Document rxZLgdbYBQLJxrwJGa68ero3v

Environmental Health Penpecliuet Vol. S3, pp. 49-66.1983________ ST0096990 Critical Review of Epidemiologic Studi Related to Ingested Asbestos ty Gary M. Marsh* Thirteen epidemiologic ttudie* of bigected ubestoe conducted in five areae at the United States and Canada were reviewed and evaluated for the deflnitiveness and applicability regarding the development of ambient water quality standards. One or more studies found male or female associations between asbestos in water supplies and cancer mortality (or incidence) due to neoplasms of the esophagus, stomach, small intestine, colon, rectum, gaitoladder, pancreas, peritoneum, lungs, pleura, prostate, kidneys, brain, and thyroid, and also due to leukemia. Several methodologic weaknesses and limitations were found in each study, leading to the determination that no individual study or aggregation of studies exist that would establish risk levels from ingested asbestos. A binomial probability analysis of the eight independent studies suggested that, while the level of male-female agreement was generally low, the number of observed positive associations in males and females for neoplasms of the esophagus, stomach, pancreas, and prostate was unlikely to have been generated by chance factors alone, and thus, may have a biological basis related to Ingested asbestos. Cancers of the small intestine and leukemia were implicated to a lesaer degree in this analysis. The patterns of integrated findings for most gastrointestinal cancers were somewhat consistent with patterns ob served among asbestos-exposed occupational groups, whereas the patterns found for pancreatic cancer, kidney cancer, and leukemia were not consistent It was recom mended that the integrated ecoiogic data to date be used to generate a rough priority of specific etiologic hypotheses that should be tested in the original settings or in indepen dent study populations using studies designed at the more definitive individual level, such as case-control studies. The Bay Area (California) and Puget Sound (Washington) were deemed to be the existing study areas most suitable for future research. Introduction In 1982, the U.S. Environmental Protection Agency commissioned a critical review of the major epidemiologic studies that were germane to tne question of possible adverse health effects c iused by ingested asbestos. Thirteen published and unpublished studies conducted in five areas cf the United States and Canada were included in the review (1-13). This paper presents the mqjor fadings and salient points of the more detailed review found elsewhere (14). background The genesis of all of the studies included in this i eview was the 1973 discovery of large amounts < f amphibole asbestos fibers in Lake Superior, the 'Department of BioaUtiatia. Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261. source of municipal water for Duluth, Minnesota, and five small communities on the lake shore. The first epidemiologic study to appear alter this discovery was conducted in Duluth by Mason et al. In 1974(1). Mason's study of 1950-1969 cancer mortality rates was followed by two studies of cancer incidence rates in Duluth, the first in 1976 by Levy et al. (2) and the second in 1981 by Sigurdson et al. (3). The two Connecticut studies of Harrington et al. in 1978 (4) and Meigs et al. in 1980 (5) were prompted by the possibility of studying reliable cancer incidence data over a 35year period through the Connecticut Tumor Reg istry and linking these data with information collected on asbestos-cement pipe studies done by the U.S. EPA. In Canada, the mortality studies of Wigle in 1977 (6) and Tbft et al. in 1981 (7) were induced by the extent of the asbestos mining done in Quebec and by environmental surveys that revealed high concentrations of asbestos fibers in the drinking water supplies of certain cities. The ttsimuci San Francisco Bay Area cancer incidence studies of Kanarek et al. in 1980 (8), Conforti et al. in 1981 (9), and Tarter in 1981 (10) were motivated by the fact that several drinking water supplies come from aquifers or are