ) A Ni l | ) | R. ) A |I97X> The C U M S t cm K c l c a w R oyal Sl.ilistie.il S>Kiely. O xford HI V \|\. l i e ; . HKOWN. I H A s M Illl. M -v (1969) O v arian finielioii -itier hysterectom y w ith eoiiserv.iiion o f ilic ovaries in prem enopausal w om en. J O h xici i . i <(..<>/ Br ( u t i l l . 76. 9f9 HtKUH.M * HI r a i . v ( |9 x 5 ) Ilie epidemiology o f o varian e.ineer In Or.irnm tu rn e r. H udson. C N ted), p 22 O xford llrnvcisily Press. Ovlord liKISIOiV N I- A l)AV. N I ( I9XI) S l,n i\tn < it l / . - / W t mi tu r n e r Rcw o rth . I HI. /. The .4nuh%n C a w c n n lro t S tin lic \ I ARC' Scienulk Publica lions N o 32 IA R (' I yon t VNl'KR AND STEROID HORMON) SIHDV Ol r i l l O N I KI.S I OK DISt.ASI- CONI ROI AND till- NAIIONAI IN S M IH II Of1*1111 D III A l.Ill AND HUMAN DPV H O l'M E N \ (19X7). The reduction in risk o f ovarian cancer associated with oral-coniracepnve use .V Engl J M e t. 316, 650 t xSAORANDt. J 1 . PIK E . M f . ROSS R K . I O lilh . I: W . R()3 . S & HENDERSON. B E (1979) 'Incessant ovu latio n ' an d ovarian cancer. I.tu itri. ii. 17(). xixAMKR. I) W . DEVESA. S S & W U < II W ft ( |9 X |) T rends m the incidence o f endom etrioid and clear cell cancers o f (he ovary in (he United S u te s. .4/. J. F.puknuol . 114, 201 i hamek.i>w .wn.cn. w r s<ui i y. r e &w ornec/iiow.ski.t a (19X2). O varian cancer an d la k . a case-control study ( mu er 50, '72 < KAMI R. D W . I l l i l t IIISON < H W | I ( | | . W R . S< UI n R I A RYAN. K ) ( I9 x ') D cicrinm ants o l ovarian eancet risk I. R ep roductive experiences and family history. J Airi l t urn e r Ih m 71 711 I I I SWOK III. I K . A l l ! N. II II A NIS'kl R. J -V | |9 R l | O v .u ian Innction a lte r radical hysterectom y for Stage I If c.iicinom a ol cervix. Am. J Oh.xtct t i i ii c c o t . 145, 1X5 I A I'll At I A. M l ( l `>7|) Incessant ovulation a ta ilo r m ov.ui.in neoplasia * I in n e r ii, 163 IK ANt'l:.S<'III. S . LA V'HCIHA. ( ' . Ill I M RU'll. S P . M VNOIONI ( .v ltK iN O N I. (i (19X2). Risk factors tr epithcli.il ov aria n cancel m Italy, twi, J E p id e m io l.. 115. 714 llAR K iE. P . HOOVER K . I.I.SHIK. I P .V M .t.OW AN I ( |9 K l| I ale an d o varian cancer J Am Mc<t A w tw . 250, 1X44. H it DRI TH. N ( i , Kl-l SKY. J l IIV O IS I. V A A $ others ( |9X 11 An epidem iologic study o f epithelial carcin o m a o f die ovary Am J E p id e m io l .1 1 4 , 39X JOI Y. 0 7.111 U NI I I D A M . DIAMOND. 1.1 a HRDSS. I D J (1974) An epidemiologic study ol the relationship ol reproductive e.vpertenee to ea rn er o f the ovary. Am. J E p id e m ia l. 99. 190 I.LSHLK. L . M diOW A N. I .H A H H ih .P & IIOOVLR. K (19X5) Age.U first birth an d risk o f epithelial o varian cancer J .Yn it t\ir n c r fm l . 74. 1361. I.O NtiO. I) I. A YOUNO. R V (1979) C osm etic tak and o v arian callee I Mm c l. ii, 349. M. COWAN. I PAREN | . I | KDNAR. W * NORRIS. It J (1979) The w om an at risk fot developing o v arian cancer t .'i m i a t O m n i 7. 325 J. i wucr (19X4). I. 59m 600 The Maculili.. css I id . IVMM s, M V . . , . A M IV .V H H < 1 ^ 1 ........... ...... . "1 >. nuci ni I.iii.in ( jfiiiv. SJ. 2746 s vsi' n l. R I I N ^ l l > r C m o r o s i. s RK H ^ r K . * l A N H O ' , , l J l Anepulf...... wl.u iivc' ijvlKi *" J /'juleiiuol. 119, .05 M M IR IIW M IIM ' l . l RUSH IO N . I * I H U ,'-,,K M ,, s I . 1 ^ 1 IRK .>1 by S| ' i h .'ir h .n O l-- ' m l M U S b .'S HS IIS77I A tJsc control slu j) *il cjrunonu / . m o n u n s u s I's " ,, I f s i . R v m N 1h" !. ... irpr.slii.loe Nfr ,nd .he me,de,ree ol eP,,hel,al ,,vana,, earner 4m J fpstmr.d . 117. IW KO, I J . | W ) ( . ,, ,, . o v e r e y e ,, - m e ,, e laie a ,, d o v a n a n c a n e e , I> * 4 . ,, h e ,, l l d M , T h e eptdcn,toit,Ry tri osar tan eaneet tnlrrecee. a .aee-eoniro .lu y r/r. J ( amer Chu Orno! 20, IIM5 BT SH U H ` ' S l l ' r Z . 1 1 1 . * > P * H l,N B A R l,l R. R s 4 7 ,, r n e M m ill Personal and cnvrronmcnlal character,Mies related lo epii e ta ovariTn cancer I R .p ro d u c v e and m en.lrua, even,, and O,a, ^ cancer of ihc ovary <nnrer 23. 52 Eptdetntnlogy o . . SHORT COMMUNICATION ( ' Pulmonary giant cell carcinoma: the relation to smoking R II Depue1& B R Ballard- `.V612 Bunnell Dr.. Potomac. M D 20X54: !Department o f P athology. Universit o f M ississippi School ol M edicine. Jock .son. M S JV2/6. USA The relation o f sm oking to the occurrence o f the most com sm okers am ong men with gianl cell carcinom a o f the lung mon types o f lung cancer has been examined and found has a 95% confidence interval o f 11.86 0 96 (R othm an & positive in epidem iological studies. Types that have been Uoicc. 1979). This confidence interval docs not com e near to causally ascribed lo smoking in such studies have included overlapping any o f the estim ates o f the prevalence o f sm o k squamous cell, small and large cell, and adenocarcinom as ing am ong U S m ales and. therefore, is highly statistically (US Dept o f Health & Human Services. 1982, International significant (Table I). M ost reports did not include inform Agency for Research on Cancer. 19X6). H owever, rarer types, ation on the years or am ount sm oked, so wc could mu such as gianl and alveolar cell carcinom as, have not been exam ine the dose rcponsc. subject to separate study, not only because o f their rarity, but Smoking prevalence data, based on a US national health because som e pathologic classification system s for lung survey (U S Dept o f H ealth A Human Services. 19X3). show ciiwer include these types with other histologies (Yesner A that 52%. 42% and 3K% o f all males were current sm okers Carter. I9K2). While all histological types o f lung cancer in 1965. 1976 and 19X0 respectively, while 21)%. 30% ami studied thus far have been related to tob acco use and no type 31% were former smokers. Therefore. 72% . 72% and 6 9 " u has yet been found to be unrelated, there have been questions o f U S males had a positive sm oking history in those years raised about the relationship o f sm oking to these rarer types. The prevalence o f a history o f sm oking in males is then Giant cell carcinom a o f the lung was first described by about 71% over the period 1965 19X0. U sing that rate, one- Nash and Stout in I95K. Som e investigators classify it us a can calculate a relative risk o f 5.3 for pulm onary giant cell separate entity (Shin et al.. 19X6). On the other hand, some carcinoma due to ever having smoked. M any o f our 23 studies have included it as a sub type o f adenocarcinom a o f the lung, did not distinguish between current and form er smokers. while both the World Health Organization and the Armed Therefore, the risk estim ate o f 5.3 is lo o low if the smoking forces Institute o f Pathology include it with large cell status in som e o f these studies included only current smokers. undifferentiated carcinom a (R az/u k e t u f. 1970). These T o obtain an upper limit to this estimated risk, wc can several schemes lead to non-uniform criteria in the literature com pare our tabulation to the 44% mean prevalence o f for diagnosis o f pulm onary giant cell carcinom a. A survey current sm oking in the U S data from those three years. Such such as ours cannot resolve this difficulty and. therefore, we calculation yields a risk ratio o f 16. This range o f risk musi accept the cases reported as genuine for our purpose. (5 3 16) is equal to or higher than the relative risks pub Tins survey is an attempt to provide som e inform ation 11shed for other histological types o f lung cancer (U S Dept ol hearing on the aetiological relationship o f gianl cell car Health A Human Services. 19X2; International A gency for cinoma o f the lung to sm oking We searched the literature, as Research on ('nicer. 19X6). Our use o f U S smoking indexed by M edline through 19X6. for case reports o f pul prevalence for the years 1965 19X0 as a population com monary giant cell carcinom a which included sm oking his- parison is jsutilied by the fact that the reports w e used were IMI ICS. from I95X to 19X6. essentially the same period, ami (hat X.l% We found reports o f 119 cases o f pulm onary giant cell o f our reported cases were from the US carcinoma with sm oking histories o f the patients. O nly eight There is a possibility that the authors o f the case histones of these cases were female, live sm okers and three non may have m entioned a positive history o f sm oking more smokers. which are lo o few to analyse separately Therefore, readily than a negative one. This hias cannot be com pletely we confined our analysis to males. The 111 male cases with excluded. However, in order to have the 95% confidence smoking data were found in 23 independent studies (Bcndel limits (R othm an Sc Hotce, 1979) o f our observed proportion & lshak. 1961; Broderick et al . 1975; Dailey Sc. Marcuse. overlap the expected 71% prevalence o f a positive history ol 1969; DeA ngelis e t al., 1961; Flanagan & Roeckel. 1964; sm oking, at least 20 non-sm okers must have been Inedberg. 1965; Gajarai et o f , 1971; (rutilai) Sc Zclman. unreporicd. This would mean that the non-sm okers would 1966; H athaway et o f . 1969; I(clistron) Sc Fisher. 1963; Ilo n e have to have been at least 70% under-reported lor our result & Ohla. 1981; Kallcnbcrg Sc Jaque. 1979; K ennedy. 1969; to become non-significant, winch is an unreasonably high t Lerner. 1967; Naib. 1961; M atsuo et al., 19X6. Nash Sc Stout. proportion I95X; PfefTer Sc Sloven. 1978. Pfuzer Sc K noblich, 1975; Shin In summary, we abstracted sm oking data from 23 pub it ul., 19X6, Thom as. 1962; W ang et u f. 1976). (Although not lished clinical reports o f giant cell carcinom a o f the lung in reported as such in the publication, on e wom an in the study 111 men. o f whom 9.3% were or had been smokers. This of Shin et al. (1986) was a smoker, as were 8 o f the 13 men. proportion is significantly higher than the 71% prevciance of Shin, personal com m unication.) Wc also found 99 other a history o f sm oking am ong US men in the corresponding cases. 81 male and 18 female, in other studies which did not con'ain smoking inform ation (12 references not cited) therefore, the male: female sex ratio for pulm onary gianl cell Table I Prevalence o f sm oker* am ong m ales carcinoma is 7 .4 :1 0 ! the I 11 men with giant cell carcinom a o f the lung, only Smokerx Odds ratio P rulin' fight (7.2%) had been non-smokers. This 0 93 proportion o f G u m cell lung cancer cases 93% (9 8 /1 11 ) Correspondence. B Ballard. Departm ent of Pathology. University of leva* M edical B ranch, G alveston. TX 77550. USA. Received 3 M arch 1989; and in revised form 9 M ay I9XM U S population 1965 KQ Current smokers Tver smokers 44% 72% 53 16 < (io o | < 0 001 UiO l i l t N KIV L M il.A N l) JO U R N A L OK M K D IC IN K Nov. 18. 1W3 REVIEW ARTICLE MEDICAL PROGRESS__________J CANCER OK THE OVARY S tk pm p.n A . C a n n is t k a , M l ) . EPITHELIAL cancer >r the ovary is ilie HIih most common malignant Condi(ion among women in the United Slates, with an annual incidence of '22,1100 new cases.' litis disease predominantly alire is post menopausal women in their sixth decade, accounting lor approximately 13,.100 deaths each year and tor over half of all deaths from genital cancer. The highly lethal nature of this tumor is related to the absence of symptoms in the majority o f women with early stages of the disease Seventy jierreiil of women present with advanced disease in which the tumor has spread to the peritoneal surfaces of the upper abdomen Extensive iniraabdoininal disease is difficult to eradicate com pletely by surgery, and many patients have only a partial response to postoperative chemotherapy. This review will discuss selected issues concerning the risk factors for and diagnosis of ovarian caneer, followed by a detailed discussion of treatment options lor pa tients with newly diagnosed disease and those with persistent disease. R is k F a c t io u s f o r O v a r ia n C a n c c m The average lifetime risk for the development of ovarian cancer in women in the United States is 1 in 70 Women with uninterrupted ovulation appear to be at higher risk lor malignant transformation o f the ovarian surface epithelium. For instance, a higher risk of ovarian cancer is associated with nulliparity or a first birth alter the age o f 35.2Conversely, a lowei risk is associated with childbirth, especially at age 25 or younger, or with the use of oral contraceptives.v How ever, a family history of ovarian cancer, especially if two or more first-degree relatives have been affected, is the most important risk factor.** At least two welldescribed familial ovarian cancer syndromes exist, known as the hereditary breast-ovarian cancer syn drome and the Lynch syndrome II.4 The hereditary breast-ovarian cancer syndrome is characterized by a familial predisposition to breast and ovarian cancer, with an autosomal dominant pattern o f inheritance derived from either maternal or paternal genes The f .ynch syndrome 11 is characterized by a similar mode of inheritance and a familial predisposition to ovarian Fr\*ii ihe D ivixun if M u tu a l Oncok^y. Dana Kurbcr Cancel liu liiu lc. ami IIarvan) Medical Sihoul. Boaion AJJrcu repnm icquc in Or CannikUa mi (he Dana Fait) Cancer Inxitvic. ** Binncy Si . M A 021 IS. tam er, endometrial cancer, and colon cancer without polyposis. Familial ovarian cancer accounts lor less than 5 percent of .II ovarian cancers and lends to occur m women at a youngrr age than its sporadic counterpart. Although the res|>oiisible gene (or genes) has not yet been identified, linkage analysis suggrsts that a candidate gene referred to as HRCA-! is located <ii the long arm of chromosome I7 at site I7ql2-q23.J For women with a strong family history of ovar ian earner, management typically includes frequent screening and consideration o f prophylactic oophorec tomy after age 35 and when childhearing is complete.4 It is not yet known whether currently available screen ing methods will permit the efficient detection of dis ease at an early siage or improve surv ival. A screening method that uses the UA-125 serum tumor marker (see below), followed by confirmatory pelvic ultraso nography lor women with elevated levels, lias a speci ficity of 99.9 peiccnt in postmenopausal women, al though the sensitivity of this eomhiued approach is still only approximately 50 percent.oVFurther investi gation of such screening strategies will be required before firm recommendations can be made regarding their general use. C l . l N I C A i . P m IlS K N T A T I O N The most common symptoms of ovarian cancer arc related to extension of the disease outside the pelvis. Many women present with abdominal fullness and early satiety due to ascites and omental tumor im plants (Fig. I). Occasionally, a woman will undergo an extensive gastrointestinal evaluation for presumed hiatal hernia or gallbladder disease before the true cause o f these nonspecific symptoms is identified. Ab dominal examination may reveal ascites, upper ab dominal fullness from omental involvement, umbilical hernia due to increased abdominal pressure, or um bilical lymph-node involvement (Sister Mary Joseph's node). A minority of patients (30 percent) with dis ease still conlined to the pelvis present with occasional pelvic pain due to intermittent ovarian torsion, but most patients with limited disease are asymptomatic and are identified by routine pelvic examination. Pel vic examination may reveal an ovarian mass in a pre menopausal woman or a palpable ovary in a postmen opausal woman. Although ovarian canc er is a disease (hat initially spreads throughout the alidominal cav ity, examination may sometimes reveal a pleural effu sion or involvement of the axillary or inguinal lymph ncKfes."1In addition, patients with abdominal disease frequently have asymptomatic involvement of the pel vic and retroperitoneal nodes, usually documented at the time of surgery." Patients with ovarian cancer may occasionallwprcsent with various types of para neoplastic coudiJlryis, such as humorally mediated hy percalcemia (clear-cell histologic feature's),,7n cereIx'llar degeneration (associated with antibodies to ft Vul IV N .. VI M K J M C M r i( ( :i'i SS < ' N N f . lS I M A IVl dries not g u a r a n te e ih r p r e s e n c e >,l ovarian c a m rr, and a norm al (i.V I'5level d oes nui p reclud e ih r n m l for u i^ n y il ih r p e lv ic e x .im iiia iiou and ultrasound study s u r r s t cancer. S i A f.lN i. AND S llM tiiC A I. TKfcATMfc.NI A n ex p lo ra to ry la p a m tn m y is (lien necessary in d iagn ose nv.ui.in cancer, since ovarian m asses may also he caused by henigu cysts; by other prim ary ovai luu Illin ois, such as germ -cell nr strom al-cell neo p la sm s; o r hy m e ta s ta sis to tin- ov a ry Iroin a prim ary g a siiu m icsiiii.il or hreasl i.lli< rf.7,i,,T lie path ologi cal distinction betw een epithelial Figure 1 Intraoperative Appearance ol Stage III Ovarian Cancer Involving the Upper Abdomen. The exploratory laparotomy is performed through a vertical, mkJlme incision to permit adequate visualization ol the upper abdomen. Several serosal implanis are shown in a patient with stage 111disease (Photograph courtesy ot Dr. Jonathan Nilofl, Department ol Gynecologic Oncology, Beth Israel Hospital, Boston.) o v a r ia n c a n c e r a n d th e se o th e r | m s sib ililies is u su ally straight loi w aid ; the com m on histologic features aie s h o w n hi T a b le I . O nce epithelial o v a iia ii cam ei h a s h ern c o n fir m ed , a to ta l iiImIoiii- ilia! h y sterectom y, b ilateral sal- Purkinje's c e l l s ) , t h e sudden appearance of sclw>r- pi n g o -o o p h o r c c to i u y , a n d n iiic n le c io m v .tie u sn a ll) rheic keratosis (sign of Leser-Trelat),,b or chronic in perform ed , a lo n g w ith a careful e x a m in a tio n ol all travascular coagulation leading to arterial and venous serosal su rfa ces, b io p sie s of g io ss ly in v o lv e d a ic a s , thrombi (Trousseau's syndrome), (n rare cases, pa a n d c o lle c tio n ol a sc ite s or p e r ito n e a l w a sh in g s loi tients have presented with brain metastasis as the first c y to lo g ic stu d ie s, i f ih e d ise a se a p p e a r s to he lim sign of disease.17 ited to (hr ovary, the letropetitoiiraf nodes are also Once a pelvic mass has been detected, confirmatory e x a m in e d , sin ce e v id e n c e of ly m p h a tic in v o lv e m e n t studies include a pelvic ultrasound examination and a alters th e p o sto p e ra tiv e iie a liu e n l a p p r o a c h (se e hr CA-125 determination. Pelvic ultrasonography, espe lo w ). O c c u r r e n c e ol iu lia o p e ia iiv e in p lm * ol the cially when performed Iransvaginally,11'1* is a more p rim a r y -o v a ria n m a ss is a lso n o te d l'`in ally. .m a t sensitive technique than computed tomography (CT) tem p t is m a d e to d clm lk a ll g r o ss d isea se , a n d the for imaging the ovary, and it provides information that a m o u n t o l resid u al p e iiio n e a l o r sero sa l im p la n is ir - is predictive of the presence o f cancer. Specifically, m a in in g at d ie c o m p le tio n o f su rgery is n oted T h e benign cysts usually appear as single, simple struc s ta g in g sy ste m d e fin e d hy the In tern ation al K rd ci- tures without internal echoes, whereas an ovarian tu a iio n o f (iy n e t( lo g ic O n c o lo g ists (se e T a b le 7) a s mor may appear as a niullicyslic complex mass. The su m e s that an a d e q u a te s ta g in g o p e r a tio n h a s b een CA-125 antigen is a serum tumor marker that is ele perfi nurd. vated in 80 percent of patients with advanced ovarian A dverse prognostic lactors in ovarian can cer in c a n c e r .C o n s e q u e n t ly , this tumor marker often c lu d e an a d v a n c e d sta g e ol d isea se , a h igh iin n o i provides confirmatory evidence of ovarian cancer in a g ra d e, th e p r e sen ce o f residual d ise a se a lte r in itia l s u r woman with a pelvic mass and suspicious findings on g e r y , o ld e r a g e , c lea r-cell or m u c in o u s h is to lo g ic Jra- ultrasound evaluation. In one study of 158 patients tu rcs, o v e r ex p re ssio n o f die ep id ci m a l-g r o w ili la cto r with pelvic masses, a CA-125 level above 65 U per milliliter was highly suggestive of cancer, with a jKsilive predictive value of 98 percent in postmenopausal Table 1 Common Histologic Fealuies and Clinical Correlates ol Epithelial Ovarian Cancer women.*1However, the CA-125 antigen is not a specif ic marker for ovarian cancer. Elevated levels o f the antigen may also be associated with several other tu mors, such as breast, lung, and gastrointestinal can cers; with norinialignant conditions, such as benign cysts and endometriosis; and with menstruation and pregnancy. In addition, the results of this test are nor mal in up to 50 percent of patients with early-slage ovarian cancer.,v-30 Thus, an elevated CA-125 level The nwM oMiiiffl y|c I i.vanan cancel OcCai.MiaUy ncialeaJ w ill) ja im a jy c m S .iK liia l i ik c i im ciMi.MTKinio.il. bU aicial pee veutai m i in u p in III pci cent oi cue. O iieu u w iia lC il w ill) uuniial m unljf m inim ally clcvairO C'A 125 level Slajtc I.M lge, lhe w ia a |M.>kimk>u . t *aiiv c le U liv c iiK ru iiiv iiy hi itK iiuM lierap y. a\w>cialcil will ciMh> in c in im and hu iin x a lly mediated hypercalcem ia -- -,-------- - I5.V2 I'llK NfcVV KNtil.AiNli JOUKN.M. Oh MKDK'.INh Table 2. FlGO Staging System lor Epithelial Cancer ot the Ovary.* Su< I: lunxwliimicdu>it* jiio IA One uv a iy . r>> awn. in i t , i ^ n k IBBiahvine*, noikik>. mai capMik 1CRuplurxdiapuk. va|ulai utiilvcinriM. hmjt. .a nialijtitaMskiIo pcnumcal Sue ll; ovananlumo* withpelvn IIAPcWiv oicmiun lo uicr\j>.* iuhe> 118Pd>k' eiinuiiAn>otherpelvicfan. iMaUrr, iccium. vajiiiul IICPcIvk' eaicruion. plunrvjiiiji indicaicd Iih1C Maft III. lumoruuimk the pelvi*. with pukiiivr ivOet IIIAMicrokcopic tccJing iMivnie the true |li DIBCroiv Jepxkiit .' cm. ItICGioaJcpmHt >2 cm.t puaiuvc ncvlo Slat* tV: Jixant <f*ninvoheincnl, includinglivci paicrkhyiiuot pkural pace 'F K H IiIi k w i Iiwci'uivwuI N i k i a t * <>.A .ji I W . C ^ ku receptor, and amplification ol the c-iieu proto-oitcogene J< r> Paticnls with stage 1(1 disrasr and residual peritoneal or serosal implants under U 5 cm in diame ter have a median survival of 40 months, and those with residual implants ol 0.5 to 2 ent have a median survival of lb months. Patients with bulky residual tumors lhai are over 2 cm in diameter have a median survival in the ran^r of 6 to 12 months, and overall long-term survival is less than 10 percent.*111 The strong correlation between ihe amount of residual dis ease and survival in patients with stage 111 disease has been use d as dir main justification for delmlkiug, al though it is possible dial this correlation simply repre sents a selection phenomenon whereby patients with more indolent oi clieiiiosriisiiive disease are ideniilied on tile basis o f (heir ability lo undergo successful dr.bulking, which is partly a function of the volume and location of the disease. ,J Not surprisingly, (he ability to debulk tumors in patients with stage IV disease is not necessarily associated with prolonged survival.11 although selected patients may still iieuelit from the relief of imminent bowel obstruction before chemo therapy is started. Postoperative T reatment Combination chemotherapy that includes a plati num analogue -- either ei.splaiin or carlx>plntiu -- is (he standard postoperative dieiapy lor patients wiili residual disease or those at high risk On recurrent dis ease alter initial surgery. Although alkylating agents such as melphalan induce responses in al>ou( 30 per cent of patients wiih advanced disease," platinumbased chemotherapy has resulted in a substantially improved response rate of approximately ((I percent and prolongation of median survival by alxmt six months.1' 16 LImiltd OImam (Stag* Iand II) Patients with stage IA or lli tumors (Table 2) lli.it are well or nuxleraiely well diflereuliatrd have a liveyear disease-free survival of 91 to i)H percent, which is not improved by adjuvant (iraim ent.17 In selected pa tients with stage IA tumors dial arc well or moderate ly well differentiated who wish lo remain fertile, a unilateral salpingo-nophorecioiiiy may be performed, without evidence of compromised survival." Patients with either stage 10 (e g., capsular involvement or positive washings), ptrorly differentiated stage I, or any stage II disease have a higher risk of relapse, and these patients arc often considered for postoperative adjuvant tlierapy.:,,,v In a trial by the Ovarian Cancer Study Croup, such higher-risk patients were random ly assigned to receive cither 12 months o f oral ntelphalau or one dose of intraperitoneal chromic phosphate MP, resulting in equivalent 5-year disease-free survival rates in the HO percent range.*7 However, prolonged melphalan treatment is associated with an increased risk o f leukemia, and the distribution of intraperitone al WP may not be uniform in patients who have under gone surgery and in whom intraabdominal adhesions subsequently develop, in a current study by the Cynecologic Oncology Croup, such patients arc randomly assigned to receive either one dose of intraperitoneal **P or three cycles of intravenous cyclophosphamide and cispiatin. Patients with stage 1C, poorly dillcrentiated stage I, or any stage II ovarian cancer who are not enrolled in clinical studies are often treated with adjuvant platinum-based chemotherapy as described below. A dvanced Dlaeaaa (Stagaa III and IV) 'The majority of women with ovarian cancer present with stage 111 or IV disease. The treatment of ad vanced disease with platinum-based chemotherapy has resulted in an improved response rate and longer median survival, but the overall survival rate is still only al>oui 20 percent. Since this is a heterogeneous group o f patients, overall survival is a function of both the stage and the amount of residual disease after ini tial dehulking. Patients with stage III disease who have undergone optional dehulking ( < l to 2 cm of residual tumor) have a four-year survival rate o f ap proximately 30 percent*"; those with stage IV disease or stage 111 disease and sulxiptimal dehulking (> 2 cm of residual tumor) have less than a 10 percent chance o f long-term survival.*1*? Although many types of platinum-based combina tion chemotherapy have been investigated lor the treatment o f advanced ovarian cancer, the regimen of cispiatin plus cyclophosphamide is just as effec tive as more complicated and toxic regimens, such as those containing altrctamine or doxorubicin hydrocliloridc. *"* **However, the loxit effects of cispiatin, including renal insufficiency, peripheral neuropathy, ototoxicity, and nausea, have prompted the investiga tion of less toxic platinum analogues such as carboplatin In two reccuL randomized trials comparing cis piatin plus cyclopLffisphamide with rarboplatin plus cydophosphainidiAa,treat advanced ovarian cancer, the carboplatin regimen was better tolerated, with less nausea, less renal toxicity, and less neurotoxicity.** *5 /> ,fSIMHifritiliii<iiiiiii Vol 3W No. 21 M F.DIC.Al. l*K()(;kKS.S - C A N N IV I KA |V il Thr primary toxic effect o f cartxiplaiin is bom* mar row suppression, willi cumulaiive thrombocytopenia that occasionally interferes with the chemotherapy dosage. Nevertheless, the two regimens have been shown to he equally effective," and combination therapy consisting o f carlxplaiin plus cyclophospha mide lor six cycles (which can he given on an outpa tient basis) is now a widely accepted regimen for the treatment o f advanced-stage ovarian cancer. Continu ing chemotherapy beyond six cycles does not extend survival.41 The majority of patients with advanced-stage ovar ian cancer have a complete clinical rcs|xmse (as de fined by a normal physical examination, CT scan, and CA-I2S level) to siirgic.il dclmllnng and postop erative chcmmlui apy (Tig. 2). Ilowevi-r, olii) Ml | m*i rent of these patients have a compiete pathological response (as defined hy negative biopsv speeiincns obtained during a s<voml-l<ok laparotomy).41' The high false negative rale of cium al evaluation jcl.iic.s to the fact that jwisistrnt disease is usually mam test ed hy either small-volume gross de|xsiis or micro scopic disease, both being below the level of detec tion hy C T scanning or CA-125 testing. Tims, the second-hxik laparotomy is the most sensitive way ol detecting persistent disease alter platinum chemo therapy, although there are currently no Mamlaid second-line salvage regimens that .tie curative lm this Figure 2 Representative Outcome In 100 Patients with Advanced-Stage Ovarian Cancer. Exploratory laparotomy with debulking is appropriate for patients with stage III dis e a se Debulking should be used selectively in stage IV d isease for specific indications such as imminent bowel obstruction. The most common regimen is standard-dose cyclophosphamide with carboplatin or cus pidiin. Taxoi pius cisplatin is a more active regimen for selected patients with bulky re sidual disease, although the effect of this treatment on overall survival is not yet known. Several Investigational regimens containing taxoi, high-dose carboplatin. or both are also available for patients with ap propriate 6tage III or IV disease. A com plete clinical response is defined a s a nor mal physical examination. CT scan, end CA-125 level Second-look aurgery ia per formed if the finding of residual disease would make the patient eligible for innova tive second-tine treatment programs. The relapse rate tor patients with a complete pathological response is about 50 percent. Persistent clinical d isea se is usuelly manifested by an elevated CA-125 level or an abnormal physical examination or CT scan. Symptomatic residual d isease is usually treated with second-line chemo therapy regimens such as laxol Patients are occasionally considered lor secondary debulking. after which they may be eligible tor investigational intraperitoneal therapy, if appropriate. Persistent, minimal disease at second-look surgery may be treated with second-line chemotherapy regimens, intrapentoneal chemotherapy, or autologous transplantation. A rapid recurrence (within six months) 6 often treated with agents that are potentially not cross-resistant to platinum, such a s taxoi. A late recurrence (after six months) may respond to addition al standard d oses of platinum-based ch e motherapy. Like patients with persistent disease, those with recurrent disease are occasionally considered for secondary de bulking, followed by investigational Intra peritoneal chemotherapy. Appropriate pa tients who have an excellent response to second-line chemotherapy may also be considered for autologous transplantation. Advanced Disease (stage III or IV) (n = 100) i Surgery for Diagnosis and Debulking, if Appropriate I (Alternative Option) Standard Chemotherapy Platinum Dose intensification or Taxol-Containing Regimens Complete Clinical Response (n = 80) . Persistent Disease (n - 20) I Second-Look Surgery, if Indicated for Investigational Purposes Second-Line Systemic Chemotherapy Second-Line Systemic Chemotherapy Intraperitoneal Therapy Autologous Transplantation T i l t NKVV KiMil.AND.IOURNAl. OK MKDICINK Nnv 18. 1993 pal iffit group. In addition, patinas who do have a complete pathological response as determined by second-look surgery still have a .SO percent chance of relapse." ** Thus, even though laparotomy is more sensitive than clinical evaluation, the detection o f ac tive disease has little cllrci on survival, and thr surgi cal procedure has a high lalsr negative rate. ( )n (hr basis o f these considerations, second-look laparotomy is no longer j>erformcd routinely, unless thr detection of disease would make the patient eligible ioi inno vative treatment approaches such as those outlined below. M e i m o o s t o O b t a in a C o m p l e t e P a t h o l o g ic a l R e s p o n s e in A d v a n c e d O v a r ia n C a n c e r 1he inability to obtain a complete pathological re sponse with standard doses ol rarlxiplaiiti and cyclo phosphamide is the chief obstacle to cure in advauerd ovarian ranerr (Kig. 2)- Strategies to increase the late ol complete pathological responses include the use of new drugs with unique mechanisms ol action, such as taxol, ami dose intensification to improve the activity ol platinum analogues. Taxol is a ditcipeue plant product isolated from the haik ol the Pacilic yew trrr, Tutus brevijoha. It exerts cytotoxic elicits through a unique tnrcliaiiism of action involving mi crotubule pjlyuirri/.ation, rendering the mitotic spin dle dysfunctional and resulting in an arrest between (he second gap ((,) phase and the initotie (M) phase of the cell cycle. As discussed below, this drug was originally noted to he eliective in platinumrefractory ovarian cancer, suggesting that taxol and eisplatin do not share common mechanisms ol drug resistance. In a recent trial by (hr Gynecologic Oncol ogy (roup, the two drugs were combined to treat pa tients with newly diagnosed, advanced disease." A total of :m patients with residual tumors over I cm in diameter were randomly assigned to receive either cyclophosphamide plus rispl.itm or taxol plus isplatiti rvrry three weeks for a total o f six cycles. Al though the patients receiving taxol hail a greater degree of neutropenia, the incidence of documented episodes of sepsis was similar lor the two groups, and the taxol-cisplaiin regimen was griirrallv well loieratrd. The overall response rale of the taxol-eisplatin group was signiiicanlly higher than that o f the cyclophosphamicle-cisplatin group (79 percent vs 03 per cent, PCO.OI), and the median periods o f survival without progression o f disease were 17.9 and 13.8 months, respectively. The rates o f complete pathologi cal responses for thr taxol-eisplatin group and lire ryclophosphamide-risplaiin group were not signifi cantly dilVercnt (20 percent vs. 19 percent, respective ly; P - 0.07), suggesting that (he increased response rate with taxol may not necessarily result in improved long-term survival. A longer period o f Ihllow-iip will he required to address this issue. Nevertheless, this study has shown that taxol combined with eisplatin is a highly active regimen that represents a reasonable alternative to standard chemotherapy lor selected pa tients with newly diagnosed, advanced disease. Dose intensilu atiou is another strategy for improv ing the response rate in ovarian cancer. The theory underlying a higher platinum dose is that many tu mors exhibit a strep dose-respinse curve to this drug in vitro, meaning that a relatively small increase in the dose may result in a large increase in the number of tumor cells killed. Support for this concept comes from a retrospective study by larvili and Hryniuk demon strating that a higher platinum dose is associated with a superior outcome in patients with ovarian cancer.'1 A recent randomized study compared two doses of eisplatin -- 50 versus 1lM> mg per square meter of body-surliicc area -- administered with cyclophos phamide every three weeks lor a total of six cycles." The median survival was significantly improved in the higher-dose group, with die elicci being most promi nent in patients with residual tumors under 2 cm in diameter alter initial surgery (median survival, three years). Since this study used doses of eisplatin that are still within the standard range, the results do not ad dress the value of using even higher chemotherapy doses to neat advanced ovarian canter. Regimens of systemic high-dose eisplatin or earliopl.ifm may in duce additional respinse* in 20 to 30 percent of pa tients with residual disease after standard doses of platinum chemotherapy (usually in those with a previ ous rrspmsr to standard-dose platinum), providing further support lor the concept of a steep doseresponse curve." " Among newly diagnosed patients, however, con vincing improvements in the rate of complete patho logical respmses or survival have not yet been ob served with systemic dose intensification of platinum analogues." SKVi A sjodrrl! et al. suggest, this lack of improvement may reflect a plateau in the doserespmse curve, so that higher platinum doses may not necessarily yield improved respinse rates.1,0 in addi tion, some patients an* unable io tolerate sequential cycles o f high-dose eisplatin because of nonlienialologic toxicity, .such as nephrotoxicity and neuropathy, which may he severe at doses as high as 200 mg per square meter of body-surface area per cycle. Likewise, many patients are unable to tolerate sequential cycles of high-dose carboplatin because of cumulative bone marrow toxicity, usually in the lorin of persistent thrombocytopenia, which is often unresponsive to treatment with hematopiieiic growth factor.*163 Re cently, several groups have begun to treat newly diag nosed, advanced-stage ovarian cancer with high-dose carboplatin followed by an infusion of autologous periphrral-hlood progenitor cells in an attempt to in crease the dose without incurring excessive hone mar row toxicity.*4'1-*The effect of this strategy on the rate of completr pathological responses among patients with newly diagnosed, advanced disease will require further investigation. T r e a t m e n t C n riO N s k o r P e r s is t e n t o r R ^ ^ H R e n t D is e a s e Patients who have persistent or recurrent disease after initial chemotherapy are generally not curable. Vul. 329 No 21 MKDK-AI. PRtKIKKSS -- <IANNIS I KA Persistent disease is often heralded by a CA-I2.S level that docs not return to normal alter three cycles ol platinum treat men i,,`' or that remains elevated alter completion of chemotherapy.7"'7VSimilarly, a new ele vation in the CA-125 level is often the lirst sign of relapse and inay occur before a clinically apparent tumor (as detected by physical examination or C T scanning), with a median lead time of three months.71 However, since second-line treatment approaches are |xiW`ii(ially toxic and generally palliative in nature, the decision to treat patients with persistent or recurrent disease is often based on either the presence o f symp toms, such as abdominal discomfort resulting front a documented mass or ascites, or the eligibility of patients for innovative treatment approaches (see below). Patients with disease that persists after initial che motherapy or recurs within six months after treatment are unlikely to benefit from additional standard-dose platinum.73 In contrast, patients with disease that re curs more than six months after their initial treatment may have a good chance o f brnrliting from secondline cisplatin or carlxiplatin (27 to 5`) percent response rate),71 although such responses arc generally not as sociated with long-term survival. A description of treatment options for the management o f persistent or recurrent disease is presented below and summarized in Figure 2. Secondary Dafcuiking Surgary Patients with persistent or recurrent disease are sometimes treated with another debulking procedure in an attempt to improve (heir survival. Pro|K>neuts of this procedure have reported an improved median survival rate for patients in whom secondary debulk ing resulted in minimal residual disease, although most patients treated in this way have been carefully selected and arc nevertheless destined to die of their disease.O M70As with initial debulking, it is difficult to know whether the secondary debulking itself lias a therapeutic elTrct or even a durable palliative efleci, or whether the patients in whom the procedure is suc cessful are those who have more indolent disease. .Since secondary debulking is not curative and may be associated with considerable morbidity in patients who may otherwise be asymptomatic at the time, its routine use cannot be recommended. However, there are patients who may benefit from secondary debulking, including those with imminent bowel ob struction due to a single dominant pelvic mass and those who have a relapse many years after the initial diagnosis, in whom an exploratory laparotomy to ver ify the diagnosis and dehulk the tumor may lie fol lowed by several more years o f disease-free survi val. Finally, secondary debulking of gross disease may be appropriate for patients who arc eligible ibr irratment programs involving imrapcritoneal chemothera py, since it is known that this form of treatment induces res|onses only if there is minimal disease, as discussed below. In a uoninvcstigatioiial setting, howevrr, secondary debulking should have a limited role in the p alliative e a ie ol patients w u h ovai iau c a n cer, and its use should be individualized. Intraparllonaal Chemotherapy Inirap eriioucal ebeinolher.ipy eonsists of a d m in is tering di ngs such as cisplatin and etoposide through an indw elling catheter inserted into the peritoneal cavity. A lthough this procedure has been used lor sev eral years io treat patients with residual disease alter initial system ic platinum ih ei.ip y , its uupaei on sm vival in th ese patients is still unk now n o w in g in the lack o f ran d om ized (rials in ihis area u niil recently. A ppropriate drugs lor iiiiraperiione.il therapy air those that have a favorable ratio o f peritoneal io plas m a drug concent rat ion, those that are nouii liu tin g w h en in stilled , and those in w hich a sm all increm ent in thr d o se con centration m ay result in an ad d ition al cy totoxic rcs|Xuse (i.c ., drugs w ith a steep d o s e rrsp onse curve.). C isp latin and taxnl d isp lay such characteristics, and dose scheduh's have been esta b lished that provide a p liam iacologic ad vantage with a c c e p ta b le t o x ic ity ./7 Wl S in c e u itr a p e r iiu iir a l c h e m o therapy is a lim n o f topical treatm ent, drug diffusion in to tissu e is lim ited to the lu st several iiiilliiu e le is ol tum or at best, suggesting that bulky tum or m asses cannot be expected to respond (although icspou scs can still be im b u ed w ith sy.sieniically ab sotb ed d in g ) S everal p h a se I an d p h ase II stu d ie s u sin g p la tin u m based inlrapcriloiie.il regim ens have reportrd an ac cep tab le level o f toxicity and a com p lete surgical re s p o n s e ra te a s h ig h a s 2 0 to '.id p ercen t in p a tie n ts w ith residual sm all-volum e disease w ho have previously ha d a r e s|m u s e to s y s te m ic p la tin u m " 1,1 N e v e r th e less, the im portance ol these espouses, in term s ol both palliation and cure, rem ains uncertain, and m ost patients eventually die ol progressive disease. A d di tional stu d ies have show n that the patients w ho .in m ost likely m benefit from inti apei itoneal li raim ent arc those w ith m inim al Illinois ( < | to 2 cm ) who do not have diflusc carcinom atosis, evid en ce of progres sive disease during therapy w ith system ic platinum , extensive iiitraalxlom inal adhesions dial would p re v ent h o m o g e n e o u s d ru g d is tr ib u tio n , o r d is e a s e in inaccessib le sites such as the lym ph nixies or liv n par e n c h y m a .HJH` T h e s e restr ic tio n s su g g e st ilc.it im i.ip erilonral therapy m ay be appropriate for only a sm all subgroup of patients w ith persistent or rapidly irt H i ring disease. It is (Xissible dial a dosc-imensilii ation strategy such as inuaperitoneal therapy will hr most effective early in the treatment of thr disease, when a substan tial fraction ol tumor ceils air still cliriimsrnsilive There aie ai least two randomized nials undei wav that should drtriniiue the value, il any, ol inliaprntoncal eliriiiodicrapy as a means of dose uuensi fication in this selling. The <yui-cologic Ontology Group is conducting a randomized trial ol mu aperitonral versus intravenous cisplatin (with a stand ard dose o f intravenous cyclophosphamide in huili arms) in patients with newly diagnosed stage III dis ease that has been optimally debuikrd I lie f'.tinipr.ui IS!*> THK NtVV KN<.I-ANI) JOUR NA L OF MKDICINK Nov. IH. IW3 Organization for the Research and Treatment o f Can cer is conducting a randomized trial comparing intraperitoneal platinum-based therapy with no further treatment as a means o f improving the complete pathological response in patients with advanced-stage disease. Radiotherapy There is no convincing evidence that whole-aUiomen radiotherapy can improve survival in patients who have persistent residual disease alirr platinum c h e m o t h e r a p y .P a t ie n t s with macroscopic resid ual disease, csj>ecially if poorly dillerrntiatrd, clearly do not benefit from whole-aUlomen radiotherapy,*' and the toxic effects o f (his treatment include smallbowel obstruction, radiation enteritis, and myelosupprrssion. Methods to overcome mechanisms o f re sistance to radiotherapy after failure ol platinum therapy or to decrease toxicity through hyperfractionaiimt techniques are currently under investigation."" Although radiotherapy cannot be routinely recom mended as a salvage treatment in ovarian cancer, it occasionally provides palliation lor patients with re current pelvic disease who are not candidates for sys temic chemotherapy. $ytmic. Non-CroM-Raslaiant Chamotharapy Systemic treatment with agents that are not crossrcsistant to platinum represents (he mainstay o f treat ment lor patients with platinum-refractory disease, as defined above. Although several drugs exhibit activity in this patient population, response rates are low (10 to I.1) percent), and most responses arc partial and short-lived. Potentially useful drugs in treating platinum-refrac tory disease include altretamine,*1-9* mclphalan, ilbsfamide,91 tamoxifen,'*4 and taxol.95'97 As stated above, the most active agent in this category is laxol, which induces response rates of approximately 25 percent in patients with platinum-refractory disease. In con trast to most other chemotherapy agents, taxol may require at least three or four cycles of administra tion before a response is evident. The drug is generally administered intravenously over a period o f 24 hours on an inpatient basis, with drxaineihasonc, cimeiidine, and diphenhydramine prophylaxis to prevent the hypersensitivity reactions frequently observed in the past, thought to he due to either taxol itself or its Cremophor vehicle.9" The main toxic effect of taxol is neutropenia, often without suppression of red-cell or platelet counts, as well as cardiac-con duction disturbances and peripheral neuropathy. The dose of taxol can be escalated through the use of hematopoietic growth factor," although it is un clear whether higher doses o f taxol result in im proved rrs|>oiise rates for patients with platinumrefractory disease, 'laxol, which is now available commercially for the treatment of platinum-refractory ovarian canerr, has recently been shown to produce a substantial improvement in the overall response rate when used in conjunction with cisplatin to treat patients with newly diagnosed disease who have re sidual tumors over 1 cm in diameter.51 Autologous Bono Marrow Transplantation Techniques to administer very-high-dose, myeloablative chemotherapy lollowed by hone marrow rescue are now well established in many treatment centers. This strategy is potentially curative for patients with tumors such as lymphoma, in whom both chcmosensitivicy and a minimal volume of disease before trans plantation are prerequisites for obtaining an optimal response. "u,wl Most ablative regimens used in ovar ian-cancer transplantation programs include (Kitenlially active agents such as an alkylating agent (i.e., melphaluu, cyclophosphamide, or ifosfamide), a plati num analogue such as c.arlioplatiu, and other drugs such as eio|M>sidr. The benefits of transplantation in ovarian cancer are dilhculi to assess owing to both the relatively small numbers o f patients who have undergone this proce dure and selection hias."<7l<"` Transplantation has generally been reserved lor patients with persistent residual disease after standard chemotherapy, and resjxmse rates as high as 75 percent have been report ed IU> Some researchers have even rrjxjrtrd small numbers o f long-term survivors, although many of these patients received transplants for persistent mi croscopic disease, a group that may have had an indo lent clinical course regardless of the treatment.*6,IU7 However, much of the early transplantation experi ence involved patients who had received extensive pri or chemotherapy, resulting iri a level of drug resist ance (hat could not be surmounted by dose escalation alone. These patients tended to have either no re sponse or a short-lived partial response to transplanta tion, at the expense of considerable toxicity (due to a relatively small marrow reserve) Therefore, the most appropriate patients for transplantation are those who have had a clear response to initial platinum-based therapy hut nevertheless have persistent, nonbulky disease; who are otherwise young and healthy; and who understand (he risks of autologous transplanta tion (including a 5 to 10 percent chance of treatmentrelated death). S u m m a r y a n d F u t u r e D ir e c t io n s Treatment of ovarian cancer is both frustrating and encouraging. In view o f its propensity to remain in the abdominal cavity and its responsiveness to chemotherapy, ovarian cancer would appear to be a curable disease. In reality, however, by the time most patients are given a diagnosis, the tumor has progressed to a stage in its natural history at which it is partially resistant to drug therapy. It is pos sible that the recognition of familial cancer syndromes and the use of iinpuived methods of early detec tion will eventually Jifrniit diagnosis at an earlier, more limited stagrapjf the disease, when drug re sistance has not yet developed and the majority of patients can be cured by surgery and standarddose adjuvant chemolhera|>y. Voi. 329 No. 21 M K D 1C A I. PK fX iR H S S -- C A N N IS I KA i .r.*i ; For paiienfs with advanced disease, (he inahilily to achieve and maintain a complete pathological re sponse with platinum-based therapy is the chief ob stacle to cure. Ongoing efforts to improve the rate of complete pathological response include intensification of the platinum dose and the use o f new drugs with novel mechanisms o f action, such as taxol. Treatment of persistent residual or rapidly recurrent disease con tinues to be a major problem, but there is a promising role for non-cross-resistant drugs such as taxol, a pos sible role for intraperitoneal therapy, and an evolving role for autologous bone marrow transplantation. Sev eral other innovative approaches (hat may eventually benefit patients with advanced disease include rever sal o f platinum resistance with huthionine sulloximiue, intraperitoneal administration o f autologous lymphokine-activated killer cells,,,w or iniraperitonral administration o f tumor-directed aniilxKlies conjugat ed 10 toxins or radioisotopes.I,w,m Thus, although ovarian cancer can be a particularly difficult disease to treat, we now have several promis ing drugs and strategies that may eventually result in improved survival for patients with this disease. How ever, it is likely that cure will result only from a combi nation of treatment strategies tailored to the needs of individual patients. For example, it is possible that in the future patients with advanced-stage disease may best be treated hy surgery, followed by postoperative chemotherapy with taxol and high-dose carboplatin, with immediate consolidation o f an excellent response hy autologous hone marrow transplantation or by inira|>eri(oiieal administration o f antilxxly-toxin con jugates. The impact of these new treatment ap proaches on survival will be determined only through the continued enrollment o f patients in well-designed clinical trials. R kfcrences 1 Boring CC, Squires TS. Tong T Cancer statistics. 1993 CA Cancer J Clin 1993,43:726. 2 Negri E. F ranccidu S. Ttonou A. ci 1 Pooled analysis of 3 European CMC-control slut)ICS I. Reproductive (acton and risk of epithelial ovarian cancer Ini J Cancel 1991:49.30-6 3. SchilJkraur JM . Thompson W l) Familial ovarian cancel, a populationbased ca-control Study Am J Epidemiol 1988.128:456-66 4 Lynch HT. Lynch JF Hereditary ovarian carcinoma. Hcmntol Oncol Clin . North Am 1992:6 7S3-III (( 5. jjN arod SA, Feunicun i. Lynch HT. el aJ Familial hrem -ovarian cancer ' -'lo cu s on chromosome I7ql2-q23 la n c e t 1991,3.18:82-3 6 Jacobs I. Stabile I. Bndgcs l. et al. Multimodal approach to screening for ovarian cancer. Lancet 1988:1:268-71. 7. Jacobs I, Davies AP, Oram D Role o f CA-125 in acre*rung (or ovarian cancer In: Sharp t: . Mason W p, Creasman W. eds Ovarian cancer 2: biology, diagnosis and management. London: Chapman A Hall, 1992:26375. 8 Bourne TH. Whitehead M l. Campbell S. Royalon P, Bhnn V. Collins WP Ultraaound screening for fsuuliai ovanan cancer Gynecol Oncol IV9I;43: 92-7. 9. Schwartz PE. Chambers JT. Taylor KJ, cl al. Early detection of ovarian cancer, preliminary results of the Y ak Early Detection Program Yak J Biol Med 1991:64:373-82 10 M cG om gk KF. Dwdnruki MR Endometnoid .arcirsnm* of the ovary presenting with an enlarged inguinal lymph node without evidence o f ahdoininal carcinomatosis Gynecol Oncol l9 9 2 :4 j.225-8 11. burg hardl E, Girnidi P. lahuusen M. T am utsino K, S'lcltncr II Pailcm iol pelvic and pnranunic lymph node involvement in ovarian cancer. Gynecol Oncol 1991:40.103-6 12 Allan SG. Ixxkh an SP. IxonarJ K C . Smyth JF Paraneoplastic hyperval.x n w a in ovanan raremoni* H M ) I9H4 ?X8 1714 3 13 Nusahauio SK . ( ia / K l ) . Arnold A I lypcm lcciiua a m ln topic secretion ol paraihyioul boom aie by an ovanan c jic iii. miu with rearrangement ol i Ik gene fie paiaihynnd hormone N Engl J Med 1990.12 I 1324 8 14 Case Record ol ihc Miiwmhuseii General Hospital (Case 14 l'/H 9 i N Engl J Med 19X9. 