stored in reservoirs that are exposed to serpentine, the parent rock form of chrysotile asbestos. The single unpub lished epidemiologic study of cancer incidence conducted in Utah by Sadler et al. in 1981 (21) was based upon the fact that several Utah com munities were known to have used predomi nantly asbestos-cement pipe for periods exceeding 20 years. Finally, the two studies of cancer inci dence and mortality in the Puget Sound area by Severson et al. in 1979 (22) and Polissar et al. in 1982 (13) were motivated by the fact that three of the largest metropolitan areas of western Wash ington state have been almost constantly serviced since the early part of the 20th century by water supplies containing a wide range of chrysotile asbestos fibers. Individual Reviews and Qualitative Integration of Findings For each of the 13 studies, a determination was made of an overall positive, negative, or lack of association between ingested asbestos and cancer mortality or incidence. Determinations were based on general epidemiologic considerations while accounting for the strengths and weak nesses of the underlying study designs. Due to the subjectivity inherent in the assessment of re search findings, the interpretations made were not always those of the authors cited. Ihbles 1 and 2 show for gastrointestinal and nongastrointestinal cancer sites, respectively, a summary ofresults from the 13 studies. As shown here, one or more previous studies have found for males or females some association between asbes tos in water supplies and cancer mortality (or incidence) for neoplasms of the esophagus (1,8,9), stomach (1J2.6-9), small intestine (13), colon Ibble 1. Summary of studies of gastrointestinal cancer risk in relation to inflated asbestos by cancer site.* Gastrointestinal cancar site. <ICD 7th revision code*) Maaon i/i Duluth Connecticut Quebec Levy Sigurdaon Harrington Meip Wigie tbft 12) (3) (4) (5) IS) (7) Bay Ana. CA Utah Puget Sound. W'A Kanank Conferti Tarter Sadler Severson Potife-ar >4) (9) 110) (//) (/2) /J' All him combined (150-159) Eeophague i150i Stomach (151> Small intestine ( 152) Colon 11 S3' Rectum i154) Biliary paaeap liver 11 Si-1 MAI Gallbladder' 155. U Pancrea* '157) Peritoneum lSt 1**1 (V -) l*1 ns lOOl <*-*> lOOl i- -i i00) (- 0) i00) (001 (001 ns (On ns (001 I 1 tOOl 1001 100) (00) 100) toot (00) 100) 1001 lO <001 (IS ns (00) 1 *0) l 1 <-*- +\ { 1 ns (00) ns ns BS (00) (00) (0 + ) (+ +) ns ns ns (00 (00) (001 (-0) 1 + 0) ) (+ +) ns 100) (00) ;00. ns M as ns (00) (00) ns (00) ns 1 - - 1 100) (00) (00) (00) (00) ( + 0) ns (0-1 1- ~) (OOl 100) (00) 100) (00) (00) (00) ns (00) ns <00< ns ns ns ns (00) (00) ns ns ns 00' ns OS ns ns (0 + 1 (00) ns (0+) ns i00> ns 1+0) <0-1 lOOl <0-1 < -el ns lOOl ns '00' ns ns ns ns (0+) ns (00) ns >00' `'Mile, female * association with ingested asbestos: + positive, 0 none, - negative, ns * not studied. Table 2. Summary of atudiee of nongastrointestinal cancar risk In relation to ingested asbestos by cancar site.* Nonga.*trointe*iift*l cancer ;ute *ICD 7th revision codes! bal's 4 4B IS animait.* >, dence spj>* the two D\l Maaon li) Duluth Connecticut Quebec Levy \Sigurdaon Harrington Meip Wigte Tbft \2) (J) (4) (5) (81 in i ns ns ns 100) (00) Bay Aren. CA Utah Puget Sound. Kanarek Conibrti Tkrter Sadler Severson PoliSJ. 18) (9) (70) i/i) 1/2) i/2< ns ns ns ns ns (00> s <00) ns lOOl ( *0) < +0) ' *01 (00) ns ns ns (00 88 s na ns s ns # ns ta ns IS ns ns ns ns na (0-1 (0-e 1 ns ns ns ns ns ns 0 0 0 - ns ns ns . ns (001 (OOl (00) (0-1 (00) ns \*0i `00i ns (00) (00) (00) (001 (00) ns ns ns ns ns (00) (00) (OOl <001 ns ns ns 1 - ns ns ns ns (001 (001 ns ns ns > - - M ns ns ns (00) lOOl (00) (001 ns i +0) ns 1 - - ted aabattoa: - poaitiv*, 0 none, - negative, ns * not studied. REVIEW OF EPIDEMIOLOGIC STUDIES 51 *.2J),11, 12), rectum (1), gallbladder (8,11), pan creas (1-3,5,8,8,9), peritoneum (5,9), bronchus, trachea, or lungs (1,6-8), pleura (8, 9), prostate 9,13), kidneys (8,11), brain or central nervous system (13), thyroid (13), and leukemia or aleuke mia (11,13). The large variability in findings evi dent among the studies