21 324 15 15 Greenke JL. Illu d im i IIK Antibodies io icich rllai Kwikuig; .e lls in pa ncnts with param-opu>ii< in e b r ilo degeneraiion ami ovanan i aicmoitw Ann Neurol I9 K I.I4 N I I 16 Hulguin T . I'adiJIa NS. Ampucfo F Ovarian .. ...................... ... preventing wiih the sign ol le ver - retai. Gynecol Oncol l9 K n ./3 12* (J 17. IZ t|U ic n h iM A .fl|o la ll. P h llu irv r. A h m a rM C .M a iia v -G M u X Ovanan carvnaana preceded hy cerebral n e io ia c u review ol tiir liiciatuic t iyuc col Oncol 1992.45 2116 III 18 van Nagell JK Jr. Higgins K V, Ikniaklvnn l b . ci al I ransvajrmal v.Nmgia phy M screening meihod lor o vm m h cancer, a ir p a i ini lise In d linai cues screened t im e r 19911.63 371 7 19. Bast KC Jr. Feeney M . la /a n n I I . N.idler I.M . Colvnt k H . Knapp KC Heacov iiy o| a nwNHXlonal a n tib o d y with h u m a n o v a ria n i an moina I ( Ini Invest I9 K I.6 8 I i l l -7. 20. Bast KC Jr. Klug I I . . Si Ji4M F., cl al A radioniiiiiiimwssay living a irsnruxtonaJ MilNKly lo imunior ihc course ol cpniK-lml ovarian la m c i N Engl J Med 19*1.109 MKI 7 21. M alkatiaii G O Ir. K n ap ' KC. I avin P I', cl al J'lroper alive evaluation ol aeruni C A 123 kvcls in preinrmipauval a ilj p.n4im:iMi|iau>al pallenti willi pclvu' inavses J im.m musi mo benign lim n malignarli doe a x * Am J Obirel Oynocol IV R 8 .I5 9 .U I 6 22. Perni k . Picket If . Hcydalladai M . r i al Nongrmtal earners o KlaM alk So (be ovry Gynecol Oncol 1992:44 K l 6 23. H olU F . fla il W K Krokcnhcrg iimimns ol Hie ovaty. a ilim cnpaiholngii analysis o f 27 cases Cancer I9K2.3! 24 lK 47 24 Omura G A . Brady M F . Ilnmcvlcy I I I ) , ci al lo n g lenii follow up and prognostic lactor analyviv in advamed ovanan la fiiin N ii* He Gyrsei ohign Oncology Group esperiem.c J ( Ini O m u l I9 V I.V 11 <R Ml 23 Caoniura SA W hy slmwM the ilnnngcnic cell awav br ixognovin allv important in ovarian cancer' I f l u Oncol 1991.9 Wx /is. 26. Kndriguet G C . Ilc riliik k A . Whitaker K s. .VhlmviHun D . ( larke IVaiv.Ni 1.. Basi R i Ir I pukrmal growth Inclot icn-pon espressimi in inuiual ovanan epithelium and ovarian carnei II Relun.Ni*lip h riw rrn iiie p r .* c api essi.ni and resp>mx to C|Hikrmal growth la n > Ain I t ibvlcl i .y o n >! 1991:164 745 30 27. Bau KC Jr. Jacohv I. Bcrchuck A Maligiumt translm malum ol iivaiian cpillicliuni J Nail t micci Invi IW 2.K 4 33n H 21 Stanimi l>J. <n M |4 ii i . W . k m . L A . r i al Nim In-v .4 I I I K 2'nru pin lotmcogcnc in human breavl and ovarian tam e i V n - m r Pan').744 In/ 12 29. Hcicliuvk A . Kamel A . W kilakci K . ci al Ovi-ie-ijNi-wion ol I II K .'o rni rv associated with |nnn survival in advam c l cjmliclial ovarian i ai ci < am ci Ken 19*41.5*1 41 *7 91 H I. Hacker N F. (tcrek JS. la g a w 1.1). Niehcrg K K . Flavh.iii N M I'tumuy cytm d uciivc surgery fur cpnhclial ovarian carnei Ohsict Oyncm l I9K t. 61 411 20 31 Rcrek JS. lin k e r N F. legasse 1.0. Pmlr T . Kcvrikk H . Ntchcrg k k Second-look laparotomy ru stage I II rprllwlial ovariait cancer, clinical vai tahks associated w ith disease status Obslcl Gynecol I984.r*4 2H7 12. 32. Hunter N W . A lc i amici N I)E . SiNitlcr W |* Mela analysis of surgery in advanced ovanan carcinoma is inacniuini iykncdm.nvc smgciy an in dependent determinant o f prognosi'.' Am J Obwcf Gynrvol 19*12.160 3im II 33. (iuudman I IM . Itarlow H I.. .Sheets F.k. el at D ie role ol cyl.Ncdu.live surgery Ui Ihc management ut stage I V cprlhrItat ovaitan i arum mia lagne col iJm.nl 1992.46 .Vi7 71 34. Whan, ni r r . Ile i von J Surgery for ONItinun cpnhclial iu iik n s ol ilie ovai) Cancer l98l.4H.Snppl 382-9 35. Omura G . B ksvm gJA. FJu In liC K . et al. A raiuhniured inai o f i ) i Io|4 m>n pbajnrde and ik iu m ih tiin wnh <n wribuut ospfatin in advam cd ovai rari carcinoma Cancel JVH6..37:1 723 MI 36 Groppe InicrogMNtale Cooperativo Omologico ( imci.>k>gia kamk>Mucd coinpaotoo nf crsplalin with cyckiphoiplunindc'Cisplalin ami with c y ih i phus|ihamKkAkisorol>icirv'osplaiin in advam r.l ovarian t u k c i tornei 987.2:1319 37. Y.wng KC. Wallmr I.A . F.llciibcig SS. et al Adtuvani tlictapy in stage I and stage II epithelial ovarian Carnet ii-viilts ol two |Nospective mndoiii i/cd trials N Fjtgl J Med 19**1.422 11)21-7 38. M iya/akl 1 . liN la d a Y . (Mila M . Kalni T . Mwuuo K . Sakaktbaia K Pleservatnai of ovarian turn inni ami re|N.lm live altlny m |m ik ih > wtlli malignant ovarian tuimes (y m .o l (Nic.it ItH K .h l 179 41 39. Pive/ M S . Malictano J. Itakci IK . Ileinpim g K l. l ive yeat survival lor stage 1C (M stage I p a le 1 rp iilx lia l ivariail carnet Healed w illi nvplalin based chetili alici apy tiynccol IN k h ) 1992.46 13 7 Mi 40. ffutura GA lim ujy U N . Heiek JS. t o ily S. Uclgmhi t i . M .n tel K Kau d o u ii/rJ in .d o l . y. (..|4ii.pli<<in<ic pluvi ivpi.itm wnlr wiilaNit J u im iila cm in ovai u n u in m n iu i a f iynciologi. (in . .logy tiiim p Study J ( In. (In o li 1989.) 137 r.s I6SH T H K N h \v K N < ;|.A N | J MKNA1. f i t M Kl MCI NI. Nov. IH. IV.J3 41 lotnc K G . Om is MK. M y c n CE. ci al Ir>(t term rouit ni cisplatmcontaining combinai ion chemotherapy rrgimcn lu the treatment of ad vanccd ovanan carvmoma J Clin Oncol 19*6.4:1579-85. 42 Hainwoih )U . Crush W W . Bumcfl 1 5 . lune H W || |. W o lll SN. Cucco hA Advanced ovarun cancel long term ic m iIu o/ treatment with mienmv citplalm -haxd chemotherapy ni boel Ju/ao-xi Ana Intent Med iv n it.iiu . i i t 70 43 Ncij JP. ten bokkcl Huinmk W W >en ,1er Hurg M fcL. ci al Kaitdoiuucd mal companng i * u combination chentraherapy regimeru ( C IlA P 3 v CP) in advanced ovanan carcmonta I Clin Oncol 1987.3.1137-66 44 Albert U S. Grren S. Haitmgan t V , cl al Improved therapeutic m dci of cartopUnn plut cycU-phmpbannde vertu cisplatm plu cyclophospha mide final report by the South! OncuUigy Group uf a phase I II random wed mal m ugc 111 and IV ovanan cancer I C lin Oncol IW 2 .10.706-17. IKnalum. J Clin Oncol IW 2 .M I MJ3 | 43 Swenenon K . Jellrey J. Stuan .e t al Cuplaim cyclophosphamide vertu caitupUnn-cyclophosphamide in a J v.iK d l ovarun carvel: a randomued phase 111 tiudy ni the -Natamal Cancer Institute *>( Canada Clinical 'fn a lt Group J Clin Oncol IW 2 .il) 718 2b 46 Advanced Ovarun Cancer T lu lu t Croup i tiemotherapy m advanced ovanan cancer an overview ol randomued Ointe al iriait BMJ I W l;3 03 : 884 93 47 Ptver M S. l-clc SB. Marchrtti 0 | . Baker TH . Isukada 3'. l.innch l J The impact 1 aggressive dchulkmg surgery a n j cuplaim bated cbemntherapy on progression tree turvival in stage III and IV ovarian carcinoma J Clin Oncol lu g H A V H l^ 48 Haket T B . Q u ia H. Hcrvkm WJ. cl al Kandomired prospective m al of 3 vertu IH cycle o i tyclophotptumiJc. dosuruhuin and cuplaim m advanced ovaiian carcinome Gynecol Oncul IW 2 .4 5 .2 M -9 49 Brenner DP. Shalf M l. Jonc H W . Grosh W W . Grve PA. Bumcii 15 Ahdummopclvic computed Irara-graphy valuation m panent undergoing second kxik Upa/oiomy for ovanan care monta Ohticl Gynecol 1983.63 713-9 50 Stchman PB. CaJkttu A K . Was JL. Sm in l.K . Sutton GP. L h rlu n t t A companton if findmgt al second look Uparutomy with ptcopcrativ cornpuled liauograpfty mpanenu with ovanan cancer Gynecol Oncol 1988.29: 37 42 51 McGuire W p. Mokm WJ. Brady M p . ei al A p h ax I I I m al comparing citplaiiivcyiotan tP C l and ciip la im 'u to i |K T I in advanced ovanan cancer <AOC> Proc Am Soc Clin OncU IW 3 .1 2:235 abstract 52. I w m L . Hryniuk W M Dote micnnty analym o f chemotherapy regiment in ovanan carcinoma J Clin Oncol 1917:5.756-67 53 Kaye SB. le w is C K . Paul J. el I. Mandomivcd tludy of two dote of cuplaim with cyclophutphimide in epithelial ovarian cancer lancet IW 2.340.329-3J 54 Ifamtwonh JL). Bumeti LS. June H W JJI. Groh W W , Johnton O H . Greco PA Htgh-dotc cuplaim combination chemotherapy in the ircatmeni (4 advanced epithelial ovanan carcinoma J Clin Oncol IV9.8.3026 55 O w l i RP. OtKhcga Y , Cun G . Young RC High d o x carbopiatm in refractory ovanan cancel patnt J C lin (htcol 1987.5 197 201 56 tond B. Hansen M . Hanten OP. Hanaen HH llig h -d u x plaimuni con tienne of combined cartwpUlin and civpUtm in prcvioutly untreated ivm un cancer panent wuh residual d ix a x J C lin Oncol 1989.7:146971 37 Mem Combined high dote carbopiatm and cuplaim . and iluiamidc m previously untreated Ovanan cancer patient with rctidual d ix a x J Clin Oncol 1990.8 1226-30 58 Rothenberg M L . Urol RP. Cilai>iein k . Sicinbcrg SM . Reed k . Young RC Dote intcruivc mduenon therapy with cyclophosphamide. cuplaim. and contolidalivc abdominal radiation m advanced stage cpiihclial ovarian cancer J Clin Oncol I W , 10.727 34 3v Colombo N . Pniclli M R . Panna G . M artola M . In m W . Mangiom C Copiai ir> (P* d o x ni(cnity m advaikcd ovarian cancer lA t X 'l a ran doinued ludy of convenlu mal d o x ( IX 't v d u x - intensive tL)l) enplaim monochemotherapy Proc Am Si C lin Oncul IW 3 .I2 .2 3 5 bttracl Ml Jodrcll D I. Fgonn M J. Canctla R M . ci al Rclalionthip betweencarbopla lin cipusurc and tumor rtsponx and toaiciiy in pattern with ovanan cancer J Clin Oncol lW 2:IO:520-8 61 Pcylan-Ramu N . Deur.seh V . Heshing Y . Caianc R. kUJor A. The effect of carbuplatm (CPi on megakaryocyte (M K ) progenitor and platelci produc no Proc Am Sic G in Oncol IW 2 .1 1.124 abstract 62 ten Bokkcl Humink W W . van WannenJani L. flclmerhict T . cl *1 Mammally high d o x chemotherapy (or high in k ovanan cancer Plue Am Soc Clin Oncol I W 2 .I I I.H) ablract 63 O 'D u ye t PJ. 1-aCrrU FP. Schrtdci R. ci al Ph ax I m al o l duotepa m combinait. with recombinant human granulocyte macrophage colony tumulaimg (actor J G in Oncol I W 2.IO . 1352 K 64 Gianni A M . Siena S. Hrcgm M . cl al Granulocyte maemphage colony tnmulntmg (ai tor to harveo circulating haemopoietn item e d it for auto transplantation Lancet 1989.2:580-3 65 t h a t A D . Ayach L . Aude n o n K C . cl al. M o b ilisa tio n o f pcnphcral blood prgcniim cell by chemotherapy and granuUxytc-m acrophage colony stim ulating fa ctor fo r hem atologic support aller hig h-d o x intcm ification fo r break! cancer Blood I W 2.7 V .3036-44. 66. She T C . M a x * JR. Storm olo A M . et al Sequential cycle o f high do x cartuiplaiin adm im tlered w ith iccoinhm ani human granulocyie-macio- phage c o lo n y-tnm ulating iK lo r and repealed infu tion o f autologous pc- npbcral blood progenitor cell: a novel and cflc v u v c m ethod fu r delivering n iu h tp k c u u rx * u f d o x -in te n iiv e therapy J C lm O ncol IW 2 .10.464-73 67. Teplcr k .C annutra S A . Frei E 111. c l al U x o l peiipberat-blood progenitor cell abrugate the m yckRoaiciry o f repetitive outpane ni high d o x c a r t plattn and cydophoaphantidc chemotherapy J C lin Oncul IW 3 :ll.T 5 8 J - 91 68. Fcnnclly I) . C row n I . Hake T . et al Accelerated delivery o( m ultiple c u u r x t o f h ig h -d u x (H D ) chemotherapy ( C f a new role fo r peripheral blocal progenitor cells (PBP ) Proc A m SoC C lin Oncol I9 9 3 ;I2 :2 6 0 ab- *4ta il 69 1a vin PT. KiU |*p R C . Malkaian G . W hitney C W . Bcrck JC. Bast RC i t C A 123 lo r the n>onih^tng o f ovanan carcinoma du rin g prim ary therapy Ohtet Gynecol 1987.69 223 7 70 N iJofi J M . K napp K C . I Jivin PT. cl al The C A 123 u t a y a a predictor ol clin ic a l recurrence in ep ithe lial ovarian cancer Am J 0 6 lei Gynecol 1986.153:56-60. 7| Nik>(( JM . B a il RC Jr. Schaci/I k M . Knapp R C . Prcdiclive value of C A 125 antigen level in xco n d -lo o k procedure fo r ovarian cancer A m J O te lei (iy n e co f 1985.151 981-6 72 Berek JS. Knapyi K C . M a lk a tu n G f) . et al C A 125 x r u m level currelaicd w ith x c o n d look operation among ovarian cancer patient O tm c l Gyne col 1986:67 685 9 73 Martiman M . Rothman R. Hake T . et al Second line platinum therapy in pa ticnii with ovanan cancer p rcvioucly ircatcd w ith cuplaim J G in Oncol 1991.9 389 93- 74 Berek JS. H acker N F. L a g a a x L D . Nteherg K K . Elahof( K M Survival o f p n irtn u fo llo w in g Ccorsdary cytorcducuve urycry in uvanan cancel 0 6 - ici Gynecol 1983.61 189 93. 75 Schw anr P t . Sm ith JP Second-lo* operation in ovanan cancer Am i O b i let Gynecol 1980.138.1124 30 76 Hoki>u W J. Rubin SC. lArlaney E. et al Influence o f ccim dary cytore duct am at the lim e o f ic w n d -lo o k lapatotom y on ihe urvival o f pal tend w ith epithelial ovanan carcinom a Gynecol Oncol 1989.34.365-71 77 H ow ell SB, Z im m S. M arkman M . ci i l Irm g term urvival u f advanced refractory ovanan carcinom a pa ixn i w ith m ail-volum e d i x a x treated w ith m iraperilim eal chemotherapy J C lin Oncol 1987.5 1607 12 78 Franc P. R ow tn thy F, Hake T . et al Pfcax I u ta l o f w eekly m irapcrno ncal tlP ) Laiol in patients (pis) w ith residual ovanan carcinoma (O C j: a OO G study Proc A m Soc C lin Oncol IW 3 ;I2 '2 5 7 abstract 79. Kirm a m S. Lucas W E . K u n S. et al A phase II tn al o f intrapcnioncal cisplatm and etuposide as salvage treatm ent (or m in im a l residual ovanan care ira w u J C lin Oncol I W I.9 6 4 9 57 80 Piver M S . L^ lc SB. M archetti 0 1 .. Baker T K . F nirich | J . Hartm an AB Surgically Jocuiucnted resp on x to iniraperitooeal c u p la im . cylarabinc. and bleom ycin alter intravenous ciplatm-haeJ chcuaMherapy in advanced ovarian adenocarcinoma J C lin Oncol 1988,6 1679-M 81 Rcichman B . M arkm an M . Hake T. et al. Intraperitoneal cuptatm and ctoposKte m ihe treatm ent o f refraiiury'recurT cnt ovanan carcinoma. J C lin Oncol I98V.7 1327-32 82. M arkman M . Berek JS. Blessing 1A. M c<iotre W P. B e ll J. Home icy H D C harBCicnsliccof pa tx n is w ith n u ll volume residuai ovanan cancer unre sponsive to cisplatm based ip chemotherapy: lesson learned from a G yne cologic O ncology G rou p p h a x II trial o f ip cisplatm and recombinant alpha-interferon Gynecol Oncol IW 2.4 5 3 8 83. M v k o ia n M . Rcichman B. Hake T . ct at Response to second-IttK eu- platin h a x d tiHnipentoncal therapy in ovarian cancer influence u f prior reponx to iiitraveiuiu cuptatm . J C lm Oncol 1991.9 1801-3 84 Peters W A III . Biasko JC. Bagley C M Ji. Rudolph R H . Sm ith M R . R jvkin SE. Salvage therapy w ith w hole -abdominal ina dia iton in parient with advanced carcinoma o f the ovary previously treated by com bination che motherapy. Cancel 1986;58:880-2. 83 Fuk Z . R ire l S. Biran S. Chemotherapeutic and tu rg tca j indueIkmi of pathological complete rciiHssnm and w hole abdom inal irradiation fo r con- m laJatioa d ie s not enhance the cure o f stage III ovarian carcinom a J G in Oncol 1988.6:509 16. 86 Copeland U . Gershenson D M . W nan on JT. et al M icnucupK . d i x a x at second look laparutotny in advanced ovanan cancer. Cancer 1983:55.472- 8 87 Schray M F , M a n tte r A . Ilow e s A F . ei al Advanced epithelial ovanan cancer: salvage whole abdom inal irradiation lo r patient w ith recurrent or p c rs iix n t d i x a x a f t^ com bin atio n chemotherapy J C lin Oncol I9H8.6 I4J3 V . 88 La i G -M . OzoU R A ^ o n g R C . H am ilton T C . K fle ti o f glulalhionc on D N A repair in c n p l M rcsisU nl human Ovarian cancer cell line. J N all Cancer Inst 1989;81:533-9 1Vol. 321 No. 1 MEDICAI. rKiK.KhSh CANNISI KA 89 Louie KG, Behrens BC. Kinscila TJ. cl al. Radiation survival parameters of antincoplasitc drug -sensitive tod - m o u n t human ovarian cancer cell line* and (heir modilicalinn by Kithioninc sulfoaiminc. Cancer Res 1985;45:21 10-5 VO. Ejfcl P i, Gcrhn*m OM , Dele ho I.. Whanori JT. Petei* 1J Tw ice-daily. iplu-cuursc abdoniinopclvi. radiaiMxi therapy aflci chemotherapy and fu g itive second-look laparotomy for epithelial ovarian carcinoma Int J Radial Oncol Biol Phy l9 v i;2 l:IO l3 -8 91. M anena A. MacNcill C , l.yler JA, cl al Hexamethylmelamine at a single teeond-line agent in ovarian cancer Gynecol Oncol l99fl;.36-93-6. 92. Sunon G P, Bkssm g JA . Homeiley HD. Berman ML. MaKelano J. Phase II inaJ of ifnafamuk and m esna in advanced ovanan carcinoma: Gyneco logic Oncology Group Study J Clin Oncol 1989.7:1672 6 93. Hillcoai BL, Campbell iJ. Peppcrell R. Guinn MA. Bishop JF. Day A. Phase 11trial of VP-16-213 in advanced ovarian CMVinoma. Gynecol Oncol 1983.22.162-6 94. Ahlgrcn ID . Ellison N, lo h ic h 1, el al high dose tamoxifen (TAM): ealended palliation in palienit with chcruacsistani epithelial ovarian can cer (EOC): a mid-Atlantic Oncology Program (MAOP) study. Proc Am Soc O io Oncol I993;I2:258 abstract 93. E iiuig A l. Wiemik Pfl. Sasktff J. et al. Risse II study of Taxol (T) tn patients fPis) with advanced ovarian cancer. Proc Am A>oc Cancer Rci 1990.31:187 abstract 96. McGuire WP. Rowinsky hK , Rotensltein NB. et al Taxol- a unique anuncoplasiu agent with xignilicant activity in advanced ovarian epithelial neoplasm Ann Intern Med 1989;111:273-9. 97. Thigpen T, Blessing 1. Ball H. hum m el S. Barrel K. Phase If inal of laxol as second-1me therapy for ovarian carcinoma: a Gynecologic Oncology Group study. Proc Am Soc Clin Oncol 1990.9:136. abstract 91. Weiss RB. Donchower RC. Wiemik PH. cl at. Hypcrxcnailiviiy reactions from u s o l J Clin Oncol 1990:8:1263-8 99 Sa/oayG . Kohn E. Stone DA. cl si Phase I study of laxol and granulocyte colony-stimulating factor in patients with refractory ovanan cancer. J Clin O n e d 1992; 10.1165-70. 100. Takvonan T. Canelloa GP. R iu J. el al. Prolonged disease-free survival her utologons boor m a n u uansplanutiun in p atents a h non-Hiklgktn's lymphoma with a poor prognosis N Engl J Mod 1987:316:1499-303 101. 102. 103. 104 103. 106. 107 108. 109. 111) III. Philip T , Armitage 11). Spitrci t i. et ul Ihgli-d o * ihciapy and suini, ig n i % hone marrow transjiUniaimn altri failure nf conventional c hcinuihcnijiy in adults with intermediale-grude nr high grade non H odgkin'. lymphoma N Engl 1 Med I9H7.316 149 K Shea 1 C . Flaherty M . ITis> A , el al A |>Iu m ' I >lini.al sud |4isnii.n.ikm cl ic study nf carhoplaini and aut.dng.ui> Itone uiaiiow m i|.|m i ) ( tm liin .il 1989.7-631 61 Jf-xraluiii. J C lin One. 4 I9IW.7 1 1 1 ) | Vneaendorp R. Aal.kr K i. S kijfcr D I', cl al l.il.-.nve high .lo * .liem o (lierapy with aundnginis Urne m in ti m iu .imi m icmnU iii . ju ju . an. ci Gynecol Oncol 1984.17 271-6 Daupfsi I. l-cg i.n M . Condai I'. Fenicie IP. lieu Allinei! S. Rogne K High-dose loclphslali and aulnliiguu> hnlic iiu n n w >u|^uat Ini lic.ilincili ni ovarian carcinoma with positive second U4. o}k.iaiioi> (iyn c.n l O n .n l 1989..34.294 X Viens P. Maranmchi O , I egro M . cl al High dine ii<cl|4talan and anml.. g<his iiu n n w rescue in advanced epithelial uvaiian . j i. moms. o a i m speenve analysis uf 35 paiscnis trea te d in France Hnnr M a rio 11jn>|d.uii l9W;S:227-3.3 Shpall FJ. C 'U iW I'cai.m i 1. Supei f i . et al High dose jikytam.g agent chemotherapy with aulnlngnu. hnue in a ilo * .uppotl hi |u ii.'n l. with stage 111/1V epithelial ovarian cancer (iy n c .n l O m n i f r a i. t h imo 91. Mulder HO. W ille m * PH. Aal.krs H i. el - I High d o * <liciit.hi-ijpy with si.mtog.Ki> h u e marrnw irampi anialimi m p a n n ili with ie lla , miy ovanan cancer Kui J Cancer C lin Oncol |9R 9;2.TM 5 9 Hamilton T . O 'D w ycr P. Ynung R. el al P lia * I inai uf huthmiiiiic sullosinune (B S O l plus ntcljthalan ll.-P A M j in |M iunis w ill, ad. anted cancer. Pr.H` Am Sue O i* O n e 4 199(1:9 7.1 ah .ira.t Stewao JA. Hclmsun JL. Moure A l . rt al I l i a * I mal id n.i ^ > ii. al rocumbrnam interleukin 2/lyinjtbnkine-aiiivsted killer .e lle in |iatirm> wiih ovarian cancer Cancri Re IWU.50-6.X12- III Sicwart l.SW. Hint V . Stand I) , cl al luira|ieriluneal ylliu iiii "*) labeled monoclonal antibody in ovarian i irci J Clin O n .n l iv ai.K 194 I 30. Pai i.H , Bookman M A . ( i / i b KF, et al Clinical evaluation >d iintj|k-au ncal Pseudomonas cxnrnxin minium <c.Hi|ugalc OVfJ.t PF ill palle-ili. uh ovanan cancer J C lin Oncol 1991.9:2093-103 Massachusetts Medical Sm icly Registry on C ontinuing M edical Ktlticatimi T o obtain information on continuing m edical education courses in tiir New Knglaml .h i m , call between 9:00 a.m. and 12:00 noon. M onday through Friday, (017) 1193-4610 tn in M assachusetts 1-800-322-230'J, ext. 13-42. \M I M I'-\N 1>" RNM |-'l*l I . ...n>l tit ',I |I \ rt^ lii 1!>*>! Iv T lir .lu liii" 11' "ik i ri~- I !m \ * r - 11\ S* Ip m j . >l I K ^ i c n c .m>l I ' iiM u l l r . i l l l i \II n^ his rr-r\'il V..I l in. N.. * I 'nnh-i/ in / .M,\ A Brief Original Contribution > LACTASE PERSISTENCE AND MILK CONSUMPTION AS DETERMINANTS OK OVARIAN CANCER RISK DANIKI. \V. CltAMKH Cramer, D. W. (Brigham and W om en's Hospital, Boston, MA 0 2 115 ) . L actase persistence and milk consum ption a s determ inants of ovarian can cer risk. A m J E p id e m io l 1989;130:904-10. Using published data, largely from the 19 7 0 s, the author com pared ovarian cancer incidence, per capita milk consum ption, and population estim ates of lactase persistence (the ability to dig est lactose after infancy) in 2 7 countries. Significant positive correlations were noted between ovarian cancer incidence, per capita milk consumption, and lactase persistence. L a cta se persistence showed a stronger association than milk consumption or animal fat consumption in multiple regression models. The author sp e cu la te s that toxicity from the lactose component of milk and, more specifically, g alacto se, the digestion of which is facilitated by lactase persistence, may provide a biologic b a sis for the correlation. galacto sidase; lactose; milk; ovarian neoplasm s A correlation between ovarian cancer mnrlaliiy and per capita consumption of milk has been described and lias been at;ribut t-tI to ihe Ini content of dairy products il). A nutrient which distinguishes milk roducts is lactose, a disaccharide commsed ol'galactose and glucose. The ability lo digest lactose alter childhood, called Inc ase persistence, varies widely geogrnphiallyand racially. Its correlation wit h ovarun cancer might suggest the importance of actose or its component sugars in ovarian nicer etiology. This brief report describes orrelations between ovarian cancer incilence, milk consumption, and lactase per Heueivf-iJ lor publication Septem ber 2'. IPMH. and ii lim it fo rm A p ril l.'l. 1`iKi) I' r*111 ill*1 I >rp;irl u ic n l 1 <)lisli*i ricN a w l ( iv i ie r n l gv. <I b s ir i r ics ill id ( '.y u e c o lo g v K pidciiti'-IngY ( 'f ill i t , trigliiim and W im in i's H osp ital. H arvard M ed ical M'hool. B o sto n . M A . Itcprim reijucsls (n |)r. Dtinicl W. ('rauicr, Obstet ics and <iynecology Kpidemiologv (V ider, Brigham lid W omen's Hospital, 221 I.nilgWOodAA&MIIIR, B o s on. MA 021 If. T his work was supported by (iran |{OI ('A 1200H rom 11n- Nalional ('am er 1nsl il tile. sistence and speculates on t heir possibly biologic basis. M a t k k ia i.s a n d MKTIIODS Estimates, by country, for the following variables were sought; ovarian cancer inci dence, percentage of adults who are lactasepersistent, and per capita consumption of milk (estimated in fluid weight) and animal fat (from nondairy sources). Data on the incidence of ovarian cancer came from Can cer Incidence in Fire Continents (2) or from National dancer Institute data (3) for the period 1970 1980 and are staled in terms of cumulative incidence, estimated by sum ming age-specific incidences from birth to age 7f> years. Information on population frequencies of lactase persistence came from reports based upon jejunal biopsies in hospitalized patients, serial blood galactose or glucose determinations alter an oral lac tose load, or hydrogen breath determina tions after an oral lactose load (4-2(5). In formation on consumpt ion of various food stuffs was obtained from the food balance sheets published by the United Nations (27). Food balance sheets show, for primary I.ACTASK l'KKSISTKNCK ANI) OVAUIAN CANCKK KISK 905 food commodities, the amount potentially available for human consumption by total ling quantities produced and imported; sub tracting exports, feed to livestock, and other nonfood uses; and then dividing by the total population. Statistical correla tions on the above data were performed using standard linear or multiple regression techniques (2ft). R k s u i .t s Table 1 shows estimates for the cumula tive incidence of ovarian cancer, per capita supply of milk (in grams per day), and proportion of the population who have lac tase persistence. The cumulative (lifetime) risk for ovarian cancer varies from 0.4 per cent in Japan und Senegal to 2.2 percent in Sweden. The per capita supply of milk var ies from 13 g/day in China to 857 g/day in New Zealand. The estimated proportion of the population who hove lactase persis tence varies from 0 percent in Singapore and the Netherlands Antilles to 99 percent in Sweden. A significant correlation (/a = 0.29) was noted between the cumulative T abi.k 1 Cumulative incidence of ovarian cancer, /x*r capita supply of m ilk, and /tercenta^e of populatiim with lactase persisten ce in .32 countries Country (abbreviation) Cumulative incidence* t%) Per capita milk supply (g/day)t Percentage with lactase persistence Reference for column 3 Australia (A U S) Brazil (BKA) Canada (CAN) Colombia (COL) Culis (C U B ) Czechoslovakia (CZE) Denmark (D E N ) Finland (KIN) France (KitA) German Dem ocratic Republic (GDR) Federal Republic o f G erm any (FRO) Hong Kong (H K ) Hungary (H U N ) India (IN D ) Isroel (ISR ) Italy (1TA) Jamaica (JAM) Japan (JAB) Netherlands Antilles (NEA) New Zealand (Maori) (NZ) Norway (NOR) Poland (POL) Romania (ROM) Senegal (SEN ) China (Shanghai) (SH A ) Singapore (ethnic Chinese) (SIN ) Spain (SPA) Sweden (SWK) Switzerland (SW1) United Kingdom (Scotland) (UK) United States (Iowa) (USA) Yugoslavia (YUG) 1.21 331 82 0.94 183 31 1.58 468 NA* 1.22 168 46 0.58 422 NA 1.02 426 88 2.03 432 97 1.44 711 83 0.75 347 58 1.63 389 78 1.81 329 88 0.68 95 NA 0.84 316 63 0.85 105 36 1.66 315 34 1.47 300 49 0.91 78 41 0.43 135 10 0.55 421 0 1.37 a s 7 36 1.87 693 NA 1.00 437 62 0.79 404 44 0.38 97 65 0.47 13 8 0.69 113 0 0.63 329 72 2.12 502 99 1.91 461 84 1.50 455 9b 1.64 462 81 1.40 298 NA 4 5 6 7 8 9 10 7 7 n 12 13 14 15 16 17 18 19 11 20 21 22 23 9 24 25 26 * Based upon age-specific incidence rate in the m id-1970s ( `2, 1)- For countries with more than one regtatry ite, the registry closest to the site for the survey o f lactase in s is t e n c e wan elected, t From U nited N ation s food balance sh eets (1979-1981) averages ('27). t NA, not available. 9(K> incidence <i1 ovarian cancer and the per capita supply of milk (p --0.004) (table 21. Linear regression indicates that tor each 100-J4 increase in the daily per capita supply of fluid milk, there is a net increase of 0.14 percent in the cumulative incidence of ovarian cancer (figure 1). A stronger cor relation (R' = 0.44) was noted between the cumulative incidence of ovarian cancer and lactase persistence t p = 0.0002) (table 2). Linear regression indicates that for each 10-percent increase in the population frei|uency of lactase persistence, there is a net increase of 0.1 percent in the cumulative risk of ovarian cancer (figure 2). Some countries which appear to he out liers in the model between milk consump tion and ovarian cancer are not outliers in the lactase persistence and ovarian cancer model (e.g., New Zealand) and vice versa (e.g., Senegal). This suggests that ovarian cancer occurrence might he better ex plained liv considering both laetase persis tence and milk consumption in a multiple regression model. For this two-parameter model, (he overall R: was 0.50, some im provement over the linear models, with lac tase persistence showing a stronger associ ation (table 2). Lactase persistence also exerted a greater effect on ovarian cancer incidence than per capita nondairy animal fats in a regression model including these two variables (table 2). D is c u s s io n The ability to digest lactose from the jejunum after infancy, called lactase per sistence, arose as a selective advantage sometime around 4,000 B.C. in certain pop ulations practicing dairying (29). It is transmitted as an autosomal dominant trait and remains a characteristic of people with dairying ancestry (30). However, the majority of the world's population, along with most mammuls, do not possess this characteristic and are described as having primary adult lactase deficiency. Surpris ingly little has been written about the pos sible relation between lactase persistence and chronic disease patterns (31, 32). This report has confirmed, using incidence, a previous report correlating mortality rates for ovarian cancer with per capita milk consumption (1) and has described a stronger correlation between lactase per sistence and ovarian cancer incidence. Lac tase persistence does not appear to be merely a surrogate variable for milk con sumption or fat consumption, since it showed a stronger association than either of these variables in multiple regression models; and a model containing both lac tase persistence and milk consumption bet ter explained the geographic variation of ovarian cancer than simple linear models involving each variable separately. T a iii.k 2 Correlations mid mrarneters for regression rootlets of oennon cancer, milk supply, and lactase persistence, with the ineidenee of urariim ctmecr as the dependent uariablc* H ` lo r mod*! ( p ) 0.25# (0.004) 0.44 (0 0002) o.:w 1.00 n o. i9 (0.ooo;o 0.50 (0.OO02J In te rc e p t param eter (U )7;i 0.524 0.510 0.:w<J o :<<H M ilk supply p a r a m e te r (/) 0.0014 (0.004) 0.0008 (0.20) 0.0008 (0.0!)) I -iitiase p ersistence param eter (p) 0.011 (0.0002) 0.008 (0.009) 0.009 (0.004) A nim al fat param etert (p ) 0.008 (0.10) ll.OOfi (0.13) S e e (a id e 1 lo r s o u rc e ul iiic id e fit'g y tiid m il r ie n l in fo rm a l io n . t A n i m a l O il f r o m n o n d a i r y s m ircf^ iT 1y lis m l c n l a l e d ly s o h l r m t i n g a n i m a l f a t f r o m m i l k p r o d u c t s f r o m t o t a l a n i m a l l a i v i s in g d a t a |>r o v i d e d i n rt-r r r e n r o 2 7 . LACTASE 1`KUSISTENCE AND OVAKIAN CANCER RISK 1+ ------- ---------+ --------- + --------+ --------- + ---------+ - -- 2 2+ | 2 .0 * onP'' o$W E oOEN oSw i cl 1 8 + id <j z < 1 .6 + o1 z < 1 .4 + CL 1 < o> 1 .2 + joli_ 1 .0 + J</> CL 0 .8 + UJ *>FRG o'S R 0 GDR0USA oCAN o lT A o VUG oUK oCOL o AUS oJAM o0RA o'N O C Z E Oo p 0 L ,,HUN ,,ROM o f RA 2 H K 0 0 SlN \- 0 6 + LU I ,,SPA 0 CuB N [ A U- SHA -- \ o JA P 04 + SN oN0R FIN 907 0 .2 + l 0 _________ t _.____; 100 200 3 0 0 4 0 0 5 0 0 6 0 0 7 0 0 8 0 0 9 0 0 P ERCAPl TA S U P P L Y OF MILK IN GRAMS/DAY Fig u r e 1, C orrelation between ovarian caneer incidence and per capita m ilk consumption in .'12 countries See table I for country abbrevialiona. Readers will be aware of the limitations of ecologic studies such as this one, includ ing inaccuracies of the data and potential confounding factors. Data on the per capita supply of foodstuffs are limited by the ac curacy of the gross data provided by the countries. Estimates of lactase persistence vary because of differences in techniques used and the specific population sampled (33). Jejunal biopsy is a precise technique but is limited to a select group of hospital ized patients. A preferred test is the hydro gen breath test after oral lactose, which is based upon the observation that undigested lactose is fermented by colonic bacteria, releasing hydrogen which is absorbed in the blood and exhaled. Probably most reliable are the data on ovarian cancer occurrence, because they come from established regis tries using standardized age and disease categories. Other factors associated with both lactase persistence and ovarian cancer could account for the correlations observed, and it must be conceded that not all poten tial confounding factors have been consid ered. Despite these limitations, the author believes that these data provide descriptive support for a model of ovarian cancer as a consequence of galactose toxicity. A population with high levels of both lactose consumption and lactase persis tence is a population that will be exposed togalaetose -the carbohydrate that is com bined with glucose in the disnccharide lac tose and whose digestion is facilitated by lactase persistence. There is evidence that C 'K A M K K *-- *-I - + -- - - ------- + --------- --------- + -- f --------- + -- *+ --------- -- 2.2 * 0 SW t 2 .0 * 1 ir 1 .8 Ld | o zl <t 1 6 * o <z 14 * oa<>: 1 . 2 * UO 1.0 * CO a: 0 8* UJ 5 L sin K UJ 0 - 6 j . N L u. o S n A _J 0 . 4 * JAP 0 z\ L oDEN 0 ISR oSW I CDR0 0 U S A | - I 1 -* 0 !TA 0N7 ,,COL ,, k 0UK o FIN 0 AUS 1 + | | + I 0 8N A 0 JAM 0 iN[) ,,ROM 0 ROl 0 muN n riiA n SPA - 0 SE N + + + _____ + _____ + ______ + 0 CZE 1 + I 1 + I + I _ I 1 + I _ + + ______ 0 10 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 PERCENT LACTASE PERSISTENT r F16UKK 2. ('orrelalion between ovarian cHnrer incidence and lactase [ersistence in 27 countries. See table 1 ir country abbreviations. alactose may lie toxic to ovarian germ cells rom animal studies showing that female odents fed high-galactose diets produced ffspring with reduced oocyte numbers (34) nd that nonpregnanl females developed vulatory dysfunction (35). (ialactosemic tomen who are unable to metabolize galacose develop premature menopause (36), nd this investigator has recently reported hat carriers for galactosemia also have an arlier menopause (37). Experimental evidence supports ovarian nilure (hypogonadism) as a predisposing actor to ovarian cancer. Rodents receiving varian irradiation or oocyte toxins such as olvcyclic hydrocarbons or having congeli al deficiency of oocytes develojje ovarian umors, primarily of stromal types (38). That gonadotropin stimulation was a nec essary component of the model (hypergo nadotropic hypogonadism) was inferred from the ability of pituitary ablation to block tumor development. But the rele vance of the rodent models has been doubted, because rodents develop stromal tumors of t he ovary and humans most com monly develop epithelial tumors of the ovary. However, humans have much greater stromal epithelial admixture in their ovaries (inclusion cysts), leading to the possibility that, under the same stimu lus, humans may develop different tumor types than rodents. This model is discussed in more detail elsewhere (39). In conclusion, the author believes that these descriptive data on lactase persis- .ACT'ASK I'KUSISTKNCK AN D OVAIIAN CANCKIt RISK 909 tence and milk consumption in relation to ovarian cancer are pertinent to its etiology and that additional studies regarding the consumption, absorption, and metabolism of galactose in relation to ovarian cancer are warranted. References 1. Rose D P , Boyar AP, W ynder EL. International com parisons o f m ortality rates for cancer o f the breast, ovary, prostate, and colon, and per capita food consum ption. Cancer 1986;58:2363-71. 2. W aterhouse -I, M uir C, Shnnm ugarotnam K, et al. Cancer incidence in five co n tin en ts, vol 4. (1ARC scientific publication no. 42). Lyon: International Agency for Research on Cancer, 1982. 3. Young J L Jr, Percy CL, Asire AJ, eds. Surveil lance, Epidemiology, and End Results: incidence and mortality data, 1973-1977. (National Cancer Institute monograph no. 57) (NIH publication no. 81-2330). Bethesda, MD: US Department of Health and Human Services, 1981. 4. Bolin T D , M orrison RM. Steel JE, et al. Lactose intolerance in Australia. M ed J Aust 1970; 1:128992. 5. Seva-Pereira A, Bciguelm an B. Primary lactose m alabsorption in healthy Brazilian adult caucasoid, negroid, and monogoloid subjects. Arq G as troenterol 1982;19:133-8. 6. Fajardo LF, Leal H, Victoria F, et al. Milk in to l erance in Colom bian children: its prevalence and relation to lactose malabsorption. Arch Latinoam Nutr 1979;29:329-39. 7. Flatz G, How ell JN , Doench J , et al. Distribution of physiological adult lactase phenotypes, lactose absorber and malahsorber, in Germany. Hum G e net 1982;62:152-7. 8. Busk HE, Dahlerup B, Lytzen T , et al. T he inci dence of lactose m alabsorption in ulcerative c o li tis. Scand J Gastroenterol 1975;10:263-5. 9. Sahi T . L actose m alabsorption in F in n ish speaking and Swedish-speaking populations in Finland. Scand J Gastroenterol 1974;9:303-8. 10. O'Morain C, Louhiere M, Rampal P, et al. Etude comparative de I'insuffisance en lactase de deux populations adultes differentes. (In French). Acta Gastroenterol Belg 1978;41:56-63. 11. Flatz G, Czeizel E, Flatz D S. Prevalence o f adult lactose malabsorption in Hungary. Orv H etil 1984;125:147-51. 12. Desai HG, Gupte UV, Pradhan AG, e t al. Inci dence o f lactase deficiency in control subjects from India: role o f hereditary factors. Indian J Med Sei 1970;24:729-36. 13. G ilat T . Lactase deficiency: the world pattern today. Isr J Med Sei 1979;15:369-73. 14. Brgin GR, Flatz G, Barbera C, et al. Prevalence of primary adult lactose malabsorption and aware ness of milk intolerance in Italy. Am J Clin Nutr 1984;39:100-4. 15. Stoopler M, Frayer W, Adlerman MH. Prevalence and persistence of lactose malabsorption among young Jam aican children. Am J C lin Nutr 1974; 27:728-32. 16. N ose O, lida Y, Kai H, et al. Breath hydrogen test for detecting lactose malabsorption in infanta and children: prevalence of lactose malabsorption in Japanese children and adults. Arch Dis Child 1979;54:436-40. 17. Erkelens D W , Houman JG, Jansen W. Lactose intolerance and lactase deficiency in adult p a tients o f C urasao. Ned Tijdschr G eneeskd 1972; 116:1080-4. 18. Abl>ott WG, T asinun-Jones C. Incidence o f ac quired primary hypolactasia in three New Zealand racial groups. NZ Med J 1985;98:228-9. 19. Socha J, Ksiazyk J, Flatz G, et al. P revalence of primary adult lactose malabsorption in Poland. Ann Hum Biol 1984;11:311-16. 20. Arnold J, Diop M, Kodjovi M, et al. Lactose intolerance in adults in Senegal. C R Soc Biol (Paris) 1980; 174:983-92. 21. W ang YG, Yan Y S, Xu JJ, e t al. P revalence of primary adult lactose malabsorption in three pop ulations o f northern China. Hum G enet 1984; 67:103-6. 22. Bolin, T D , Davis AE, Seah CS, et al. Lactose intolerance in Singapore. G astroenterology 1970; 59:76-84. 23. Garcia JG , Gom ez R, Querenda A, et al. Intoler ancia a la lactosa en la poblacin espaola. (In Spanish). Rev Esp Enferm Apar Dig J974;42:36782. 24. Auricchio S, Rubino G, Semenza G, et al. Disaccharidase activities in human intestin al mucosa. Enzymol Biol Clin 1963;3:193-208. 25. Ferguson A, M acDonald DM, Brydon WG. Prev alence of lactase deficiency in British adults. Gut 1984;25:163-7. 26. W elsh JD, Rohrer GV, Walker A. Hum an in te s tinal disaccharidase activity. I. Norm al individ uals. Arch Intern Med 196(5; 117:488-94. 27. Food and Agriculture Organization of the United N ations. Food balance sheets, 1979 -1981 average. Rome: FAO, 1984. 28. Kendall M, Stuart A. T he advanced theory of statistics, vol 2. New York: Macmillan Publishing Co, Inc, 1979. 29. Sim oons FJ. The geographic hypothesis and lac tose malabsorption. Dig Dis 1978;23:963-80. 30. Sahi T , lsokoshi M, Jussela J, et al. R ecessive inheritance of adult-type lactose malabsorption. Lancet 1973;2:823-6. 31. Segall JJ. H ypothesis: Is lactose a dietary risk factor for ischaemic heart disease? Int J Epide miol 1980;9:271-6. 32. Sim oons FJ. A geographic approach to senile cat aracts. Dig D is 1982;27:257-64. 33. M cGill l)B . D iagnostic tests for lactose deficiency. In: D elm ont J, ed. Milk intolerances and rejection. Basel: Karger. 1983. 34. Chen VT, M allison DR, Feigenhaum L, e l al. Reduction in oocyte numlier following prenatal exposure to a diet high in galactose. Science 1981;214:1145-7. 35. Swartz WJ, Matt son DR. Galactose inhibition of 910 OKAMKW ovulation in mice. F enil Steril 1988;-l9:.r22-tj. 36. Kaufman FR. Kogut M l). Donnell (IN . ct al. Hypergonadotropic hypogonadism in female p a tienta with galactosemia. N Kngl .I Med I98l; 304:994-8. 37. Cramer DW , Harlow RI., Hnrhieri RL. et al. Galactose-1-phosphate uridyl transferase associ ated with age at menopause and reproductive h is tory. Fertil Steril I989;f> 1:009 10. 38. ('ram er 1)\V, Welch WR. D eterm inants of ovarian cancer risk. II. Inferences regarding pathogenesis. JNCI 1983;71:717-21. 39. ('ram er DW . Harlow B l,, W illett WC. e l al. Ga lactose consum ption and m etabolism in relation to the risk o f ovarian cancer. Lancet I989;2:f>671. & ie h ic a n J o u r n a l o f F.p i d e m i o i.o g y .pyright (C 19Kb hy T h e JnhnR H opkin s U n iv ersity S ch o o l o f HvKiene and l'u ld ic H ealth All r ig h u reserved Vol i:J. N o V i(</ in 11 S A RELATION OF BODY FATNESS AND ITS DISTRIBUTION TO CARDIOVASCULAR RISK FACTORS IN YOUNG BLACKS AND WHITES THE ROLE OF INSULIN AARON R. FOIJSOM.' GREGORY L. B U R K E ,' CAROL BALI.EW .D A V ID R. JA C O BS, Jit.,1 W ILLIAM L. H A SK E L L ,' RICHARD l>. D O N A IH IE ,* K IANG LIU ,1 a n d JO A N E. HILNEK" Folsom, A. R. (Div. of Epidemiology, U. of Minnesota School of Public Health, Minneapolis, MN 55455), G. L. Burke, C. Ballew, 0. R. Jacobs, Jr., W. L. Haskell, R. P. Donahue, K. Liu, and J. E. Hilner. Relation of body fatness and its distribution to cardiovascular risk factors in young b.':ks and whites: the role of insulin. Am J Epidemiol 1989;130:911-24. Persons whose body fat is distributed predominantly in the abdomen compared with the hips are at increased risk of several chronic diseases. This study examined the cross-sectional relation of percent body fat, computed from skinfold thickness, and fat distribution, measured by the waist-to-hip girth ratio, to phys iologic cardiovascular risk factors in a biracial sample (blacks and whites) of young adults aged 18-30 years. The subjects were persons who were examined at baseline (1984-1986) in the Coronary Artery Risk Development in Young Adults Study in four US metropolitan areas. The two hypotheses tested were that 1) after adjusting for percent body fat, waist-to-hip girth ratio is associated with several physiologic risk factors, and 2) fasting concentrations of serum insulin partly explain such associations. Percent body fat was significantly associated with all measured blood lipids, lipoproteins, apolipoproteins, uric acid, and blood pressure. Waist-to-hip girth ratio was significantly, although more weakly, asso ciated in multivariate models with blood concentrations of triglycerides, high density lipoprotein (HDL) cholesterol, HDL, cholesterol, apolipoproteins A-l and B, low density lipoprotein cholesterol (in women only), uric acid, and systolic blood pressure, but was not associated in either sex with total cholesterol, HDL, cholesterol, or diastolic blood pressure. Fasting serum insulin concentrations were significantly associated with percent body fat (Pearson r = 0.45-0.53), waist-to-hip girth ratio (Pearson r = 0.18-0.27), and most of the physiologic risk factors. Inclusion of fasting insulin in multivariate models reduced, but rarely eliminated, associations between waist-to-hip girth ratio and the physiologic risk factors. These findings suggest that obese young adults, especially those with abdominal fat preponderance, carry a physiologic profile that places them at higher risk of cardiovascular disease, and that tasting Insulin concentrations are only partly explanatory. adipose tissue; apolipoproteins; blood pressure; insulin; lipids; lipoproteins; obesity The relations of blood lipids, glucose, uric well established. Several investigators have acid, and blood pressure to body fatness are recently hypothesized that these relations Received for publication July 14, 19K8, and in final form February 13, 19B9. Abbreviations: HDL, high density lipoprotein: I.DL. low density lipoprotein, 1Division of Epidemiology, University o f Mimic sola School of Public Health. Minneapolis. MN Mil. I ANUI- l.| U l Y b , 1989 G Al .AGI O SE CONSUM PTION AND M ETABO LISM IN RI.ATION TO T H E RISK O F OVARIAN GANGER D.ANlfcl W . (iRAMHR1 W a i. it r C. W ii i i r n 2 D fbra a . Bf m W o N ti Nil* Bf r n a r d l . . H a r l o w W illia m R W fx ch w R o h h r i E . S c u i j .y4 Rohlr i C. K napp* <V>iVml J 't.1 ('oiler, Jhponmeni oj O buetm i Deponnuiu oj Lpulerwobigy toiJ Nutrition. H arvurJ S i \ *f o j Public H ealth , a n J the C.hanmni l*jtk.xotor\ ,-* unJ Depsirmuftt o j PatK'J<i(y, Women'i and P eriruiat D n u io i, lin^ham j >i.1 IF'iw irif'i H oifntal, H a rw irJ M tJ u a J .5ho>'t . ' llepai ttnem oj P a th o lw , A L in d . Museta General H oipuat, H an orJ M .Ji.ol l hi in on oj M eduaJ <renen o . (JuLircn'i flo ip u a / oj /j< i .ir^ e U t.' a n J Ih visu n >/ GvneioL<gtc ihvolog)-. D e tr im e n t o j ( tbuetru i anJ </vWeW.jcy. H ngham a nJ Women'i Hoxpttal, H o n o rJ M edical School * S u m m a ry In a case-cunt rol srudy, a s s u m p tio n o f dairy foods by 2 35 w hile w om en w ith epithelial ova nan cancer and by 239 control w om en, and activity o f red blood cell galactose-1-phosphate uridyl transferase (transferase) in a su b set o f 145 cases an d 127 controls w ere determ ined. Y ogun was consum ed at least m onthly by 49% o f cases and 36% o f controls. T h e mean transferase activity o f cases was significantly lower than that o f controls. W h e n a ratio o f Lactose co n su m p tio n to transferase (L /T ) was calculated, cases had a m ean L /T of 1 17 co m p a red w ith -98 for a m iro ls ; th ere w as a highly significant trend for increasing ovarian cancer risk w ith increasing I . T ratio. I a c tose consum ption may be a dietary nsk factor and transferase a genetic risk factor for ovarian cancer. In tro d u c tio n F or py y ea n ago, G a rd n e r' pm pcd that ovarian cancer was caused by hypergonadotropic hypogonadism -- ie, high secretions o f gonadotropins due to ovarian failure or lack of feedback control on the pituitary. In 1983 w c2 restated that theory and began a case -control srudy to exam ine novel exposures that m ight be associated with ovarian cancer through this m echanism . Because dietary galactose consum ption and a key enzym e involved in its m etabolism , galactose-1 phosphate uridyltransfcrasc, have also been linked with hypergonadotropic hypogonadism , we have inducted a ease-control study to investigate the relation between ovan an cancer and galactose consum ption and m etab o lism S ubjects a n d M ethods Subjects The study w u restricted io Hnglith-speak mg while resident of Massachusetts within an 80 km radius ol Boston who were aged between 18 and 76. Gases were women with ovarian am eer diagnosed ai ten participating hospitals ui the greater Boston area between July, 1983, and Septem ber, 1987. and uicniilicd through rumour boards, pathology registers, or the Massachusetts State Cancer Registry During the srudy penod, 394 potential cases were identified; 42 women (1 1%) could not he interviewed because the physician denied contact, 43(11% ) refused, and 37 <9%) had died or had moved before contact. For each o f the 272 interviewed cases, slides uf surgical specimen* w eir reviewed. 10 cases were excluded as rumours metastanc to the ovaries, and ilic remaining caw were classified according to histotognal criteria of the World Health Organisation ciaskificanun o f ovarian tum ours.' I"he present analysis is restneted io the 2 M women with epithelial ovarian cancer including tumours of borderline malignancy. Control women were identified through the Massachusetts Town Books--annual publications that list residents by name, age, and Kidrcu. For each interviewed case a list of 5 potential control* was prepared by random selection from those women with listed phone num bers who matched case* by area of residence, n e e , and age within two yean. These potential controls were contacted until an eligible control was identified who agreed to participate. Women who said that they had had bilateral uophorcciomy were excluded. O f 526 potential control names on our lists, 129 (25% ) could not be contacted because they had m oved, died, or were otherwise unreachable, 54 (10% ) were ineligible because of oophorectomy or incorrect age o r race m atch, and 102 (19% ) declined io participate. 'Ih u s , 2 39 controls were included m ihis study. D ie ta ry H istory T he women were interviewed io assess reproductive, medical, and family history. Dietary history was obtained with a aclfadm inisured semiquanutauve 116-iicm food frequency questionnaire developed and validated for the Nurses' Health Study.*' Cases and tim ings were asked to fvicus tm their dietary patterns o f the previous five year*. Case* were reminded nut to show changes in food preference* since the diagnosis of their cancer. Questionnaires were checked for completeness and open-ended questions were coded by a nutritionist; questionnaire* were then read by an optical scanner. Nutrient summary score* were prepared with computer algorithms based on nutrient value* for standard food portions.' Galactose consumption was measured indirectly: it is equivalent to lactose consumption since lactose is virtually the only source of dietary galactose 236 ml of milk contains about 11 g of lactose and 5 5 g of galactose. W e estimated lactose consumption from the frequency of use of eleven dairy products including whole milk, skimmed milk, ice cream, various cheexes, and yogun but it was not feasible to estimate lactose intake from fillers in food products or in medications. A VT.HSlt:R ANDl Illll.KS RF.FP.KKNtSiS--mnnm/ II Hc<r> NJ IMU. * 1(1. nutxaSn ipiiui eyi^n<ak>rirva J I W Atctk liM. It: d AlO S ot AIDS tiltin l ldiJ iu ih m f r o l-.lt. n|i tuner! ItoafauJ nA1 tu AlrJ J |<M|, 2M| 74VVl i Arrruugc P, Hoiy < S*mjmkjiI ) rntlwM mord 2nd cxJ OtKil hUOrril. II l <inlTiihi 111.Hif>nun t . Apcir.jm.uxi iKrcf pitpuix) (inti i<r<i ft' J th iu i IWU,Sli ? IS 4 | o . n JH. |-| i A. 1-nfVawl [A. I d ! JK. Uallouf MM PiT'alrew ,4 iV * w r j l in v iiilid ) i I w n t ^ u U , and Ism aw iiB S in ln io l wttfi human uinani>lrik >ijv>rrpe I lia u A u a luat. it : Ini m 5 Qi.am ll l'a I*. Mi<mul III. n 11 Sif.4. avl iinmunWyn itwln in palienl. nih Alt'S m Ajiiciu JA liu JAMA IWI7,JUi M I -21 RM-Imhel* Ml 1 AM Strain |A. I cr Hi, <I'XUllry PM. Nhema AJ llnt4vn>o,,.<.|.i.nnAII)!> u 7o 47 7 P a BKPuK. Ilnttmon N.ciil I'm!*he, <4On qu/al unmunr ikfionx) >)nJr.<iir .ic'd f nj in as>4*o id 'pi*r ti*.'<->uai men N h*glJ M<d |VK7. lite pi re, !M.M.AH.Hachen. P.In,J t'.rul y 1.4vui1.1r. Sc MJVanJihcik'd imeni |v (iru n jy J t . S V K n w ^ JA. Santlrrwm AK, (w drahi I'll m4 W ip n n aii lW , ll; W7VN 3 1IlinM ti.K am ) S C.K anm J.l'i^o')! HuangI ImnolMiecw^ ptie ic^n 4 human cyuawfhSiovui iiwixaoi.cir, dv p(-*Tcr ,4 luotan immun.lrcxcTKy >mn IVu. NmI At+i S.i USA |W7, 4,0042 -to It Sk.dmk PH. K..H1 hH. Minch MS Hh|< m*rK*wrcm i vrui lypr * **>** ''*1' V 14 II N hui JA. Mr>T>4J Kriilo (.. OUx.au MILA. Wile> A MIS' bl IK MV cixaiieci Mui erth ai patient' uuh AIDS lV.U-r> iw . Il; UK*1 II KuiaLS. r.H. Iliac PM. Ilirch MS Imrru'uai .4 .yi.anegSinn -Mti lntM n li.<11 paiuno 111mavaiuckcau due 11.>yi.cnegik..irux f /yi ( |u77, IMe 14 1S> . Dl . la l-auc. IITU Jired m i-IA N lll-.l . JUI VH, |4K9 b7 I A H I I I I I IN K A1 A N K D IW IX .K A l'I t ll <.11AKAC.1 1.8 INI II S Agt lyri All ubjcilt SuhjfriHip with q u v iliu tiv c imlCTc mcaiurcmeim No of No .if c*set i 9 , / om irub [ % i ( n - 2)5) (n 2W ) Noof (n - 145 No of o m in b / 7 (n -1 2 7 ) J5-49 50-64 2 65 I'anty 0 t 2 J 2* HthgVm Caihi>8tc Jewish PfOtCSIMII OthCT HJucjrum l %-r, S 12 12 Atunint ifolui Nevei married Ever marned Oral cmirocepJnti N e v a uvet) Ever uved 65 ,27 7, V ) , Jv6, 46 i / 9 6 . 67 2* Oi H 8 t)6 B , 50 i 20 9 J )7 #24 5, 6 7 ,4 6 21 23 r /5 9 / )2 (2 W 54,42 S 24 , IB 9 7V , / f 6 , 11 H I 2 40 , / 7 p , * (2119 < V> i IS U 4 ) i IHOi 2 7 /1 /Jl tri ,2 6 4> 4 ) i IB Ol b\ ,26 4 i 4K . II 1i 17 ,11 71 24 i I 6 6 i )2 i 22 11 24 i / 6 61 21 , 16 5 16,126 3 3 . 2b 0 29 ,2 2* ?H , 22 / 1IK i V) 2 ' 15 i 14 9 , 6b i 2 1) 16 l6 B l 1)2 i 55 2 i 21 18 B> 76 i II Hi 1 0 /4 2/ 7H 15 /K , 22 itS 2, 40 /2 76 , 5iJS / 69 , 54 i 17 , / ) 4 M ,29 1 4 t I2i 9 ) , /9 6 , 1)5 <4K / 1 57 / 19 h 142 iM 4> 124 i ' / 9) 8 H /6 0 7 , M (504 6) (496, 40 117 0i 2b 110 9) l i l l S 9l 12 ( 9 5 / 165 <H t0` 21) IHV 1/ 122 ,84 II 115 i 90 6 / 168 171 Si 157 ,65 7, 102 i 70 h 67 ( 2>l 51 H2 I 14 1/ 4 ) ,29 7! 85 /66 9/ 42 ( i ) / / T ram /trust A a n ity 213 (91*/*) c i k i and 17f(71% ) controls provided a blond specimen fur a n a l)n i of galactose-1 -phosphate uridyltram fcra (transferase). Initially finger-prick specimens were obtained to m an u re transferase hy the Rcutkr lew, a qualitative technique developed for newhom screening for galaclnuemia: in ihi* teat a fluorescent reaction is lelaycd in subject with reduced enzyme activity* These analyse* were lone at the newborn screening laboratory of the Massachusetts Department of IHiblic Health. I atc r in the study venepuncture specimen* were obtained for more precise quantitative measurements if transferase activity with a c a rb o n -U ('`<1) labelling m ethod.' Transferase activity is reotrded in micromoles of uridine diphr*phaic galactose formed from *H'.