is matched by a consider able descrepancy in results for males and females within the 13 studies. Several factors might ex plain, at least in part, the internal and external inconsistencies in results. First, the descrepant results may be due to differences in characteristics of asbestos exposure in the various study populations. These differ ences are summarized in Table 3. The relatively low number ofpositive associations found in Utah < 11) and Connecticut (4,5) could be due to the low concentrations of asbestos in the drinking water r r to the relatively short duration of community -xpoaure in several study subareas. The virtual absence of positive findings in the most recent Duluth study (3) could also be due to relatively short duration of exposures as well as the amphil ole fiber, which is fundamentally different from i he chrysotile fibers found in the remaining study r.reas. By utilizing the differences in exposure i haracteristics, the three study areas associated - ith long duration of exposures (> 40 years) to chrysotile asbestos can be roughly ranked accord ing to the concentration of fibers in their water systems. However, the resulting ranking, Bay Area (lowest), Puget Sound (intermediate), and Quebec (highest), does not appear to be related to the pattern of associations shown in Tables 1 and 2. In addition to duration and intensity, it is also likely that other exposure factors, such as the characteristics of asbestos pipe used, the concen tration of other possibly carcinogenic contami nants of water, and certain physical properties of r sbestos fiber (e.g., length), vary among and ithin the six study areas. As a second major factor, the different study designs employed in the various areas, coupled with the disparity in their underlying strengths and weaknesses, most likely also contributed to the observed variability in results. The most im portant methodologic weaknesses and limitations ascertained from the individual reviews are sum marized in Table 4. The weaknesses are listed in approximate decreasing order of importance rela tive to their potential impact on the credibility and definitiveness of the findings. By far the most serious limitation of all the studies conducted to date is that they are ecologi cal or, more specifically, geographic correlation studies by design. This drawback alone does not permit a definitive conclusion to be made from any of the studies of the possible adverse health effects of ingested asbestos. The major drawback of ecological analysis for testing etiologic hypoth eses is the potential for substantial bias in effect estimation. This problem, known as the "ecologi cal fallacy* results from making a causal infer ence about individual phenomena on the bias of observations of groups. Theoretically, the bias resulting from ecological analysis can make an association appear stronger or weaker than it is at an individual level; however, in practice, this bias ordinarily exaggerates the magnitude of a true association, if one exists (15-17). Ecoiogic study bias can be minimized, for example, through the judicious application of ecoiogic re gression techniques. Such techniques were em ployed, at least inpart, in the Connecticut study of Meigs (5), the three Bay Area studies (8-10), and the two Puget Sound studies (12, 13). How ever, the overall variability in results does not appear to be any less among or within these six studies compared to the remaining seven, which did not incorporate more refined ecoiogic analy ses. Much of the bias inherent in ecoiogic analysis results from the inability to control for confound ing factors at the individual level. Table 4 showsthat most of the studies reviewed did not directly control for confounding factors even at the group level. Notable exceptions are the Bay Area stud ies of Kanarek et al. (8) and Conforti et al. (9) and the two Puget Sound studies (12,13), which em- ST0096992 Table 3. Characteristic* of asbestos exposure* in drinking water in various itudy populations. Characteristic Duluth Type of asbestos Number of flbera/L,-i Topulation exposed Maximum duration of exposure, yr Amphibole 1.0-30.0 - 10* 100.000 15-20 *BDL below detectable limit. ^dls. date not available. Connecticut Chrysotile BDL-0.7 - 10* 576,800 23--U Quebec Chrysotile 1.1-1300 - 10* 420,000 >50 Bay Area, CA Chrysotile 0.025-36 - 10* 3,000,000 > 40 Utah Chrysotile n.aA 24,000 20-30 Puget Sound, WA Chrysotile 7.3-206.5 - 10* 200,000 > 40 ST0096993 S3 C. Af. HARSH Table 4. Summary of methodologic weakness** and limitation* aiaociated with varioua itudlaa of ingested asbestos.