-labelled galactose-1 -phosphate precursor per hour per gram of haemoglobin (pmol/h per g Hh). Ihcsc analyses were lone at the Children's Hospital n f 14 Angeles on hepannised blood shipped by u r express at mom temperature within I or 2 day* of collection. Samples were analysed within an average of I b lays of shipment in cases and I 5 lays in cum mis. Blood collection was delayed for at least 3 m onths in any patient who had had a transfusion. Finger-prick specimens only were available inert 68 cases and -13 coo.roh, venepuncture specimens only were available from ">7 eases and 79 a m t rob, and both were available from HD cases and 48 controls. Analyses o f trznslcnse activity focused on the I-15 eases and 127 onntmts who gave venepuncture samples. Statistical Analyses Differences between cases and controls in die distribution of quantitative variables such as lactose consumption or transferase activity were measured hy t tests o r hy Wilooww rank sum tests depending on the normality o f the distribution." Relative risks for any particular am*Hint of lactose consumption wtransferase activity were determ ined with multivariate togislic regression for nonpaired data to cuntml lor jioicntial confounding factors * A ram-patred technique was used because eases and atntm ls were m atched only lunsely lor demographic factors and because bkxiJ results were not always availahk fur both ease and control members o f a pair. In these analyses, nsks were adjusted for potential confounding factors including age (< 5<>, 2 50), pariry (0 ,1 -2 , '2 ), religion (Jewish, non-Jewish), education ( 1 2 years, > 12), and I A111 I I I A iq t/M I |l K it A l|V |k lS K M i K llV A K lA N < A N ll.K a s m v i a t i l t vet t it t .iiN s iiM i' i it in t u- v A K in n s i y i*i -,m u - i i a i k y i'Moiuk i.n D ir y ( 'rtiim -Am w / 2B t ' Never of < mooihly 2 MonOily <`Mligt ctutu 1 11 1g 1 Never in < iitoaiihly 2 Monthly /,r ve<iw t I H g l Never IK - moodily e M iihly he milk i l l l g i Never or < nsviihl, > M ixiihly StamrtuJ mttk I216 ml i Never or < iixuiitily 2 Monthly l a / r m / i ,216 ml / Never or < nsKiihly ^ M o n ih ty V.y/wf ( 2 2 6 /, Never * nxxirhly 2 Monthly No.4 eaves . (l> 2 51 No of > in > 219. K K t9 51 .. (.1 118 i 5in 117 , S(J< 85 ( /6 / |5I ,64> 54 / 2 b IK I , 77/ 176,75/ 5 9 ,2 5 / 112 |4K, 1 2 )(5 2 / J l h . 4Vi 1 1 9 ,5 /, 121 , S/114 .4 9 / I l l ,55/ Kin ,4 5 , 10 1 2 , OK 1 7, llf> i 44 . l U , 56, 1n M ill! 2 l| 49 , 21 / |O0 r 7V, 10 <18 ,u 5 1 b IH4 , 77, 4 5 ,2 /, 10 1 i 1 7 1 M 121 1 5 /. 1IK i 4 9 1 1n 1 <l(il 7 1 5) 121 , 5 / . 1IH .4 9 , 1 i> 1 >0 7 16) 1 5 )(6 4 / l 16 i 10 1 7 .1 1 2 4,1 RR relative r id : all n i l i arc relative in never eating the product or eating it on leu than monthly bwsti and are adjusted fur age ( 50. t V )), paniy (0, I - 2. > 2), education ( n 12. 12), oemj oxiiracepiive u>e (ever, nevei), i4al calorie Intake, and rrligmn. *p-0 tW;p-(0l. ontl contraceptive use (never o r ever). Also, total cak'nc intake was included in the inodch to try to adjust for any general overreporting or under-reporting of quantities unturned. Keaulfa T able l show* clinical tuul dem ographic characteristics ol all eases and c o m m it and o f those in w hom quantitative transferase was m easured Age, a key m atching charac teristic, did not differ between cases and controls. Cdses w ere m ore likely to be Jew ish, college educated, never m arried, nulliparous, and never users of oral contraceptives th an co n tro ls. CCharacteristics if th e s u b g ro u p o f eases an d controls in w hom quantitative transferase was m easured, reflected general differences between eases and controls: thus, there was n o m ajor bias in the selection of the group from w hom specim ens w ere taken by venepuncture. T a b le II show s risk lo r o v a n a n ca n ce r a sso c ia ted w ith regular coosum pnon o f various types o f dairy products relative to never use o r use less than once a m o m h o f each product. Risks for regular use of these products were generally m o re th a n 1; yogurt a n d cottage cheese w e re th e only products w hose regular use especially distinguished cases fro m co n tro ls, B ating yogurt at Icum o m e a m o n th was asstxiatcd w ith a relative risk for ovarian cancer o f 1 7 (p ( H ll) : for co llag e cheese tins risk was I I (p 0 OH) T o tal lactose consum ption hy all cast's an d controls was skew ed rig h t an d no t norm ally d istrib u te d `H ie m e d ia n iutd m ean lactose- co n su m p tio n for case's w ere 11 H und 14 H g/day, respectively, and for controls were I I 0 and I 3 3 g/day. Ih esc differences w ere not significant. In th e uj*|Hr ra n g e o f lactose c o n s u m p tio n , 51 (2 2 % ) isises c o n s u m e d ^ 22 g <if lactose a day co m p a red w ith 17 ( l b " ..) co n tro ls (a d ju ste d relative risk for ovarian cancer I t>; p 0 (Hi, `5";, con fid e n ce interval | ( ' l | I 0 -2 7). According to the Beuiler test, 51 ( )3%)tf 15b cases h;ul a 50% reduction of transferase- activity compared with IH I AMI I- I I I | A* AND IR A N -. Sr.MI' l ION IN A l 1 I A M : A N D u M I K O I . M I N 1111. SUNSI I l) |' Sl'H JI (.*1 S W | I I I v i-N l H U N tT U H l Nl l > l> SAMI'I I'S R Y V A R IO U S ( H a K A C I T K I V I l'-S o iM jm jv n< n r a u l o i K activity 1K1 - 2 51) t mi m b in -2191 ( ax* in -14 1) Q n m ili (n I27) N il Median* No Median* N il M eant No M ean! if' so I I M Mil 12 v 11 21 6 49 2 )2 V ) 1tv 12 1 1 Vi lU s u JO 21 9 7B 22 6 I'jnrv 0 79 1 )2 41 11 w< 22 2 21 22 0 1-2 71 t o * 0 Mi 2 41 22 1 49 2 B Bl 1 )4 in& I I 4 5o 2) 2 17 2 ) 1 R th fi.* Jewnb 11 I I 7 21 74 22 21 D 17 21 a N <l-)c*nb 21 I I H 21* n o 121 21 B n o 240 i vi S i 2 1 I t 0 IIS III 1 17 22 1 64 22 a > 12 I t.' 124 124 12 H NR 21 6 63 22 iq AliiM j / ilMui Never married a* 12 4 26 M l 21 22 9 I2 22 4 Kvei m im ed is I I K 21) M l* <b j l . .mirk rpm 122 21 6 111 22 94 Never uved I6A 11 1 117 v a j 102 21 H N1 22 6 F er uved b7 6 H2 I I 21 41 21 42 2 ) *toy ijim iJ urtdmc d ip ls ^p h n c mUebBe h pei g H b Ip %0 0 1 for difference between ih n and lb cum plenicnutry category among cum rub < n ly 4p i 0 (K between nJ tim rn M h,r ih>\ laicg ury (20"o) of 91 controls ( p * 0 ,03). This observation, suggesting that there was lower transferase activity in ovarian cancer eases, prompted further srudy with the quantitative technique for analysis o f transferase activity in 145 cases and 127 controls. Transferase activity was normally distributed in case and conrrols: mean transferase activity fi>rcases (21 8 jimol/h pci g I !b) was lower than that (22 Hjunol/h per g Mb) for control (p - 0-0)). Ai the lower end of the transferase disiribution, there were SH (40%) lA l U l - I V Kl-.l A 1IS ! R IS K S I t W O V A R IA N A W I K H I I At | ilM - l'AN Sl*M I'l IO N ANI> I KAN SI I-R A M ' A C I I V | I V ... ( Vvnbmaimn No nt \uhievii <' 1 Act* cnniumpnun (day) .Nil S ll *II I I 1 raitilore activity iu-ima 1 per g H b) ? 22 V v22 9 >22 9 22 (am ( n - 141) 2 i 201 Ml l i l l 33, 2i< 1 1 1 fan in - I27i KR* (91V. C l) 36 ,2 * , 31 ,2 * i 2 (2 i> 2 7 ,2 1 , 1 01 1 1(0 1 2 2) 1 3 (0 6 -2 7)1 2 2 |l 1-41) *AU n i k i Me rv U lu c io reference caic-gury IR iftin tc iic fiiy . Ip : (VOS eases compared with T5 (28%) control with transferase measurements below 21 pmol/h per g I lb (adjusted relative risk 1 7; p = 0 04,95% C l I 0-2-9). Further adjustment for interval between specimen shipment and analysis or for year o f analysis did not alter this association. Table ill shows median lactose consumption for all cases and controls and mean transferase activity for the subset of subjects with venepuncture specimens by various characteristics. More lactose was consumed by controls under 50 years old than by those older than 50 (p = 0 04). Possibly related to this age difference was the observation that more lactose was consumed by those controls who had used oral contraceptives than by those: who had not (p 007). Other non-significant differences included greater lactose consumption by non-Jewish compared with Jewish controls, more educated compared with less educated controls, and never married compared with ever married controls. Case subjects older than 50 consumed more lactose than older controls (p * 0 05): likewise, ease subjects who had never used oral contraceptives consumed more lactose than controls who had never used them (p=-0 0 l). There were tv* significant differences in transferase activity by demographic characteristics within the control group fas. had lower transferase activity than C ASE S ( n =)4 5 ) C O N T R O L S I '-1 2 7 ) 35 35 X 30 30 25 2 5 20 IV .* *1 i, v . .... m 4* *(i a%* 15 2 0 2 5 3 0 15 10 5 0 35 S 7 v oA \ Oo 0 . 00 o m< oo 0o oo oa Ia oo 0 o 15 2 0 2 5 3 0 G ALACTO SE t r a n s f e r a s e { fi mol /h r /g hgb ) a c t iv it y 35 h i| 1--Correlation berwern Ik io m conium ption and (ranaferaM activity In ovarian cancer ca*c* and control. I Alii I- V M il A I IV|- KIMnN !<*Mi IVAMIAN I AN* I K ANS* IA I I I VI'I I II I I t < A IH l* lK IIN 1 i vau-gi>- ........un> 1*.'] Nif aaunitv (*,.) m - M l) in - 127) K K *t*SB* i .l - 10 1 O- 1 u i n - i tu zl S 5? i M It i2l> 2S I f 7 ' M i 2fi> til t < 1/ 24 ,2 1 IA < l2l IS M 2- 1 01 1 7 III 7 -3 S> 2 1 110 - t 5) 2 4(1 4 A III A ll n A re rclaiivt' 10 r rlrrc iK T o M ( | y IR cIctcikc *.C|p*0 Ip - O f t ll (M a n irl r ilv t) M '* i l o i 1*4 (ic m l. j H .|> - OiU) controls: this was most striking among subjects younger than 50, among those with 3 or more children, among non-Jewish subjects, and among ever married subjects. There was no correlation between transferase activiry and lactose consumption. In lig I the quadrants defined by the control medians for transferase activity ami lactose consumption contain dissimilar ratios of cases to controls with an excess of cases in the quadrant itcfincil by higlu-r lactose coiisumption and kiwer transferase aaivity. Women with lactose consumption of >11 g/day and transferase activiry o f <22 V nnnl/h per g Mb had a risk for ovanan cancer o f 2 2 (p (H D , V5"/ ('I I 1-4-5) relative to women with transferase activity o f > 22 pmol/h per g I lb and lactose consumption of 11 g/day (table IV). This finding suggested that lactose consumption and transferase aaivity, combined as a single variable, might be a good predictor of risk for ovarian cancer Fig 2 shows the distribution o f cases and controls by l./T , a variable we define as the natural log of lactose consumption by an individual < * 10) divided by their transferase activiry When the log of laaose intake was used, the distribution in cases and controls became normal < 0 ? 0 ? 0 4 0 6 0 8 i 0 t ? < 4 16 l 8 2 0 2 2* VI K l| i -- D ia irih w ilo n **f ih ra tio o f U c lo w c o n s u m p tio n lu tran sfe ra s e a c tiv ity In o v a ria n c a n c e r eases a n d c o n tro ls . 1A l l l l- V I M l AN 1HA MM I I I VSI- A* | o l 1I I MA 1lS IN V 1 IIM I M i l l 1OVAKIAN * A M I K liM A O l UK 1Y U - III Il'M tH v.K '.N 1 1 Mcm iiMiM i K u U lin c iraUAliliiM- i |illh>l l l per h 1t i l l 1 1 tan Itiid A y i* ill ivr* Sernu* in - 7 Si Minin.*-. (it Si |j>.Six-ir>.J in >2i O l l ( it IS, //<><4 y i. . i/ grotU * 0m 4li K ..-7 > 2 in 47 t i n - VII 21 l a i d 21 A i ll 71 22 a I 7i 22 Old 4i 22 1 i O A i 20101171 22 -1Ill'S ) 21 2 i A i 1 2 ill tv* i 1 JtlH M i 1 1 HMMi 1 In i Hi 1 2 III 1IA1 1 1 il l 14. I 1(007) 1 2 (HIM) *< irj.lr uii|hviIu:.I in lf> . ims n il iiiv l i l li r i s ii i j i* . 1 IIIlK l-l l 'llll. l u l Iim vu/N T h e mean (SFM ) l ./T ratio lor eases was I 17 (OlHiaiul for cofumlswasl) QH(0 04), indioiimg that women with ovarian ameer tended to have consumed more lactose relative to their transferase activity than control women (p l) iNKi5). In table v this finding is expressed in icm is ol relative risks adjusted for potentialumfounders. 'Ilie trend in risk by I ,/T categories was highly significant (p < 0-001 ). 'fable vi shows transferase activity and I ./T ratios in ovanan cancer cases according it histological type or grade o f ovarian cancer. Wumcn with tumours o l'a puriiculnr cell type or grade did not have kiwer transferase activity tr higher l./T ratios, although the deviation from control means for tra/tsferase and l ./ T was least for the endometrioid tumours. D iscu ssio n I actuse, fiMind naturally tnly in milk, is a disaccharidc composed of galactose and glucose in equal amounts. Infants can digest lactose hy hydrolysing it by ihc action of the jejunal enzyme lucusc--an ability lost with age in in*isi individuals o f non T-uropcan descent S m ic aduhs retain lactase activity and can digest lactose; but those without lactase aaivity can still ahsorb galactose by eating products in which the laaose has been partly hydrolysed, such as yogurt, or by having gut bacteria that can break down undigested lactose." Once absorbed, gala*, lose may be- used in carbohydrate side chains o f glycoproteins and lipids or metabolised for ettergy hy convention l<*glucose mainly by the 1jtrkkir pathway.11 Absence or severe deficiency o f galactose-1-phosphate urjdylrransfcrasc causes galucio- saemw." ** We huve found that women with ovarian cancer used dairy products with a higher content of prehvdrolysed laaose-- ie, yogurt and ciritage cheese-- more often than conimr women and had kwer conccntraiimts o f a key enzyme that converts galactose to gluc\rse. Risk lor ovarian cancer was strongly related to (he I J T ratio: thus, eases were more likely to have consumed more fcictose relative l<<their ability to metabolise it than controls These observations are compatible with the general model o f ovarian cancer as a consequence o f hypergonadotropic hypogonadism as evidenced by unimul and cptdcntKikvgieal studies.'"'* Ovarian tu m o u rs- mainly stromal types--devekiped in nuto ils that had received ovarian irradiaiKm" or oocyte toxins,'' *u dial had * > 70 THHl-ANCt-rt'.IUl.YH, 1989 congenital deficiency of oocyte*.14 In mice, removal o f the piruiury gland blocks tumour development.1' There is doubt about the relevance of these animat models uftfun However, a stimulus that causes stromal tumours in rodents may also promote a different type o f ovarian neoplasm in human beings. Human ovancs have numerous inclusion cysts--islands of ovarian surface mcsothchum entrapped wuhin the stroma--and most human epithelial ovarian tumours arise from the differentlaiton and proliferation of this entrapped epithelium.11 Hpiiitclial differentiation and prolifrraiKai depend on the presence of homologous stroma and its hormonal receptor statu*.14 Additionally, ovarian cancers developed in women who had received radiation for cervical cancer or who were among atomic bomb survivors.*'-JI There is also expenmental and clinical evidence that there is ` an association o f galactose consumption and galactose metabolism with hyper gonadotropic hypogonadism. Female offspring o f rodents fed with a 50% galactose diet are bom with fewer oocytes" and ovulatory disturbances develop in non-pregnant females fed on the same diet;" hypergonadotropic hypogonadism also occurs in gabetosaemte women or carriers with reduced transferase.*4 ** We believe that the above observations provide a credible biological basis for the associations between ovarian cancer, increased galactose consumption, and decreased transferase activity Nevertheless, potential hiascs must be considered. We do not believe (hat the lower transferase activity for eases is accounted for by a difference in the reading or handling of specimens: subject status was unknown to laboratory personnel Adjustment for interval between specimen shipment and analysis or year of specimen analysis did not alter the risk associated with low transferase activity. It is also noteworthy that the lower transferase activity for cases confirmed by the quantitative test was initially suggested by a different qualitative test done by a different laboratory. Docs lower transferase activity precede ovarian cancer, representing a possible genetic marker for increased susceptibility, or ikxm lower transferase activity follow the disease, representing a response to the tumour or to its treatment? If the deficient transferase activity is a consequence o f tumour development, the lowest measurements might be expected in cases with higher grade tumours or perhaps in women with a particular cell type. Neither finding was observed. Additionally, because transferase activity is measured within the red blood cell and not in the scrum, a tumour marker is less likely to be identified. Our finding of lower red blood cell galactose transferase activity should run be confused with reports of raised scrum galactosyl transferase as a possible tumour marker in women with ovarian cancer.1' Galactosyl transferase, which transfers galactose to terminal oligosaccharide chains during glycoprotein synthesis,11uses the priiduct from the galactose transferase reaction: genes that code for the two enzymes arc both located on chromosome 9 * Thus, the possibility that raised serum galactosyl transferase is due to the lower galactose transferase activity and represents a related genetic marker rather than a tumour marker should be investigated. We do mu believe that eases preferentially recalled dairy product use since ils a sso c ia ted with ovarian cancer has not been previously reported. Two case-control studies of dietary factors and ovarian cancer did not find an effect of milk consumption on ovarian cancer risk;**-*0 but these studies did not examine other products such as yogurt, nor did they calculate tolal lactose exposure. Could other factors associated with both ovarian cancer and lactose consumption account for the association? The factors thai might influence lactose consumption--ic, age, religion, marital status, education, and oral contraceptive use-- may also be associated with risk fur ovarian cancer, and therefore we adjusted for these factors in our analyses. Dietary factors may also be confounders Because diets high in lactose may also be high in animal fai, it may be difficult to state which of these components o f the diet is potentially harmful. Using more limited dietary histories, we previously concluded that the principal assoctaikai was with fat." That galactose is the harmful component rather than fat cranes from our observation o f the present study that the ability of galaciosc consumption to predict risk for ovanan cancer is improved when the L /T ratio is considered. Furthermore, the principal dairy products which distinguished cases from controls were yogurt and cottage cheese, fnbds with little fat but high lactose Although this observation might be due to intention to lose weight among cases, excess risk was nut seen for use o f skimmed or ice milk, which are also low fat but lack a high content of prehydrolysed lactose. Additionally, adjustment for weight or height did not alter the increased risk associated with regular use o f yogurt and cottage cheese. The production of both yogurt and cottage cheese usually starts with skimmed milk. 1-actuse is added in the form o f nonfat dried milk followed by laciobacilli which start the fermentation process and partly hydrolyse the lactose into its component sugars." In yogurt, up t o 40% of the lactose rs already hydrolysed so that even lactase deficient individuals can absorb galactose." The specific and significant association between ovarian cancer and yogurt consumption is n o u b k because its consumption in the U S has more than quadrupled since 1970.** The issue o f gastrointestinal absorption of lactose was not addressed in this study. Generally, individuals who are lactase deficient and cannot digest lactose have gastrointestinal symptoms from dairy produces and avoid them, whereas those with lactase persistence arc symptomfree and continue to use the products throughout life. Our observation that greater lactose consumption was most striking for older eases may poini to a higher prevalence of lactase persistence among women with ovarian cancer. World wide, ovanan cancer nsk is sirungly correlated with lactase persistence and per caput milk consumption, further epidemiological evidence that lactose rather than fat is the key dietary variable for ovarian cancer ** This study provides a basis for further research on galactose consumption as an environmental factor and galactose-1-phosphate uridyltransfcrase as a genetic factor in the aetiology of ovanan cancer. Further human cpidcmiokigical srudics and relevant animal experiments arc warranted before any public health recommendations can be made. If our findings arc confirmed, however, avoidance of lactose-rich food by adults may be a way o f primary prevention of ovarian cancer, particularly in those women with low transferase activity. We acknowledge ihc gcTtrruut panuti tattoo and o f Uk Mkiw mg ilinkSani and imiiiuiaaia in the greater Boalnn arc* l) r Pjnanucl Hriedman of (he Beth I macI lloapital. D l Robert Knapp til the Dana la rbcr Cancer Im iirulc, D r George H. Hutch non o f (he Harvard School oilU N ic Health, D r Joaeph K H u rd (the I ahey Clinic, Richard Clapp >>fthe Mavttchukctti Cancer Regitiry, D r George S. Rh hardw o of tlic M n w c h u K tii General Hospital, D r Phillip G . Stubblefield uf the M ount Auburn HtMptlal, D r Harvey l.evy and M * Jane Simnxtni of the newborn wrccning laboratory of the M u ia th u tc iu Stale I aburalnry, D r J.axiihaii N iM l t,f iltc New England Dcaonncta H io p n il, D r David M antni of Salem Hospital, D r James Whellun of Sr ElirahcthS lltn p iu l and D r Stephen Curry of T u fa -N e w n i b lANCKT, JULY, 1989 71 b i | l t n d M edical O n ic r T be w ipcrb icctinka l iM n u n c c 4 M i A m y Q h e n , M I - iilr t F o o te r, M i D w n H a r v e y ,M i) u J y G cIIct, m d M ib m ily l.ub lm j> jrm lifu lly ackn<>wlrd|fcd. I t i l l n u d y wm lu p p o fic J by *rw n KOI O A 421106 from ihc National Cancer Iru u iu ic . Cixmpondcncc ihouU be n id m ic d to D . W Urifham and Women' H oapiul. Obuctnc* and Gynecology Epidonioiugy Center, 221 I .ungwoud Avenue, bowon, Maaaachuacm 021 15, USA . P R E 1X )M IN A N C E O F B O K K E1.1A B U R G D O R F E R I SP E C IF IC H C:ELLS IN C E R E B R O SP IN A L R .U I D IN N E U R O B O K R E U O S 1S S h a h id B a u ; T o m a s O i sson H a n .s L i n k Otparimrni o f blturvlogy, Koritlnuka m ntuiet, Hu-Llu^r Umveruiy Hoiptnil, S - 141 A6 flmUitigt, Stockholm, S u xd t'i Ki-HiKhNOiS 1 <Un*nnW> I l.m amal rJulana. 1 n ( W n l , ! > M I I 2 C n tn a IJW, 'rich WR. D m nnkw na o i ovarian canon r phcanM 7 Naif Caww b m i9B, 71. 717-21 1. S o w Sp. Scully KE. Sotm l-H Im 9 Htoankec <ypwi oi 1*7 4 T d l * C Sampan 1-, Sampta M |. antiumuaaiva fuoJ frequency Am J fc ^ v --tV ISWi; 122: 1-67. . W ilm W C Smtoa L. Utomrn M L . Sompfc. MJ, Homer B. Hcnnekma CM, Spaiarr FB- The* dtc pan Am J bpUmmei I9M, I D . IM 6 Ikw Iare.BakaidaM A A umpb apn Hjaanmctaat for calacroaairaa. ^ Cm O n 1*66; 41i I 7-41. 7 Laa JfcS. N f W fi S m m a ii MdmMfun far Ihr fer^aypm* el i M <ww t and ^ h ra -a r-l tdwapbalc w v frl tramfcrmc C k C t a i A u 1 9; 114. )l-6 Kendall M , Stumi A. T V advanced Btcor). o4MauaQca, *l 2 New Yeefc Mm i i Uwi, 1*7* 9 L>ay N il, By* o r I <v.ve tiypuahem In caac-<nnt>ul e i d a cqunviciK ol M a a d H n n a M l w a i k ia a l b f e i i u t iM i H v . 1*7*. Ih a Z IU M u r . lajknahi M . Juaaeki ] , Laaauaie K, Pyiaala K Rwnaive tohemanor el ada* type latttne mdabwrptaei I mum 1*71, -26. I I. Savawa. I>A. Abwi h. A n u a A. S n u* OK. L m n M O I trim yafun. paecufiord yefurt. aaaei aod^duku .raft, lanuC -dfleiail n d v d a l i A m J ('Jm Nm r IWM; 4tk 1219-2 12 l^fcar LP I V aurymaoc timaAxmancn ol undine ifc p k a p lw gfuane Im u a p t v M . derbanve A m h H V . Hwp*yi |9i; IBa-40 1C (a la a a e nwtabnfwin. h a rd e iiy dtlarti and dtrfa Butman O. Ilotton JB. PdUKxll CA. n k Interned d a in k ii e l cartaAyOrair mnafaoiim H d o n m UMw m ) l*ar* P im , I W 14 h u d ). H annacnll JS N a y tM a dneaaaa axteHng m g owe m b a la l x* (encral umlianun with X-ray Am J Carnot |96; 21: *6-9 1 Huwvtl JS, Marrharw J, Orr JW T*r v*h*Tin ,4 uw aorum our, m with 9-10 dfcnrthyt I 2-bm/anthracenc Ur J Cmom l*4; It 6tv-46 16 MuryUiy hJ>.Kuaacll KS CXanan tumaipmsa lilkn>mt>Tik.-drmcn <4trim crlla m hybnJ m u A m I Jmb o ( V n 1961. IV 77*-. 17 March* ) T V eflccl o l hpr**yir" Tun the Jrvefcymmi ol ovarian nanram at trot im ard v a h linnaatiylhenianthrarme Ar 7<iaar* 1961,' I II 921-27 I* Hadiaa*lievK SV The padacneMi u l ovanan mchnKn cyan avl eyatanaa. Otiwi Ir'mrtal 1*77, 49. 424-2*. 19 tunha CM. Bigaby KM, Caa*e PS. S niuk <egau t -AI n * /rr I*. |7i 1 17 Y Sunmal epjlhell irucraclaeii ai 20 UayNK.ihaccIO SrunrvJ I ARC PubUartaa. no 2 Wurld Heath Orgmiaaocn. 19*1. 21 Darby SC, NataAena K. ICaauH ApmdMMMtyiacItmocrmurtaluy anaeigauarK bumb aurvtvun id paoenli with ankylimng ifaeidylnn P 'm X-ray therapy Kadaaoon hflecu ReaCaivh b u n d a n n , Teehrucal fry*in 4-A4. Hiruahwna 22 (h e n V T .M tm e in R .K ^ e n b a -n l .h u tu iH .S A u tm m JI) Rnk>aiceim..ay*r number hdfcnvwig prenatal tafaiaun *o dart high 1 gdaouac Snmrr IW l, 214: II47-47 2 Swam W | , M artian DM Caiactow inhatatacne l avuUlam n mice f-Vu/ .Vim<l*MS; 4*t 22-26 24.1k fr w p i D , V i r r o -1lilt l>. h i l l li l thanan laituar in gabfimacmra /a r u i 197*. ii. 1197 2 Kaufman I K. K.lc M O, IVamrll CiN, lauhehmam U. Manh C Koch H H)pig<UcliUijpa h y |*i> n a lia ii n (cmaWpaurrui with p k lin c in a N h.<*rlJ M ol 19; 04. 9*4 2 (jam ar I * , IV k iw HI., Hatbam R|., Ng 'I i CMacmae-l r* -T *- undyt (ran*feme actmiy aaarawted with age M moaipauie ami rrjHrahjcrivr hniory fm til Stm t/ l*ffr, ! 609-1 , Oaneryac SK, Bcaa J| Urnknr S-dipbaphair-gabcruar rtafnaar activity m ihc ovanan cancer patient C w ffn m V, ( ] i | H , Kirath IK, llodia CP Tbe human m J in m w m r 9 at band p i J 'm r e f r A A M lim i 19*6, 12.611 M> 29 Uyar* T . Marlhall J. tWaham S. Mental .warm M A la r e m w l itudy *4 deary arvJ raai-dwtary iH im In ovanan carver J Nj i/ (ua>> but I9n|, 7|i 01-06. La Vcrxhwt...iMvarWA, Nrgnfc.el aJ l)m y lati.vr arwlchr n il of cpethrludovarian (M tca .J Noiltmom but 1967, tot-ert. I.(c n n L)W, W dJi VRJ. H a d u v a CM, it al Oanary animal fat a. relnavi lo ia m c a w ir a l Oaurt ( . * . 1 IWU. 61: Ml Ml 12 M .9 llygiciw. Ccncva W.aU IlaaMl t V*..va.cn, l*C I) M .9 Kacet Wmhaiglia. IX !' M .9 liaJtrwry h w d r l n i. IWJ 4 Cnencr I W Ikciaac pcfutinac and malt ceieunipeav. drtmiunwiiv .4 n'anan cancer r id Am J f>a/rww.Wiu, prrwi S u m m a r y A niimccliukisc immuiupoi assay that allows the aninting o f culls secreting IgG, IgA, nr IgM antibodies to Borrtlia burgtLrrftri was used to compare B cell response lo B burgdorferi ai the cellular level in cerebrospinal fluid ((^SK) and blood from patients with ncuroboneliosis with that in patients with aseptic meningoencephalitis (AM ) or non-inflammatory neurological diseases. 13 o f the 14 patients with untreated ncuroborreliosis had C S F cells secreting IgG antibodies to B burgdorferi (mean 17 cells per I04C S F cells), whereas 8 of 12 pabents examined had cells secreting IgA antibodies (mean 6 cells) and 10 o f 12 had ceils secreting IgM antibodies (mean 6 cells) per I04 CSF cell. IgG antibody producing cells predominated except in 2 patients with mainly or only IgM secreting cells. Cells scctcling antibodies to B burgdorferi were rarely found in the Mood and then at very low numbers, which reflects preferential compartmentalisa!ton o f the specific H ceil response to the CSF. T h e cells were not detectable in C SF or blood from the two control groups. Evaluation o f humoral immunity at the cellular level is a novel approach to the detection and localisation o f immune events in neuroinnanimatory disorders. Introduction Lyml disease may cause mcningoradiculiiis, meningitis, or neuritis o f one or more cranial nerves.1 I .css frequently, there is clinical evidence o f neuronal lesions in the central nervous system (C N S).2C.crebrospinal fluid (CSF) findings characteristically consist o f mononuclear pleocytosis and blood-brain barrier (BHB) damage (both o f which may be pronounced), and evidence o f central nervous system synthesis o f IgG , IgA, and especially IgM *4 A history of hefc-bite and o f erythema migrans may point to the diagnosis, which is usually verified by demonstration o f IgG and IgM antibodies against Borrelm burgdorferi in scrum, but sometimes in CSF' only.* Since levels o f specific antibodies present in free form in body fluids such as the C SF may be influent cd by a variety of factors, including binding to target antigens and altered antibody catabolism, we have proposed that counting o f cells secreting specific antibodies o f different ismypes might reflect the B cell response.*' Patient* and M ethods Paired pecimcn of CiSF nod peripheral Wid were obtained from 14 consecutive patients (9 females) with elimuil and CSF findings (table l) that suggested neuroborrcIR'ktx. II pauenis had serological evidence in the scrum and CSF of B burgdorferi infection, and 3 had such evidence in die C.SF only. M.n>i patient* had prorkunccd mononuclear plcoeyi<>sis m the CbF, and <ly 1 had a C S F cell iunt which waxborderline Albumin and IgG were measured in CSF and the uirresponding plasma ample by F n m u r v and Stuuutt Coeyn** * 1M4 The Aa n e a fotti*? Society ' -.LS ' ' *-.* r -_*r". i. . . - M odem trends T ;,vx"-<rB6rard.BLWiallach, MLD.Assodate1Editor ' ' W s t e i . - 94-TA-50 Voi. M . N o. 3. S i a t o 19*4 Printed an arid-free pepoin U. S. A. Epidem iology, etiology, and fertility drugs in ovarian epithelial carcinom a: W here are we today? M 35 Zeev Shoham , MJD.* . Department of Obetetnc* and Gynecology, Kaplan fjoepitai, Rahovat*Itraeft O b jective! To review studies that have exam ined the epidemiology and etiology of the develop m ent of epithelial carcinoma of the ovary. D a ta Id en tifica tio n : Important published studies related to the topic were identified through a computerized bibliography search. C onclusion: A review o f the literature reveals that the etiology of epithelial ovarian cancer is probably m ultifactorial and that genetic, environm ental, hormonal, and viral factors appear to be directly or indirectly related to the development of the disease. An attempt to im plicate specific agents has not produced conclusive results. However, based on large epidemiologic studies, it seems that there is a clear trend of decreasing risk with increasing number of pregnancies, deliveries, use of oral contraceptives, and the duration of breast feeding. An increased risk was found to be asso ciated with ovarian dysfunction leading to infertility and exposure to asbestos and talc. The recent observation that infertile women who used fertility drugs might experience an increased risk for the developm ent o f epithelial ovarian cancer should be examined very carefully because of the small number of patients in the study, lack of appropriate information about the type of infertility, drugs used, dosage, and duration of treatm ent. Because there are no screening tests that are consistently accurate enough to detect ovarian cancer at an early stage, translating the current information into disease prevention requires careful clinical evaluation with a routine follow-up of patients at risk. Fertil Steril 1994;62:433-43 K ey W ords: Epithelial ovarian cancer, infertility, oral contraceptives, breast feeding, preg nancy, fertility drugs, asbestos Ovarian cancer is the sixth moat common malig nancy in women in the United States (1). The mor tality for ovarian cancer in Northwestern Europe and the United States is approximately 7.3 to 13 per 100,000 women, making ovarian cancer the most frequent cause of death from gynecologic malig nancies and the fifth leading cause of death from cancer in women (2). The cells of origin of ovarian epithelial cancer in human beings, which comprises some 94% of cases of ovarian cancer in the United States (3), are gener- Received March 4. 1994. * Repnnt requests: Zssv Shoham. M.D., Department of Ob stetrics and Gynecology, Kaplan Hospital, Rehovot 76100, Israel (FAX: 972-8-410991). , t Affiliated with the Hebrew University, Hsdasaah Medical School. Jerusalem, Israel. Voi. 62. No. .7, September 1994 ally agreed to be from the surface epithelium of the "vary (4-6). In the adult, the ovarian surface epithe lium arises after invagination of the coelomic m sothlium cover of the embryonic gonadal ridge and consists of cuboidal to low columnar pseudostratifled epithelium (7). This msothlial layer is set off from the hormonally active stroma by a basement membrane and a thin connective tissue layer, ho mologous to the tunica albuginea of the testis (8). The age-specific incidence rate increases from 2/100,000 for women in their 20s to 55/100,000 at the age of 70 (1). From reports by Barber (9) and Coppleson (10), it can be estimated that approxi mately 1 out of 424 women will experience ovarian cancer before age 40, and, according to the Third National Cancer Survey, it is predicted that 1 of every 70 newborn females will eventually have this disease (1). The reported incidence of ovarian S bobam P athoH tnenu o f ovarian cancer 433 cancer during pregnancy varies from 1 in 6,429 (11) toriaL Attempts to implicate specific agents in th to 1 in 25,000 (12). In our experience the incidence etiology of human ovarian cancer have not J"~ was 1 in 47,115 deliveries (13) considering the rarity duced conclusive results. Epidemiologic and ei of such events during gestation. logic studies have identified a group of risk factors Ovarian cancer typically remains clinically silent that appears to be directly or indirectly related to until it is far advanced; morbidity and death are human ovarian cancer. These factors can be generally sequela of intraperitoneal carcinomato grouped into three main categories: genetic, envi sis. Therefore, epidemiologic studies of ovarian ronmental, and hormonal. cancer have attempted to identify risk factors that might help define groups of women at unusually high risk as it has become clear that improved pre The Genetic Hypothesis vention and detection of these groups of women are The occurrence of familial ovarian cancer has necessary. been increasingly recognized (20-23), emphasizing Attempts to implicate specific agents in the etiol the tiologie influence of genetic factors in a pro ogy of human ovarian cancer have not produced portion of cases. Familial ovarian carcinoma is conclusive results, and review of the literature re transmitted as an autosomal dominant trait (20). veals that there is currently no evidence to impli After analysis of 10 families in which 25 women cate any single tiologie factor and that genetic, en developed ovarian cancer over three or four genera vironmental, hormonal, and viral factors appear to tions, Lynch et aL (24, 25) concluded that the dis be directly or indirectly related to the disease. ease is heterogeneous, with at least three genotypes Recently, in a series of publications by Whitte- predisposing to distinctive hereditary syndromes: more et aL (14-16), the authors described findings [1] site-specific familial ovarian cancer syndrome, for invasive epithelial tumors in white women where women in these families are at risk of devel based on a collaborative analysis of data from 12 oping ovarian cancer only; [2] ovarian carcinoma in case-control studies of ovarian cancer conducted in association with carcinoma of the breast (26, 27), the United States. Accumulating data from 3 (17 where a woman may have 50% risk for ovarian 19) out of the 12 studies, an increased risk for inva cancer if her mother or sister had breast cancer sive epithelial ovarian cancer was found in women and/or ovarian cancer; and [3] cancer family who used fertility drugs. By contrast, infertile drome (Lynchsyndrome II), characterized by her^^- women who did not use fertility drugs experienced itary nonpolyposis colorectal cancer with proximal no increased risk. This new epidemiologic associa colonic cancer predominance, endometrial cancer, tion between fertility drugs and ovarian cancer and ovarian carcinoma (28, 29). Consistent with raised much discussion, because, if verified, it could autosomal dominant inheritance factor, the risk for have far-reaching implications for current clinical ovarian cancer among first degree relatives may be treatment as well as public health. as high as 50% in these families. -S i As a result of the discussion and disagreement These syndromes are characterized by signifi with the results of Whittemore et aL (15), it was cantly early age at onset and an excess of multiple found appropriate to review the current literature primary cancers. A review paper by Hientz et aL (2) regarding the different epidemiologic risk factors summarizes 140 families who have been reported and the suggested mechanisms for the development with familial ovarian cancer. Out of the 140 fami of ovarian epithelial cancer, as well as to summarize lies, 94 were reported by the Familial Ovarian the new epidemiologic results and the comments to Cancer Registry (21). In these reported studies, in these findings, as discussed in the literature. 131 families, 321 women developed ovarian cancer the number of observed generations varies from two . r; EPIDEMIOLOGIC FACTORS AND MECHANISMS OF ACTION to four per family. A tendency toward earlier onset, with a mean age of 47.7 years compared with 59 years for epithelial ovarian cancer in the general .-V* The events leading to malignant transformation population (30), was noted. In addition, a predomi A i -, '-M and progression of ovarian cells are not known. It is nance of a serous type of tumor with a trend toward a likely that multiple events are necessary for trans more poorly differentiated adenocarcinomas and a formation. In the present state of knowledge, it ap high proportion of bilateral disease was described. * pears that the etiology of ovarian cancer is multifac The hereditary phenomena in these families is con 4 3 4 Shohmm PathoMtrmu of ovarian carterr Ftrtiiity and Stm hcy sistent with an autosomal dominant gene with vari able penetration, transmitted through man or woman (24, 25, 30, 31). The Familial Ovarian Cancer Registry, which was established at Roswell Park Cancer Institute in 1981, assessed through December 19901,336 cases of ovarian cancer in 587 fam ilies (32), which made it clear that familial ovarian cancer is not a ra n occurrence. The exact pathogenesis of these syndromes is not understood. Three theoretical explanations are pos sible (24): [1] different genetic susceptibility of ovarian tiseus to carcinogenic influences, [2] a ge netic defect that induces a hormonal imbalance that may be potentially carcinogenic, and [3] a ge netic defect causing an altered immunologic re sponse resulting in ovarian cancer. The most reproducible and specific structural changes in ovarian cancer are short-arm deletions of chromosome 3 (del[3][p21-pl3]) and long-arm deletions of chromosome 6 (del(6][ql5-p23]) (33). Similar abnormalities of both chromosomes have been seen in several other malignancies, with breakpoints clustering near the locations of some oncogenes (33). Variable number tandem repeat probes were used to analyze DNA from a bank of over 50 primary ovarian tumors and peripheral blood lymphocyte matched control DNA. Allele loss was seen on the long arm of chromosome 17 in over 80% of malig nant ovarian neoplasms (34). The finding of sub stantial allele loss on 17p has been confirmed by others (35). Mapping of the short arm of chromo some 11 identified allele lose in 45% of patients. It can be speculated that allele loss on chromosome 11 may represent deletion of tumor-suppressor genes (36), which appear to inhibit cellular expression of the transformed phenotype (37,38), contributing to the development and progression of ovarian cancer. Crickard et al. (39) reported two patients with bor derline serous tumors illustrating trisomy of chro mosome 2,7 and 10. Trisomy of chromosome 10 was also found in five of six serous borderline tumors in comparison with normal karyotypes seen in 18 be nign cystadenomas (40). Epidemiologic evidence raises the possibility of defective cellular repair after ovulation as a major risk factor in ovarian cancer. The molecular evi dence suggests that this repetitive proliferation al lows promotion of cells already bearing allele loss, and, by inference, those carrying inactivated tu mor-suppressor genes will lead to uncontrolled cell division and malignant transformation. Voi. 82. No. 3, September 1994 Support for the genetic hypothesis was given by the observation of decreased serum activity of a-Lfucosidase in the serum of ovarian cancer patients (41). This observation, along with the report that women with ovarian cancer may possess alleles that lead to a deficiency of a-L-fucosidaae activity (41), suggests that the enzyme deficiency may be heredi tary and associated with an increased risk for the development of ovarian cancer. If the current knowledge of the genetic epidemiology and mecha nism of pathogenesis are to be translated into dis ease prevention, more attention and effort should be invested into the identification of women at ge netic risk. Environm ental Factors and Carcinogens The fact that the highest incidence of ovarian cancer is found in industrialized countries suggests a relationship between this tumor and environmen tal factors associated with industrial chemicals and diet. This hypothesis has been strengthened by the observation that the incidence of ovarian cancer in Japanese women who were born in Japan and emi grated to the United States increased (1, 42, 43). Epidemiologic, experimental, and clinical data seem to focus on a possible relationship between ovarian cancer and materials of the talc-asbestos group (hydrous magnesium silicates) (44-49). The dear relationship between asbestos exposure and pleural and peritoneal mesotheliomas, the similar ity ofovarian cancer to mesotheliomas, and the abil ity of talc to enter the pelvic cavity are in support of this theory. Graham and Graham (47) found particles similar to talc in ovaries with ovarian carcinoma. They also were able to induce ovarian neoplasms in Guinea pigs with asbestos, which suggested that ovarian cancer could be related to asbestos exposure. Newhouse et al. (50) found a slightly elevated fre quency of ovarian carcinoma in female asbestos workers. Cramer et al. (44) found that 42.8% of ovarian cancer patients regularly used talc powder in the perineal area, either as a dusting powder on the perineum or on sanitary napkins, in contrast to 28.4% of a control, yielding a relative risk (RR) of 1.92 (P < 0.003) for developing the disease. In addi tion to this, women who had regularly engaged in both practices, powdering the perineum and sanitary napkins, had an adjusted RR of 3.28 (P < 0.001) compared with women with neither ex posure. S h o h a n Pathotferm u of ovarian cancer 4 3 5 It has been suggested that cosmetic talc powder (46) on a dusty contraceptive diaphragm can enter the vagina and comes in contact with the ovaries after passage through the uterus and fallopian tubes (51). That this route ofcontamination is feasi ble was shown in 1961 by Egii and Newton (52) using carbon particles and in 1981 by Venter (53) using labeled human albumin microspheres. If ex posure to talc on contraceptive devices is associated with increased risk of ovarian cancer, then observa tions of decreased risk among parous, lactating, or pill-using women may somewhat reflect their non usage of such methods. The "pelvic contamination" theory proposed that environmental carcinogens might ascend the genital tract and act upon the surface of the ovary (51). Talc has been proposed as one such substance (44, 46) as it was suggested that the talc (perhaps contaminated with asbestos) alters the responsive ness of the ovarian epithelium, which ultimately proliferates autonomously. In an experimental model, in mice, ovarian tu mor was induced by administering carcinogenic polycyclicaromatic hydrocarbons, like 7,12-dimethyibenzfajanthracene (54), which undergoes hor mone-dependent metabolism to reactive interme diates in the rat ovary (55). It was also shown that monooxygenase activity in proliferating hu man granulosa cells caused 7,12-dimethylbenz[a]anthracene to metabolize under the influence of go nadotropins and raised steroid concentrations (56). Uniting the above observations into a theory, it might be suggested that when granulosa cells are in the phase of rapid growth under gonadotropin and estrogen stimulation, microsomal cytochrome P450-dependent hydroxylase systems convert 7,12-dimethylbenz(a]anthracene to reactive epox ide intermediates, forming covalent linkages to protein DNA that may induce follicular cell destruc tion or transformation to a cancer cell (56). One of the major dietary differences among peo ple in industrialized countries is the high intake of meat and animal fat, which has been reported to be associated with an increase ofovarian cancer in sev eral studies (57, 58). Using published data, largely from the 1970s, Cramer (59) compared ovarian cancer incidence, per capita milk consumption, and population estimates of lactase persistence in 27 countries. (A nutrient that distinguishes milk prod ucts is lactose, a disaccharide composed of galac tose. The ability to digest lactose after childhood, called lactase persistence, varies widely geographi cally and racially.) Significant positive correlations . were noted between ovarian cancer incidence, p*' capita milk consumption, and lactase persiste! In 1989, Cramer et al. (60) proposed that increased dietary galactose consumption and low serum levels of galactoee-1-phosphate uridyltransferase, which prevents the degradation of galactose to glucose, may be linked to an increase of ovarian cancer. To explore the possible importance of the animal Cat content of milk in milk-ovarian cancer associa tion, a case-control study of303 ovarian cancer cases and 606 age-matched nonmalignant-diaeaae con trols sean between 1982 and 1988 was performed (61). Total frequency of usual milk intake was not associated with increased risk. Usually drinking more than one glass of whole milk daily relative to never drinking whole milk was associated with a RK of 3.1 (96% confidence interval [Cl] 1.8 to 5.5). Con sumption of reduced-fat milk was associated with reduced RR. Among persons who reported drinking milk regularly, persons reporting drinking only whole milk were at increased risk (RR * 2.6,95% Cl 1.7 to 4.0) relative to persons who drank only skim milk or 2% milk. Therefore, it. was suggested that milk is related to ovarian cancer when it is the vehi cle for animal fat consumption (61). Hor mones sad Related Condition* Because it was noted that women with ovanan cancer had an increased incidence of nulliparity, lower rate of pregnancies, increased incidence of miscarriage, and high rate of infertility, it was sug gestedthat malignancies ofthe ovary, especially epi thelial carcinoma, are caused by endocrine and physiological disorders associated with ovulation that impairs the ability to conceive. According to this theory, pregnancy, oral contraception, and lac tation, all of which suppress ovulation, protect against ovarian cancer. In this section, the associa tion between conditions that suppress ovulation are reviewed along with other conditions associated with infertility. Parity Many studies have shown that pregnancy confers a high degree of protection against the development of ovarian cancer compared with nulliparous women (50, 62-68). Most of the authors agree that protection increases with the increasing number of term pregnancies (50, 63,66-68). A pooled analysis of three European case-control studies (69), provid- 4 3 6 9 hoham Pathoittntau of ovarian cancrr Ftrtility and SUt ing a total database of 1,140 cases and 2,724 con trols, shows that the risk for developing ovarian cancer decreases with increasing number of births, and the trend in risk was significant (P < 0.01). In comparison with nulliparous women, those who re ported four or more births had a 40% reduction in risk of ovarian cancer (RR of 0.6,95% Cl 0.4 to 0.8). In each study, as well as in the overall database, an inverse association between the number of abor tions and ovarian cancer was noted. The estimated RR for those who experienced two or more abor tions was 0.7 (96% Cl 0.6 to 0.9). This study also presents data regarding the protective effects of full-term pregnancy upon the age at which ovarian cancer occurs. An estimated RR of 1.4 (95% Cl 1.1 to 1.7) was found for age 35 or more at first birth compared with age 25 or less. The high risk of first delivery at age 35 is comparable in this study to the RR of nulliparous women. The influence of age at the time of delivery and the association of ovarian cancer received support from the study by La Vec chia et aL (70). In their study late age at first birth was associated with increased risk in comparison with those who had a child before the age of 22 years. The RR for those who first gave birth at age 22 to 24,25 to 27, and 28 or more were 2.7,3.2, and 4.0, respectively. Comparable results were found in a large prospective study in Norway (71). The risk for ovarian cancer decreased significantly with in creasing parity. A parity of one or two was found to be linked with a risk reduction of 10% to 20%. The estimated odd ratio for women with five or more births compared with one was 0.46. Lactation There is no consensus in the literature regarding the association between breastfeeding and ovarian cancer, but the majority of studies observed a de creasing risk. Four studies found no association (18, 67, 72, 73). In another study, a 50% reduction risk was found in association with a history of lactation (74), and in others, lactation was found to be a pro tective factor (15, 17, 19, 75, 76). Based on the re cent studies by Whittemore et al. (15, 16), each month of breastfeeding was associated with an overall risk reduction of 0.99 (P < 0.01). It was also noted that the risk reduction per month of breast feeding within 6 months of delivery exceeds that for subsequent breastfeeding. B irth Control P ill The hypothesis that the use of an oral contracep tive (OC) reduces the risk of epithelial ovarian cancer has been consistenly supported (50, 64, 66, 76-78). The risk of epithelial ovarian cancer in re lation to the uae of OCs was evaluated in several case-control study (50, 66, 76). Detailed analysis by Weiss et aL (76) revealed a significantly decreased risk for ovarian cancer at age 35 to 55, especially if OCs had been used for more than 3 yearn. A case-control study by Cramer et aL (79) of 144 white women under the age of 60 who had ovarian cancer and 139 white women under 60 who were selected from the general popula tion revealed a decreased risk for ovarian cancer associated with the use of OCs in subjects 40 to 59 years of age at the time of the study. The RR ad justed for parity waa 0.11, with 95% confidence lim its of 0.04 to 0.33 (P < 0.00001). In a case-control study reported by tbe Centers for Disease Control (80) regarding cancer and ste roid hormones, the major finding was that women who had ever used OCs had a significantly lower risk of ovarian cancer. The decrease in risk was di rectly related to the duration of OC use. After a minimum of 3 months, the RR declined from 1.0 to 0.6, and it declined further to 0.4 for women who used OCs for more than 5 years. In a report (75) that evaluated the results of data gathered from a case-control study of 546 women 20 to 54 years of age with ovarian cancer and 4,228 controls, it was found that women who had used OCs had a risk for epithelial ovarian cancer of 0.6 (95% CL 0.5 to 0.7) as compared with those who had never used them. This protective effect was seen in women who had used OCs for as little as 3 to 6 months, and it continued for 15 years after OCs were stopped. Furthermore, the results indicated that the use of OCs even for a few months reduced the risk of epithelial ovarian cancer by 40% for women 20 to 54 years of age. The effect probably takes 5 to 10 years to become apparent, but it per sists long after the use of OCs ends, and protection exists regardless of the drug combination used for contraception. Looking at the association between nulliparous women who usually are located in a high risk group revealed low risk for developing the disease after the uae of OCs. Their risk for develop ing the disease was even lower than that for parous women. A protective effect of OCs among nullipa rous women also was found by Rosenberg et al. (77) and Cramer et al. (79) but only in very small sam ples. Gwinn et a i (81) estimated the effects of cumula tive months of pregnancy, breastfeeding, and OC Vol. 62. No. 3. Soptambor 1994 S h o k a a PathoKtnmu of ovarian cancer 43 7 use on the risk of developing epithelial ovarian cancer, following the hypothesis that all three con* ditions cause cessation ofovulation. They used data from the Cancer and Steroid Hormone Study, a multicenter case control study. Estimated RR of epithelial ovarian cancer was 0.6 (96% Cl 0.5 to 0.8) for women who had been pregnant, 0.6 (96% Cl 0.5 to 0.8) for women who had breast fed, and 0.5 (95% Cl 0.5 to 0.7) forwomen who had used OCs. Logistic regression analysis revealed a strong trend in de creasing risk of epithelial ovarian cancer with in* creasing cumulative months of pregnancy. In con trast, a marked reduction in risk was associated with having breast fed or used OCs, whereas the decrease in risk from additional months of either of these exposures was less than that for pregnancy. Possible Mechanisms ofPathogenesis. Several hy potheses to explain the mechanism of this biologi cal action have emerged. This is based on epidemio logic observations that increased parity, use of contraceptive pill, and breastfeeding, the three ma jor causes of anovulation during reproductive life, and to a lesser degree repeated miscarriage, early onset of menopause, and late menarche give some protection (50, 67, 75, 77, 79, 82-85) against the development of ovarian cancer. Fathalla (86) was the first to suggest that the key step in the pathogenesis of epithelial ovarian cancer is the disruption of the ovarian epithelium, occur ring at the time of ovulation, leading to malignant transformation, named the "incessant ovulation" theory. Further support for the ovulation theory came from studies showing that in the absence of ovulation, before puberty, and in patients with go nadal dysgenesis, ovarian neoplasms of surface epithelial origin are extremely rare, whereas germ cell and mesenchymal tumors occasionally arise (87, 88). Zajicek (89) have further hypothesized that epithelial tumors of the ovary may arise as a result of epithelial inclusions of the surface epithe lium of the ovary, which frequently occur after ovu lation. Pathologic review of the epithelial inclusion cysts found in the ovary at autopsy suggests that, after follicular rupture, the follicular cavity can be come lined with epithelium from the ovarian sur face, from the fallopian tubes, or from the endome trium (90). Cramer and Welch (8) have tried to unite the different etiologic factors into a model of tumorogenesis that attempts to reconcile the existing epi demiologic and pathologic data regarding ovarian cancer. According to this model, the first step to ward malignant growth is the formation of an in clusion cyst by entrapment or invagination of sur face epithelium in the ovarian stroma, disruptin. the connective tissue that separates the surface epi thelium from the ovarian cortex and allowing the cells to be in close proximity to steroid producing cells. This most probably occurs as a sequela of ovulation or indirect stimulation through inflamma tion due to chemical irritants or carcinogens. The second step is direct or indirect stimulation of the included epithelium by extragiandular or in traglandular estrogen, growth factors, follicular fluid, or gonadotropins resulting in differentiation, proliferation, and sometimes malignant transfor mation. However, the final stimulus to malignant transformation remains unknown. Support for the incessant ovulation hypothesis comes from studies looking at the side of ovulation that has been found to occur significantly more of ten in the right ovary (65%). In a study performed by Potashnik et aL (91), 64% of ovulations were detected in the right ovary (P < 0.01). Based on this observation, Cruickshank (92) raised the hypothe sis that ovarian cancer should be found more fre quently in the right ovary. In a study of women with epithelial ovarian cancer in Scotland, he found sig nificantly more tumors on the right side than on t* * left (59% vs. 41%), the 96% Cl for tumors on i right being 51% to 67%. However, an important ob servation of this study is that 60% of the patients with left-side tumors presented with disease at an earlier stage, in contrast to 58% of those with rightside tumor, who presented with disease at a later stage. Going further with the assumption that forma tion of inclusion cysts is the mechanism through which frequent ovulation leads to increased risk of ovarian cancer, one might expect to find more ger minal inclusion cysts in women with ovarian cancer. However, in a study done by Westhoff et al. (93), the authors counted the germinal inclusion cysts in a single slide of sections from ovaries of 148 women who underwent incidental oophorectomy and from contralateral ovaries of 37 women with unilateral ovarian cancer. The mean number of germinal inclusion cysts was 2.7 for cases and 3.6 for controls. Conditional logistic regression analy sis showed that germinal inclusion cysts were not associated with ovarian cancer (odds ratio - 0.98, 95% Cl 0.92 to 1.04). These findings do not support the hypothesis that increased formation of inclu- 4 3 8 S boham Pathogenetic of ovarian cancer Fertility and Sterility aion cysts par se is a risk factor for the development of ovarian cancer. A hypergonadotropic state baa also been enter tained as a possible component in the etiology of epithelial ovarian cancer (94). It is also well estab lished that some of the protective effects against ovarian cancer are associated with suppressed go nadotropin concentration. Among these are preg nancy, use of OCs, and lactation. Evidence that go nadotropin may be involved in ovarian neoplasia is indirect. In general, gonadotropin levels, particu larly FSH, increase gradually with age, peak in the immediate postmenopausal years, and decline in the oldest age groups (95, 96). This pattern corre lates with the age-specific rates of ovarian cancer (94), although the peak in incidence is some 15 to 20 years after the peak in gonadotropin levels. Support for the gonadotropin theory can be implied from experimental ovarian tumorogenesis. McGowan and Davis (97) have observed that after intras plenic ovarian transplant, tumors developed be cause of increased pituitary gonadotropin secre tion, resulting from hepatic inactivation of estrogen in the venous blood before it could exert feedback on the pituitary (98, 99). However, these tumors generally originated from the stromal cells of the ovary and do not duplicate the complex epithelial patterns observed in human ovarian tumors (100). There are several physiological conditions that can cause hypersecretion of gonadotropins early in life and therefore might be expected to increase the risk of ovarian cancer. Among these are premature ovarian failure and postradiation state. Pelvic irra diation sufficient to cause menopause may be linked to ovarian cancer. However, several recent studies have found a nonsignificant excess of women with ovarian cancer who had prior irradia tion (18, 65, 101). Viruses, particularly the mumps virus, may be a cause of premature ovarian failure and hypergona dotropic state. However, a case-control study of ovarian cancer found that patients were less likely to recall a history of mumps parotitis than were the controls, suggesting a protective effect of mumps (73). It was also observed that during lactation, FSH rises and values remain elevated until the return of ovarian estrogenic function. Therefore, it might be expected that breastfeeding will increase the risk of the disease. Yet, the data show reduced risk asso ciated with breast feeding (16). In addition to the above information and along with the abovementioned hypothesis, women using noncontracsptive estrogens would also be at lower risk of ovarian cancer, because exogenous estrogen administration is known to decrease gonadotropin concentration (102-104). However, the available data reveal no effect of hormone replacement ther apy on ovarian carcinoma risk (2, 24, 28, 29, 105). It is postulated that gonadotropin receptors on the surfoce oftarget cells regulate cellular processes by transducing the binding event into the forma tion of the intracellular messenger, cyclic adeno sine monophosphate (106). This process is accom plished by receptor-coupled activation of adenylate cyclase in the plasma membrane. The data of Graves et aL (107) suggest that ovarian cancer cells maintain*^ a functional nucleotide regulatory pro tein-adenylate cyclase unit reminiscent of normal cells. As such, failure of gonadotropins to stimulate adenylate cyclase may be due to a lack of gonadotro pin receptors or an error in receptor activation of the nucleotide protein. In the Graves et aL (107) study, none of the ovarian tumors classified as epi thelial, whether well differentiated or not, exhib ited gonadotropin-responsive adenylate cyclase. The epithelial tumors are morphologically analo gous to epithelia of various portions of the muller ian tract, for example, the fallopian tube, endocervix, and endometrium (108). It is not surprising, therefore, that the epithelial tumors are gonadotro pin insensitive, because the normal tissue ana logues have not been considered targets for these hormones. The absence of a gonadotropin-respon sive system in epithelial tumors (107) suggests that the most common type of ovarian cancer is not a target tissue for gonadotropin. The ability or inabil ity to bind gonadotropin does not establish defini tively that a tissue is responsive or insensitive to these hormones. Another study (109) trying to eval uate the incidence and localization of specific go nadotropin binding in the various types of ovarian tumors, with both the biochemical binding study and surface-binding autoradiography, demon strated that none of the 11 epithelial malignant ovarian tumors examined contained the binding sites for gonadotropins. As such, due to the various inconsistencies, the contribution ofpituitary gonad otropins to the etiology of cancer of the human ovary remains poorly supported at present (110-- 112). Mechanism* of Protection. The proposed biologi cal mechanism of protective effect for pregnancy, contraceptive pill, and lactation involves suppres- Vol. 62, No. 3, Sptmbr 1994 S h o h a a Patho^tnenvi nf ovarian cancer 4 3 9 " sion of ovulation (16, 83, 86) or alterations in the production of pituitary gonadotropins (94). The hypothesis has been supported by the studies of Ca sagrande et aL (66), Hildreth et aL (67), and Fran ceschi et aL (68), who showed that the index of "ovulation years," the total time of exposure to ovulation, also contributed significantly to the risk of ovarian cancer. However, this observation was not confirmed in a study conducted by Gwinn et aL (81), who compared the month by month effect of pregnancy, use of contraceptive pilL and breast feeding on risk of ovarian cancer. These authors noticed that equal months of exposure do not neces sarily result in equal periods of anovulation. If suppression of ovulation was the single under lying protective mechanism, one would expect a given period of a particular exposure to produce the same reduction in risk, regardless of previous expo sure. This was not the result in the Gwinn et aL (81) study as it was suggested that each month of preg nancy was associated with a 2.6% reduction in the estimated relative risk, each month of breastfeed ing reduced risk by 2.4%, whereas each month of OC use reduced it by 0.8%. The observation that anovulation due to pregnancy is more protective than anovulation due to the use of OC (15,81) indi cates that aspects of ovarian function and suppres sion other than ovulation may be more relevant to the risk of ovarian cancer. This notion is also sup ported by the estimation that a woman who had one pregnahcy and breast fed may have skipped approx imately 3% of the ovulations that would have oc curred if she had never been pregnant. However, a single pregnancy is associated with a reduction in incidence by approximately one half (113). Another theory involves the speculation that pregnancy may induce changes in the pituitary gland that in some way permanently altered the se cretion of gonadotropins that stimulate the ovary. It was shown that pregnancy induces anatomical changes in the pituitary, which seem to correlate with the number of previous pregnancies (114). It was also known that the protective effect increases with increasing parity (2, 115). However, the risk reduction among parous associated with each term pregnancy declined with age. which contradicts the notion that pregnancy induces permanent changes that continue to reduce ovarian cancer risk throughout life (16). One other possible interpretation would be that the endocrinologic status after pregnancy and the use of contraceptive pill protects against ovarian cancer and the lack of that protection makes infer tile women at high risk for ovarian cancer. It was observed that one of the hormonal effects of child bearing is the sustained rise in the concentration of sex hormone-binding globulin after pregnancy (113). Hormonal contraceptive regimens induce a similar effect. This may explain benefits extending beyond the period of hormonal consumption. Local Hormone Production as a Mechanism of Proliferation. There is some evidence suggesting that hormones can cause differentiation and prolif eration of the surface endometrium. Henderson et aL (116) recently stated that "neoplasia is the con sequence of excessive hormonal stimulation of a particular target organ, the normal growth of which is under hormonal controL" Estrogen increases granulosa cell proliferation, and hence the fre quency of mitotic activity may lead to malignant phenotypes (116,117). It is known that the surface epithelium of the ovary can differentiate into endosalpingeaL endometrial, and endocervical-like epi thelium (118). It is likely that this differentiation is mediated through estrogen and/or progesterone sensitivity of the ovarian surface epithelium. Cyto plasmic receptors for estrogen and progesterone have been found in most benign and malignant ovarian tumors (119-121). Support evidence for this proposed mechanis* was found in an animal experimental model. Thv earliest report, in 1962, concerned eight female dogs treated with stilbestrol (122). Within 19 months, ovarian carcinoma had developed in six dogs. It was also demonstrated that cell cultures from in situ ovarian cancer tissue secreted estrogen and/or pro gesterone and that this secretion was regulated by cyclic adenosine monophosphate in the majority of cultures and by hCG and FSH in approximately 30% of the cultures (123). In addition, cultured ovarian cancer cells from an omental metastasis continued to produce estrogen, indicating that FSH and hCG have time- and dose-related positive ef fects on cell division in ceils from metastatic ovar ian carcinoma (124). However, because epidemio logic data found a decreased risk for epithelial ovarian cancer with the use of estrogen in the pre menopausal or postmenopausal period (125), it might be possible to speculate that gonadotropins promote ovarian cancer cell growth in vivo under conditions where ovarian estrogen secretion is drastically decreased or virtually niL However, the epidemiologic observation that estrogen replace ment therapy reduces risk for the development of 4 4 0 S b o h a a Pathottenesu of ovarian cancer Fertility and Sterility epithelial ovarian cancer waa not supported by others (16). Recently, a new family of low-molecular-weight peptide growth factors has been described (126). These growth factors are secreted by normal tis sues, and they act to regulate growth in the local environment in which they are produced. Accord ing to this hypothesis, constitutive growth factor production would permit autonomous growth of cancer cells. It was shown that growth of ovarian epithelial cells, in vitro, is greatly enhanced by the presence of hydrocortisone and epidermal growth factor (EGF), suggesting that specific growth factors can stimulate epithelial proliferation. It is hypothesized that such factors may promote repair of postovula tory defects in the ovarian epithelium. This hy pothesis received some support from the observa tion that some virally transformed cells produced growth factors and did not require exogenous growth factors in culture (127-131). The observa tion that rabbit ovarian follicular fluid stimulates proliferation of cultured rabbit ovarian epithelium supports this hypothesis. Lately, in vitro studies of cancer cells have demonstrated abnormalities of both growth factor production and growth factor receptor number and function (126). In a study performed by Berchuck et al. (132), the response of four epithelial ovarian cancer cell lines, previously established from ascites of patients with ovarian cancer, were tested for proliferation after stimulation by EGF, platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and transforming growth factor-beta (TGF-0). All four epithelial ovarian cancer cell lines were found to differ in their response to the mitogenes tested. Two of the four cell lines increased proliferation in response to EGF. One cell line responded to FGF, and none of the cell lines responded to PDGF. These results are consistent with a previous report (133). The action of each of these peptides is me diated by a distinct membrane receptor, which in turn activates a postreceptor signaling mechanism and leads eventually to a mitogenic response (127). In another observation, EGF and TGF-a induced in vitro growth of a wide variety of tumor types, including human ovarian carcinoma (132, 134). Bauknecht et al. (135) detected EGF receptors on 45% of 141 ovarian tumor specimens, and this has been reported to correlate favorably with response to chemotherapy and overall survival in ovarian cancer patients (136). This correlation suggests that EGF, TGF-a, and/or their receptors may con tribute to the pathogenesis of the disease. However, the infrequency with which autocrine growth is seen in nonmalignant cells suggests that additional mechanisms exist to prevent uncontrolled cell pro liferation. I n fe r tility It is postulated that a common factor, probably of gonadal nature, is responsible for both ovarian car cinogenesis and decreased fertility. Ovarian cancer risk among nulliparous (but not parous) women was positively associated with a history of unsuc cessful attempts to conceive (137). The finding that married nulliparous women have a higher risk of ovarian cancer than do single nulliparous women (19,63) suggests that infertility may be a causative factor rather than childbearing as a protective factor. Several authors have identified infertility as an independent risk factor on top of the nulliparity effect (19,63,74,137). It was speculated by Whittemore et aL (137) that it is not the number ofyears of uninterrupted ovulation that increases the risk of developing cancer but the inability to conceive, due to endocrine disorders that exposed these women to a doubled risk of developing ovarian cancer com pared with women who were fertile. This group of investigators noted that among all women, risk in creased with increasing years of unprotected inter course (P value for trend - 0.02). Risk among women having 10 years or more of unprotected in tercourse was 1.8 relative to that among women having <2 years (P 0.01) (137). In another observation among nulligravid women, those with 10 years of unprotected inter course had a risk of over six times that of nulligra vid women who reported <3 months of unprotected intercourse (138). In a study performed by Harlow et aL (74) regarding patients with borderline tu mors, the results suggested that married nuiliparoua women who reported a history of infertility were at a substantially greater risk, over six times more, of being diagnosed with a borderline tumor than married nulliparous women with no reported infertility problem. However, a noncausal explana tion for this association might be that an ovarian tumor during its early preclinical development in terferes with fertility in some manner. The above findings are supported by a study of Lais et aL (139) in which a silent ovarian carcinoma Voi. 02, No. 3, September 1904 Shoiiem P a th o g tn tiu o f ovarian ca n ctr 441 was found in 6 out of 571 patients undergoing mi crosurgery for infertility. The prevalence of ovarian cancer in these patients was 1% (95% confidence limits of 0.4 to 2.3%), compared with only 1of 5,806 patients undergoing nonspecific abdominal sur gery. Five of 6 patients were nulligravid, and only 1 patient had a previous pregnancy. This patient spontaneously aborted one preterm infant and had one tubal ectopic pregnancy. M echanism of Pathogenesis Virus Infection. Several mechanisms of action were suggested. It was postulated that with the abil ity of the rubella virus to persist for years after the initial infection, one could suggest a theory that a change in ovarian function or morphology is pro duced that, although it does not interfere with ovu lation, might alter cell surfaces, making impregnantation difficult or impossible (64). Interference with Normal Process of Ovulation. It was also hypothesized that luteinized unruptured follicle syndrome, which represents another abnor mal concomitant of ovulation, may be involved in ovarian carcinogenesis. The luteinized unruptured follicle syndrome has been associated with endo crine abnormalities that may play a role in the neo plastic process. Luteal phase peritoneal concentra tions of progesterone and E, are lower in luteinized unruptured follicle cycles than in cycles character ized by an ovarian stigma or by subsequent concep tion (140). The unruptured follicles contain high concentrations of progesterone, estrogen, andro gens, and LH (140). They can form cysts that re main in the ovarian stroma (141), and they may be related to the inclusion cycts found at autopsy. Infertility Drugs. The question whether there is a possible causal relationship between ovulation in duction and the development of ovarian cancer is an intriguing one. Several cases were reported in which an association between treatment for infertil ity and epithelial ovarian cancer of low-grade malig nancy or epithelial ovarian carcinoma have been described. The first report came from Bamford and Steele in 1982 (142). They described a case in which a pa tient with primary infertility was treated with CC and hMG for 11 cycles, in which 2 years were al lowed to pass between the first 8 cycles and the last 3. During the last treatment cycle, a mass was palpa ble that was found to be ovarian endometroid tumor with encroachment of the serosal coat and a well- differentiated endometrial adenocarcinoma. In the same year Atlas and Menczer (143) reported a ct of borderline serous papillary cystadenocarcinoma after treatment of one cycle with CC. Ben-Hur et aL in 1986 (144) reported two cases in which papillary serous cystadenocarcinoma were diagnosed after treatment with CC and bromocriptine in one pa tient and CC only in the second one. Carter and Joyce (145) in 1987 reported a case of papillary serous adenocarcinoma after treatment with CC, hMG, and hCG for an IVF program. In 1989 Kulkami and McGarry (146) reported a case of serous papillary cystadenocarcinoma in a patient 8 years after several treatment cycles with CC and cydofenil, which at that time ended with pregnancy and normal delivery. Dietl in 1991 (147) reported a case of papillary serous adenocarcinoma in patients with endometriosis who received treatment for sev eral months with CC, hMG, hCG, and two types of GnRH agonist (GnRH-a). Goldberg and Runowicz (148) in 1992 reported three cases of borderline ser ous adenocarcinoma of the ovary associated with infertility and ovulation induction. In two patients treatment for two and three cycles with hMG for an IVF program was administered. In the third patient two treatment cycles with hMG were administered for ovulation induction. The last report in the literature came from N man et ai. (149) about two patients who developed a borderline malignancy of the ovary after treatment with hMG, GnRH-a, and hCG in the context of an IVF program. Even though such individual reports do not prove a direct connection between stimula tion and the induction of ovarian neoplasms, it is nevertheless possible that the stimulation or other associated factors substantially stimulated an exis tent tumor, as in most cases the tumor was found shortly after ovulation induction treatment and it developed with remarkable rapidity. Most recently, a series of articles published by Whittemore et al. (14-16) has focused attention on the association between infertility drugs and inva sive epithelial ovarian cancer. The authors de scribed findings for invasive ovarian epithelial tu mors in white women based on a collaborative analysis of data from 12 case-control studies of ovarian cancer in the United States. The original studies were based on data collected between 1956 and 1986 and included 2.197 women with invasive epithelial ovarian cancer, 327 women with ovarian tumors of low malignant potential, and 8.893 con trols. The study involved personal interviews with 4 4 2 Shoham 1`nthnttetuma nf ovarian cancer Fertility and StenlL_. study subjects. In general, most of the study find* ings were consistent with the current consensus; the risk of ovarian epithelial cancer decreased with increasing parity, the use of OCs, and breastfeed ing. The risk increased with increasing duration of unprotected intercourse regardless of parity and gravidity. However, the most striking newest data regard the use of fertility drugs. Of the 12 case-control studies, 3 studies appeared that contain data on past use of fertility drugs (17-19). The global popu lation of these three studies was 2^278 women, of whom 711 women with invasive ovarian cancer and 1,211 controls were included in the analysis. Thirty-one women were found to have some history of previous use of fertility drugs. The specific drugs involved are not known nor are the dosage, dura tion of therapy, or type of infertility treatment. Regarding the use of fertility drugs, it was found that the risk for invasive epithelial ovarian cancer among women who were diagnosed as being infer tile and who had received fertility drugs was almost three times that of women who had no history of infertility (odds ratio (OR] " 2.8, 96% Cl 1.3 to 6.1). Infertile women who had not used fertility drugs had no increased risk (OR - 0.91, 95% Cl 0.66 to 1.3). Those infertile women who used fertility drugs and became pregnant did not have significant in creased risk for ovarian cancer (OR 1.4, 95% Cl 0.52 to 3.6). However, among infertile women who had never become pregnant, the use of fertility drugs was associated with significantly increased risk of cancer (OR 27.0, 95% Cl 2.3 to 315.6). Fertility drugs had been used by 12 of 34 nulligravid cases compared with one of 23 nulligravid controls. The authors proposed the hypothesis that the in creased risk associated with infertility may be due to the use of fertility drugs. The data and the hypothesis presented by Whittemore et aL (15) and by Harris et aL (150) have prompted a worldwide concern for women's health and therefore raised an extensive discussion among epidemiologists, endocrinologists, oncologists, and gynecologists. It is not the purpose of this review to argue with the methods or the statistical analysis nor the results of the study of Whittemore et al. (15). However, some points regarding the use of fer tility drugs, as raised in the Whittemore et al. (15) paper and those published in the literature, are dis cussed. The most extensive research concerning the asso ciation of treatment of infertility and carcinoma of the ovary was published by Ron et aL (151). This study, conducted in Israel, gathered information about 2,632 Israeli women who were treated for in fertility between 1964and 1974. Analysis by infertil ity diagnosis demonstrated no significant excess of total cancer incidence. The standardized incidence ratio was 1.3 (96% Cl 0.8 to 1.8) for infertility due to hormonal deficiency, 0.7 (96% Cl " 0.3 to 1.4) for mechanical infertility, 1.6 (95% Cl " 0.6-3.6) for infertility of the male partner, and 1.1 (95% Cl " 0J5--2.2) for unclassified diagnosis. Drugs that were evaluated in the study were hMG, CC, and hCG. However, because women received many different hormones during their course of treatment, it was difficult to distinguish independent treatment ef fects. Among women with the diagnosis of hor monal infertility, neither hMG nor CC was signifi cantly associated with a cancer risk greater than the risk of infertile women receiving other hor mones or no hormones. The authors summarized their paper by stating that no evidence of associa tion between ovulation induction drugs and cancer of the ovary should be found. Under the auspices of the International Federa tion of Fertility Studies, a group of clinical and epi demiologic experts (152) has made several com ments regarding the study of Whittemore et al. (15). One of the comments is that the new fertility drugs were registered in the United States quite late in relation to the study period. Clomiphene citrate was registered in 1967, hMG was registered in 1969, and bromocriptine in 1978; before that period, treatment for infertility included pituitary irradia tion, pregnant mare serum gonadotropins, conju gated estrogen, and diethyl stilbestrol, none of which is used any more (152). In the Whittemore et aL (15) study, where infer tility treatments were analyzed, the average age of women with cancer was 53 years, where the cancer was diagnosed between 1977 and 1981. If it is as sumed that infertility treatment took place between the ages of 30 and 40, those women were then treated between 1954 and 1967 on average. These results are not compatible with use of the new fertil ity drugs that were registered from 1967 and on ward. Caro et aL (153) raised the point that at the time the original studies were carried out, infertility drugs were used almost routinely in women diag nosed with ovulatory infertility, which represents 20% to 40% of female infertility. For example, among 708 infertile women (154,155), 47% had an ovulatory infertility and 90% of those received VoL 82, No. 3, SopWmbar 1994 Shobtua Pathogeneeu of ovarian cancer 443 drugs. Caro et aL (153) also raised the possibility that women who received treatment for infertility in the past who did not conceive were more liable to remember that they had been treated than those who did conceive (156). It might be possible that infertility is due to a condition that in itself could be associated with ovarian cancer, and infertility is an expression of a common underlying phenomenon. It cannot be to tally ruled out that ovulation induction might acti vate certain types of pre-existing ovarian cancer. In summary, a number of potentially important risk factors have been identified. These are a family history of gynecologic cancer and cystic ovaries, nulliparity or low gravidity, difficulty in conceiving and infertility, and lack of exposure to OCs. If the above studies of ovarian cancer are to be translated into disease prevention, in view of the relatively high risk for infertile women to develop ovarian car cinoma, there is a need for careful clinical evalua tion with ultrasound or other modern techniques before and during ovulation induction treatment. There is an urgent need for a prospective multi center trial in which the scientific community would be able to confirm or refute the data regard ing fertility drugs. Until then, use of fertility drugs should be handled with care, following pretreat ment and post-treatment monitoring of the pa tients. REFERENCES 1. CulUr SJ, Younc JL , diton. T hird N ational Caacar Sur vey: incidence data. W ashington. O .C j U.S. Governmental Printing Office; 1978. N atl Cancer Inst Mooogr, DHEW publication. 1973;41:74-8. 2. H einu AP. Hacker NF. f m e a t LX). Epidemiology and etiology of ovarian cancan a review. Obetet Gynecol 1986;68:127-33. 3. Weiaa NS, Homoncbuk T , Young JL . Incidence of the his tologic types of ovarian cancan th e U S . third national cancer turvey. 1969-1971. Gynecoi Oncol 1977;8:161-7. 4. Abell M R The nature and classification of ovarian neo plasms. Can Mad Aiaoc J 1966:94:1102-24. 5. H artig AT, Gore HM. Ovarian cyatomaa of germinal epithe lial origin. Rocky MT Mad J 1958:55:47-50. 8. Baaaia ML. An em bryolopcaily derived classification of ovarian tumors. JAMA 1960:174:170-3. 7. Peres RP. Godwin AK. Ham ilton TC, Otole RF. Ovarian cancer biology. Semin Oncol 1991;18:186-204. 8. Cramer DW, Welch W R D eterm inants of overian cancer risk. It. Inferences regarding pathogenesis. J N atl Cancer inat 1983;71:717-21. 9. Barber RK. Cancer of the ovary. In: Nagetl JR Jr, Barber HRK. editors. Modern concepts of gynecology oncology. Boston: John W right PSG Inc.. 1982:23-40. 10. Copplaaon M. Gynecologic oncology: fundam ental principlea and clinical practica. Naw York: Churchill L iv in ' tons, 1981.-668-71. 1L T hornton JG , W alls M. Ovarian cysts in pregnancy: does ultrasound make traditional management inappropriate? Obetet Gynecol 1967;69:717-21. 12. Chong A, Birnbaum SJ. Ovarian cancer com plicating pregnancy. Obetet Gynecol 1973;41:211-4. 13. D g u i R Shofaam Z, A tar R Zosmer A, Lancet M. Ovarian carcinoma during pregnancy: a study of 23 cesse in Israel between the y ean 1980 and 1984. Gynecol Oncol 1960-.3&326-3L 14. W hittem ore AS, H arris R Itnyre J, Halparn and the Col laborative Ovarian Cancer Group. C haracteristics relating to ovarian cancer risk: collaborative analysis of 12 US ra n control studies. L Methods. Am J Epidemiol 1992:138:1175-83. 15. W hittem ore AS, H arris R Itnyre J , and the Collaborative Ovarian Cancer Group. C haracteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies, 0 . Invasive pitbeiial ovarian cancers in white women. Am J Epidemiol 1992;138:1184-203. 18. W hittam ots AS, H arris R Itnyre J. and tbs Collaborative Ovarian Cancer Group. C haracteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control sruriia IV. The pathogenesis of epithelial ovarian cancer. Am J Epidemiol 1992:136:1212-20. 17. H artge P, Schiffman MH, Hoover R McGowan L, Lasher L, N om s H J. A case control study of epithelial ovarian cancer. Am J Obetet Gynecol 1989:161:10-6. 18. Cram er OW, Hutchison GB, Welch W R Scully RE. Ryan KJ. D etarm inants of ovarian cancer risk. I. Reproductive experiences and family history. J N atl Cancer I 1983;71:711-6. 19. Nasca PC, Greenwaid P, Chorost S. Richert R, Caputo T. An epidemiologic case-control study of ovarian cancer and reproductive (actors. Am J Epidemiol 1984:119:705-13. 20. Lynch HT, Schueike GS, Wells IC, Cheng SC, Kimberiing W J, Biacone KA. et aL H ereditary ovarian carcinoma: biom arkar study. Cancer 1985;55:410-5. 21. Piver MS, M ettlin CJ. Tsukada Y. Nasca P. Greenwaid P. M cPbee M R Familial ovarian cancer registry. Obetet Gynecol 1984,-84:195-9. 22. Lynch HT, Bewtra C, Lynch JF . Fam ilial ovarian carci noma. Clinical nuanas. Am Med J 1986:81:1073-6. 23. Lewis AC, Davison BCC. Familial ovarian cancer. Lancet 1960&235-7. 24. Lynch HT. Guirgia HA, Albert S. Familial association of carcinoma of the breast and ovary. Surg Gynecol Obstet 1974:138:717-24. 25. Lynch HT. Albano WA, Lynch JF. Lynch PM. Campbell A. Surveillance and management of patients at high ge netic risk for ovarian carcinoma. Obetet Gynecoi 1982^589-96. 28. Lynch HT. Albano WA, Black L, Lynch JF . Recabaren J, Pierson R Familial excess of cancer of the ovary and other anatom ic sites. JAMA 1981;245:261-4. 27. Lynch HT. H arris R R Guirgia HA. Maloney K, Carmody LL, Lynch JF. Fam ilial association of breaat/ovarian canrcinoma. Cancer 1978:41:1543-8. 28. Lynch HT. Kimberiing W, Albano W. Lynch JF. Biscone K. Schueike GS. st al. H ereditary nonpolypoaia colorectal 4 4 4 S hohem Pathogenem* of ovarian cancer ! Fertility and Sterility cancan p art L Clinical description of resource. Cancar 198646234-3. 29. Lunch H T, Scbueike GS, Kimberiing W, Albano WA. Lynch JF . Biacooa KA, a t aL H ereditary nonpolypoaia coloractal cancan part IL Biom arkar studies. Cancar 198646239-5 L 30. Praum ani JF , Grandy GW, Creagan E T. Evaraon RB. Six fam.iiaa prana to ovarian cancar. Cancar 1976:36464-9. 31. L u n in JR . Ptver MS. Fam ilial ovarian carcinom a. Gynecoi Oncol 1979*186-92. 32. Ptvar MS, Baker TR. Piadmonta M. Sandacki AM. Epide miology and atiolosy of ovarian cancar. Samin Oncol 1991;18:177-38. 33. T ran t JM . Kanako Y, M itaiman F. Report of tha committaa of structural chramoaoma changes in neoplasia. Cytogaoat CaU Ganat 1988:49236-63. 34. Ruaaail SE, Hickey GL Lowry WS, W hite P, Atkinaon RJ. Allele loaa from chromoeome 17 in ovarian cancar. Onco gene 1990*1381-3. 33. Ecclaa DM, Cranston G, Staaia CM. N akam ura Y, Leon ard RC. Alleia loaaaa on chromoaoma 17 in human epithe lial ovarian carcinoma. Oncogene 1990*1399-801. 36. Omhimun M. Kugoh H. Koi M, Shimizu M. Yameda H. Satoh H. et ai. Transfer of a norm al hum an chromoaoma 11 lupprem as tum origsnirity of soma but not all tum or call lines. J CaU Biochem 1990;42:138-42. 37. Finlay C, Hinds P, Levina A. Tha pA3 proto-oocogana can act as a (uppreasor of transform ation. CeU 198947:1083 93. 38. Huang HS. Yes JK . Shaw JY . Chan PL. Bookstatn R, Friedm an T , at el. Suppression of the neoplastic pheno type by replacement of th e RB (ana in human cancer calls. Science 1988242:1563-6. 39. Crickard K, M arinallo M J, Crickard U. Satchidanand SK. Yoonaaai M, Cagiar H. Borderline m alignant serous tu mors of th a ovary m aintained on extracellular matrix: evi dence for d onai evaluation and invasive potential. Cancar G anat Cytogenet 198623:135-43. 40. KnoefT-Gaertner H, Scbtihm enn R. K raus H. Compara tive cytogenetic and histoiogir etudiss on early m alignant transform ation in meanthalial tum ors of th e ovary. Hum Genet 197748281-97. 41. Barlow J J , DiCioccio RA. Dillard PH . Blumenaoo LE, M atte KL. Frequency of an aileie for low activity of a-L-fum sirtses in ten s possible increase in epithelial ovarian cancar. J N atl Cancar inat 1981,-67:1006-9. 42. K olstad P, Beacham JC . Epidemiology of ovarian neopla sia. P m raedinp of the American-European Conference on the ovary. Montreux, Switzerland: Exerpta Medics Int Congress Series, 1974;364:56-63. 43. Lingeman CM. Etiology of cancar of tha human ovary: a review. J N atl Cancer Inat 1974*1:1603-9. 44. Cram er DW, Watch WR, Scully RE, Wojciechowski CA. Ovarian cancar and talc: a case control study. Cancar 198240472-6. 43. Henderson WJ, Hamilton TC. Griffiths K. Talc in normal and m alignant ovarian tissue [latter!. Lancet 1979:2:499. 46. Longo DL. Young RC. Coamatic talc and ovarian cancar. Lancat 1979*349-31. 47. G raham J, Graham R. Ovarian cancer and asbestos. En viron Rea 1967;1:113-28. 48. Roa FJ. Controversy: coamatic talc and ovarian cancar (letter}. Lancet 1979*744. 49. H artga P, Hoover R, Lasher LP, McGowan L. Talc and ovarian cancar (latter). JAMA 1963:230:1844. 50. Newbouse ML. Pearson RM, Fullerton JM . Boaaan EA. Shannon HS. A caaa control study of carcinoma of tha ovary. B r J Prav Soc Mod 1977-^1:148-33. 31. Parm aly TH . W oodruffJD . Tha ovarian mostbaUoma. Am J Ofaatst Gynecol 1974:120*14-41. 32. Egii GE. Newton M. T ha transport of carbon particles in th e human fruale reproductive tract. Frtil Steril 1961;12:151-6. 53. V enter PF. O varian epithelial cancer and chemical carcino genesis. Gynecol Oncol 1981:12281-5. 34. Roa AR. Effects of carcinogen and/or mutagen on normal and gonadotrophin-prim ed ovaries of mice, lo t J Cancer 1961;28:106-10. 36. Bangtaaon M, Rydstrom J. Regulation of carcinogen me tabolism in tha rat ovary by tha estrous cycle and gonado tropin. Science 1983:219:1437-8. 56. Bangtaaon M. Hambargar L. Rydstrom J. Metabolism of 17,12-dimathyibetiz (a) anthracene by different types of calla in th a human ovary. Xenobiotica 1988:18:1233-70. 37. Cram er DW, W elch WR. Hutchison GB. W illett W, Scully RE. D ietary anim al fot in relation to ovarian cancer risk. O bstat Gynecol 1984:63:833-8. 38. Rosa DP, Boyar AP, W ynder EL. International compari sons of m ortality rates for cancer of tha breast, ovary, proatata, and colon, and par capita food consumption. Cancar 1986482383-71. 39. Cram er DW. Lactase persistence and milk consumption as determ inants of ovarian cancer risk. Am J Epidemiol 1989:130204-10. 60. Cram er DW, Harlow BL. WiUatt WC. Beil DA. Scully RE. Ng WG. at aL Galactose consumption and metabolism in relation to th e risk of ovarian cancer. Lancet 1989:2:66-71. 61. M ettlin CJ, Prver MS. A case-control study of m ilk-drink ing and ovarian cancar risk. Am J Epidemiol 1990:132:371 - 8. 62. W ynder EL, Barber HR. Epidemiology of cancer of the ovary. Cancar 196923232-70. 63. Joly D J, Lilianfold AM. Diamond EL. Broas ID. An epide miologic study of tha relationship of reproductive experi ence to cancar of th e ovary. Am J Epidemiol 1974:99:190 209. 64. McGowan L, P arent L. Ladnar W. Norris HJ. The woman at risk for developing ovarian cancer. Gynecol Oncol 1979;7:323-44. 63. Annegera JF . Strom H. Decker DG. Dockerty MB. O'Fal lon WM. Ovarian cancer. Incidence and case-control study.Cancer 1979:43:723-9. 66. Casagranda JT . Pika MC. Rosa RK. Louie EW. Roy S. Henderson BE. " Incessant ovulation'' and ovarian cancer. Lancet 1979*170-3. 67. H ildreth NG. Kelsey JL . LiVolai VA. Fischer DB. Hoiford TR , Moscow ED, at ai. An epidemiologic study of epithelial carcinoma of th e ovary Am J Epidemiol 1981; 114:398-405. 68. Franceechi S, La Vecchia C, Helmrich SP. Mangioni C. Tognoni G. Risk factors for apithaiiai ovarian cancer in Italy. Am J Epidemiol 1982:115:714-9. 69. Negri E, Franceechi S. Tzonou A. Booth M. LA Vecchia C. Pmrasxini F. et ai. Pooled analysis of 3 European ease-con- Vol. 62. No. 3. September 1994 S bo ham Pathogenesis of ovarian cancer 4 4 6 trol nidiaa L Reproductive faetn and risk of epithelial ovarian canear. lo t J Cancar 1991;49:50-6. 70. La Veccfaia C, Deearii A, Fraocaachi 3. RaaUo M. Tognoni G. Ag* a t first and tha risk of epithelial ovarian cancar. J N ad Cancar In 1964:73.-663-. 71. Kvafo 0 . Hand L Nilaaan S, Barml V. Reproductiva factors and risk of ovarian cancan a prospective study. lo t J C ancar 1986;42346-51. 72. W yndar EL, Dodo H. Barber HR. Epidemiology of cancar o f tb s ovary. Cancar 196933352-70. 73. W est RO. Epidemiologic study of m alignancies of th e ova rica. Cancar 1966:19:1001-9. 74. Hariow BL, Wriaa NS, Roth GJ, Chu J . D alinf JR . Caaecon trol study of borderline ovarian tum ors: reproductiva history and esposare to esofenoue femis hormones C ancer Res 1968:483849-52. 75. The Cancer and Steroid Hormone Study of the C antan for Disease Control and th a N ational In stitu te of Child H ealth and Human Development. The reproduction in risk of ovarian cancer seaoriarad with oral-contraceptive use. N E n d J Med 1967316350-5. 78. W ane NS. Lyon JL , Liff JM . Vollmer WM, Dating JR . Incidence of ovarian cancer in relation to the usa of oral con tra ceptives. lo t J Cancar 19613&669-71. 77. Rosenberg L. Shapiro S. Slone D. Kaufman DW. Habnrich SP. M iattim an OS, at aL Epithelial ovarian cancer and com bination oral contraceptives. JAMA 19623473210-2. 78. Wu M L W hittam ore AS, Paffenbarger RS Jr, S arita DU K am part JB , Grosser S, et aL Personal and environm ental characteristics relatad to epithelial ovarian cancar. L Re productiva and m enstrual events and oral contraceptive uaa. Am J Epidemiol 1968:128:1216-27. 79. Cram er DW. H utchinson GB. Welch WR. Scully RE, K napp RC. Factors affecting the association of oral con traceptivas and ovarian cancar. N E n d J Mad 1962307:1047-51. 80. C antan for Disease Control Cancer and Steroid Hormone Study. Oral contraceptive uaa and the risk of ovarian cancar. JAMA 1983349:1396-9. 81. Gwian MU Lea NC. Rhode* P a Layda PM. Rubia GU Pregnancy, breast feeding, and oral contraceptives and tha risk of epithelial ovarian cancar. J Clin Epidemiol 1990:43359-68. 82. Sharp F, Sootier WP, editors. Ovarian cancer--the way ahead. Chichester: Wiley, 1967. 83. Beral V. The epidemiology of ovarian cancar. In: Sharp F, S o u tttr WP, editora. Ovarian cancar--tha way ahead Chi ch ester Wiley, 198731-8. 84. Riacfa HA, Weias NS, Lyon JU Deling JR , LiffJM . Events of reproductiva life and tha incidence of epithelial ovarian cancer. Am J Epidemiol 1983:117:128-39. 85. W illett WC, Bain C. Hannakena CH, Rosnar B. Spaixsr FE. Oral contraceptives and risk of ovarian cancar. Cancar 1981;46:1684-7. 86. Fathalla MF. incessant ovulation--a factor in ovarian naoptaaia (letter)? Lancet 19713:163. 87. Thompson JP . Dockarty MB, Symmonda RE, Hay!** AB. Ovarian and parovarian tumors in infants and children. Am J Ofastet Gynecol 196737:1039-66. 88. Fathalla MF. In: Rashad MN, Morton WRM. editors. Se lected topics on genital anomalies and related subjects. Illi nois: 1969:684. 89. Zajicek J. Ovarian cystomas and ovulation: a histoganstic concept. Turnon 197733:429-35. 90. Redisevijevic SV. The pathogenesis of ovarian in d u s' cysts and cystomas. Ofastet Gynecol 1977;49:424-9. 91. Potaahnik G, Inalar V, M ein er I, Sternberg M. Frequency, sequence, and tick of ovulation in woman m enstruating normally. B r Mad J 1967394319. 92. Cim ckahaak DJ. Aetiological im portance of ovulation in epithelial ovarian cancan a population baaed study. Br Mad J 1990301324-5. 93. Weatfaoff C, M urphy P, H allar D, Halim A. Ia ovarian canear marinated with an increaaad frequency of germinal inchwion cysts? Am J Epidemiol 1993:13830-3. 94. Stadel BV. The etiology and prevention of ovanan cancer. Am J O bstet Gynecol 1975:123:772-3. 96. Rayas Ft, W inter JSD , Fsim an C. Pituitary-ovarian rela tionship* preceding the menopause. L A crosa-eectional study of serum follicle-stim ulating hormona, luteiniting horm one, prolactin, aatradioL and progastaron* lavais. Am J Ofastet Gynecol 1977,129337-64. 98. C baknvarti S, Collins WP, Forecast JD , Newton JR , Oram DH. Studd JW . Hormonal profiles after tha meno pause. Br Mad J 19763:784-6. 97. M eGowanU Davis RH. Intrasplenk ovarian grafts in Syr ian ham sters and peritoneal fluid cellular distribution. Pine Soc Biol Mad 1970:134307-9. 96. Biskind MS, Bisklnd GR. Development of tum ors In the rat ovary after transplantation into tha spleen. Proc Soc Exp Biol Mad 194435:176-9. 99. Biskind GR, Biskind MS. Atrophy of ovaries transplanted to th e spieen in unilaterally castrated rate: proliferative changas following subsequen t removal of intact ovary. Soane* 1948:106:137-8. 100. Murphy ED. Major experimental models of ovarian mors: histogenesis and evaluation, in: Murphy E ^ . P iim ir WG, editors. Biology of ovarian neoplasia. Ge neva: UICC. 198036-73. 101. Sm ith PC, Doll R. Lata effects of X -irradiation in patients treated for m etropathia haemorrhagic*. Br J Radiol 1976:49324-32. 102. Lauritaan C On endocrina effaa of oral contraceptives. Acta Endocrinol (Copan) l967;124(Suppl):87-l00. 103. Cook* ID, nderten K J, Lenton E. Hormone patterns at tha climacteric. Postgrad Mad J 197632(Suppl):12-6. 104. Laraaon-Cohn U, Johansson ED. Kagcdai B, W ailenun L. Serum FSH. LH and oaatron* lavais in postm enopausal patients on oestrogen therapy. Br J O baut Gynaecol 197735367-72. 106. Chen KT, Scbooiey JL , Flam MS. Peritoneal carcinom a tosis after prophylactic oophorectomy in familial ovanan cancar syndrome. O baut Gynecol 1985;6633S--tS. 106. Stouffer RL, Grodia MS. Davis JR , Surw it EA. Investiga tion of binding lita* for follicle-stim ulating hormone and chorionic gonadotropin ia human ovarian cancer*. J Clin Endocrinol Metab 198439:441-6. 107. Grave* PE, Surwit EA, Davis JR . Stouffer RL. Adenylate cyclaaa in human ovarian cancers: sensitivity to gonadotro pins and nonhormonal activators. Am J Obetet Gynecol 1985:133377-82. 106. McGowan L. Ovarian cancer. In: McGowan L, editor. Gy necologic Oncology. New York: Appieton-Century-Crofts, 1978:283-91. 4 4 6 9 hoham P cu h o g tn tiu of ovanan cancer Fertility and Steru.uy 109. N akano R. KiUyama S, Yamoto M. Shim s K, Ooehima A. Localization of gonadotopin binding sitae ia human ovar ian neoplasm s Am J Obetet Gynacol 1969;161:906-10. 110. M ohle J , W hittem ore A, Pika M. Darby 3. Gonadotro phins and ovarian cancar risk. J N atl Cancar Inst 1968;75:178-9. 111. B sm g h ir K, Eehkol A. Brinadan P, Edwards RG. Follicu lar stim ulation; is it safa [lattari? Br Mad J 1989399316. 112. Fox R, W ardla P. Follicular stim ulation: is it sals [lattari? B r Mad J 1969399316. 113. Scott JS . How to indues ovarian cancar: and bow not to. Br Mad J 1964389:781-2. 114. Raemumen AT. The weight of the principal com ponents of tb s normal hypophysis catabri of tfaa adult human famaia. Am J A nat 193436353-75. 115. Lowry 3, Rusaall H. Hickay L A tkinson R. Incessant ovu lation and ovarian cancar. Lancat 1991337:1544-5. 116. H andsn o n BF, Ross R. Bam stain L. Eatrofons as a causa of hum an cancan tfaa Richard and H inds Roaantfaai Foun dation Award Lecture. Cancar Raa 1988:46346-53. 117. A tkin NB, Baker MC. Abnormal chromosomes including sm all m atacentrics in 14 ovarian cancers. Cancar Genet C ytofenet 196738356-61. 118. Lauchlan SC. T he secondary m ullsrian system. Obatat Gynacol Sure 197237:133-48. 119. Fribarg LG, Kuliandar S. Persijn JP , K orstan CB. On re ceptors for strogenes (E,) and sndrogenes (DHT) in hu man andomacrial and ovarian tum ors. Acta Obatat Scand 197837381-4. 120. H olt JA , Caputo TA. Kelly KM, Greenwsld P, Choroat S. Estrogen and progestin binding in cytosol of ovarian carci noma. Obatat Gynacol 19793330-8. 121. Galli MC. Da Giovanni CS. Nicoietti G, Grilli S, N anni P, Prodi G, et ai. The occurrence of m ultiple steroid hormone receptors in diaeaae-free and neoplastic human ovary. Cancer 1981;47:1297-302. 122. Jab ra AB. Induction of Canine ovarian tum ors by diethylstiibeetrol and progesterone. Aust J Exp Biol Med Sci 1982;40:139-43. 123. W unalaeena J, Meehan D. Csvaiio C. Human epithelial ovarian cancer cell steroid secretion and its control by go nadotropins. Gynecol Oncol 1991:4138-63. 124. W imalaeena J, Doatal R, Meehan D. Gonadotropins, es tradiol. and growth factors regulate epithelial ovarian cancar call growth. Gynecol Oncol 1992;46:345-50. 125. H artge P, Hoover R, McGowan L, LeSher L, Norris H J. M enopause and ovarian cancer. Am J Epidemiol 1988:127390-6. 126. G oustin AS, Leof EB. Shipley GD. Moses HL. Growth fac tors and cancar. Cancer Res 1986;46:1015-29. 127. Spom MB, Roberts AB. Peptide growth factors are m ulti functional. N ature 1988332317-9. 128. Kaplan PL, Anderson M, Oxanne B. Transform ing growth factorial production enables ceils to grow in the absence of serum: an autocrine system. Proc N atl Acad Sci USA 198179:485-9. 129. Ozanne B, Fulton RJ, Kaplan PL. K irsten murine sarcoma virus transform ed cell tines and s spontaneously trans formed rat cell-line produce transform ing factors. J Cell Physiol 1980:105:163-60. 130. DeLarco J , Todaro G. Growth factors from m urine sar coma virus-transform ed cells. Proc N atl Acad Sci USA 1975;75:4001-5. 131. Weiaaman B, Aaronaon S. BALE and K irsten m urine sar coma virussa alter growth and differentiation of EGF-dep en d tn t BALB/c mouse epiderm al karatinocyta lines. Cell 196332399-808. 132. Berchuck A. Olt G J, E veritt L, Soiaaon AP, Bast RC Jr, Boyar CM. T he role of peptide growth factors in epithelial ovarian cancar. Obatat Gynecol 1990:75356-62. 133. Bowan-Pope DF, Row R. Platelet-derived growth factor. J Clin Endocrinol M etab 1984;13:191-205. 134. Singletary SE, Baker FL, Spitzer G, Tucker SL, Tomaeovic B, Brock WA. et aL Biological effects of epidarmai growth factor on the in vivo growth of human tumors. Cancer Res 1967;47:403-6. 135. Bauknecht T , Kohler M. Jans I, Pbleiderer A. The occur rence of epiderm al growth factor receptors and the charac terisation of EGF-like factors in human ovarian, endome trial, cervical, and breast cancar. EGF receptors and factors in gynecological carcinomas. J Cancer Res Clin On col 1989;115:193-9. 136. Bauknecht T . Runge M. Schwall M. Pfleiserer A. Occur rence of epidermal growth factor receptors in human adenexai tum ors and their prognostic value in advanced ovar ian carcinomas Gynecol Oncol 198839:147-57. 137. W hittsm ors AS, Wu ML. Paffenbarger RS. Series DL, Kampert JB , Grosser S, et si. Epithelial ovarian cancer and the ability to conceive. Cancer Res 1989;49:4047-52. 138. Booth M, Beral V, Smith P. Risk factors for ovarian cancer a case control study. Br J Cancer 1989:60392-8. 139. Lais CW, W illiams T J, Gaffey TA. Prevalence of ovarian cancer found at the time of infertility microsurgery. Fenil Steril 1988;49:551-3. 140. Devroey P, Neektgeboren N. Temmerman M. Verhoeven N, Help J. Van Steirteghem AC. Origin of low peritoneal fluid steroid concentrations in the luteinized unruptured follicle syndrome. In: Thompson W. Harrison RF, Bonnar J, editors. World Congress of Fertility and Sterility, 1983. Dublin: Boston MTP Pram. 1983:79-83. 141. Knincks PR, De Moor P. Bros*ns IA. Diagnosis of the luteinized unruptured follicle syndrome by steroid hor mone assays on peritoneal fluid. Br J Obstet Gynaecol 198037329-34. 142. Bamford PM. Stasis SJ. U terine and ovarian carcinoma in a patient receiving gonadotrophin therapy: case report. Br J Obatat Gynaecol 198239362-4. 143. Atlas M. Menczer J. Massive hyperstim ulation and bor derline carcinoma of the ovary. Acta Obstet Gynecol Scand 1982;61:261-3. 144. Ben-Hur H. Dgaru R, Lancet M, K atz Z. Nisaira F. Rosenman D. Ovarian carcinoma masquerading as ovarian hy perstim ulation syndrome. Acta Obetst Gynecol Scand 198635313-4. 145. C artar ME, Joyce DN. Ovarian carcinoma in a patient hyperstim ulated by gonadotropin therapy for in vitro fertiliza tion: a case report. J ln Vitro F ett Embryo Transfer 1987;4:126-6. 146. K ulkam i R, McCarry JM . Follicular stim ulation and ovar ian cancar (le tte r|. Br Med J 1989399:740. T Vol. 82. No. 3. September 1994 't Sho he m P athogtne^ of ovarian cancer 447 147. Diati J. Ovulation and ovarian cancar [lattari. Lancet 1991:2:446. 148. Goldberg GL. Runowica CD. Ovarian carcinom a of low malicnant potanual. infertility, ad induction of ovulation-- is that a link? Am J Ofastat Gynacoi 1992:166363-4. 149. Nljman HW, Bursar CW, Baak JPA . Schata R, Vetmorkan JB , Kanamana P. Bordatiina m alifnancy of tb a ovary and controlled hyperatim uiadon, a raport of 2 casse. E ur J Cancar 199238A:1971-3. 160. H arm R, W hittamora AS, Itnyra J , Collaborativa Ovarian Cancar Group. Characteristic (latine to ovarian canca r risk: collaborativa analyaia of 12 US csss control studia. HL Bpitbalial turnara of loar m alignant potential in white woman. Am J Epidemiol 1992:136:1204-11. 151. Ron E, Lunenfitld B, M anejar J , Blum stain T , K ata L, Oalanar G. at s i Cancar incidence in a cohort of infertile woman. Am J Epidemiol 1967;123:780-90. 152. Coban J, Forman R, H ariap S, Johanniaaon E, Lunenfitld B, da Mouson J, at a l IFF9 azparts group raport on the W hittam ora study relatad to tba risk of ovarian cancer asaociatad with tba use of infartility agents. Hum R sp r^ 1993:8396-9. 153. Caro J J , Johannas CB, H arts SC, M in a R, M iattinan Ot>. RE: "Characteristics relating to ovarian cancer risk: collab orative analysis of 12 US case control studies. IL Invasive epithelial ovarian cancer in whit woman" [latter). Am J Epidemiol 1993:137328. 154. H ull MG, Savage PE, Bromham DR. Anovulatory and ovulatory infertility: results w ith simplified management. Br Mad J 1962384:1681-6. 166. H ull MG, G lasanar CM, Kailay N J. Conway DI. Foster PA. H inton RA. at aL Population study of causes, treat m ent, and outcome of infertility. Br Mad J 1985:291:1693 7. 156. M ahlstedt PP. The peychoiogicai com ponent of infartility. FartU Steril 1966:43335-46. 4 4 8 S hoham Pathuttenenu nf ovarian cancer Fertility and Stern