* Weakneselimitationt Duluth Connecticut Quebec Bay Area. CA Utah Puget Sound. WA Tbul Mason Levy SigurdsonHarrington Meigs Wigls Tbft Kanarek Conforti Tsrter Sadler Savereon Polissar icrou U) i2> iJ> (4) (5) id) <71 id) 19) HO) HI) 113) tl3) 'tudit* Ecolofic study design Insufficient latency period Death certificate date Duration and'or intensity of txposure low Uncontrolled confounding Race Sex Occupation Socioeconomic status Population density Ethnicity In/otit migration Personal habits Abaenc* lor incompistat data on do**-response Multiple comparisons problem Insensitivity of summary statistic* Absence of historical asbestos exposure data Use of at least on* questionable tututici] procadun -- m " * * * -- - - -- - -- * * -- - - 13 - -4 - -3 - -s - 10 ---- ---- -- ---- -- 1 - 0 - 10 - m - - -- - 7 - m 0- 10 m - 11 * - --s 13 " -- * P- 8 ] 12 0 -- - 10 13 _2_ _2_ 4 Tbtai 14 IS 14 12 12 u a 7 7 11 10 10 9 `Legend: asterisk ' *> indicates presence of characteristic: minus I - > indicates absence of characteristic. *!n approximate decreasing order of relative impact on dednitivsnsss of study results. ployed relatively more sophisticated multivariate statistical analyses as an attempt to control for confounding at the group level. Only one study to date, that of Polissar et al. in 1982 (13), attempted to collect data on a confounding variable at the individual level; however, since this was done only for cancer cases and not controls, it was not possible to analyze the data on a more sensitive and reliable case-control basis. Occupation was a particularly important con founding variable in the studies conducted in Quebec (6,7), the Bay Area (8-10), and Connecti cut (4), since a substantial number of males are employed in the various asbestos-related indus tries within these areas. The confounding effects of occupation are particularly evident in the two Quebec studies (6,7), where positive associations for lung and stomach cancer were consistently confined to males. Misclassification of asbestos exposures is an other serious limitation of all the studies con ducted to date. This misclassification results from several factors including: the basic ecologic de sign, which assigns specific exposures to an entire goegraphic area; tenuous assumptions regarding the extent of asbestos contamination from asbes tos pipes; the lack of any reliable historical asbes tos exposure data; and the in/out and daily mobil ity of the study populations. It is also likely that many of the associations found among the 13 studies are simply chance occurrences arising from the large number of statistical comparisons that were generally made. Whenever a large number ofsignificance tests are performed at a constant significance level, a cer tain number of tests will be significant by chance alone and the actual significance levels must be higher than those reported by the authors. Among the 13 studies reviewed, the number of separate statistical comparisons reported ranged from 33 to 336 with an average of 193. Therefore, at a 5% level of significance, the number of posi tive findings expected due to chance alone would range from approximately 2 to 17 with an aver age across the 13 studies of about 10. In other statistical terms, the probability that at least one of the n independent comparisons was due to chance alone ranged from 0.81 in a study report ing about 30 comparisons to virtual certainty in studies reporting 100 or more comparisons. (At the 5% level of significance, the probability of falsely claiming statistical significance in at least one of n independent comparisons is l-0.95\ Objective integration of Findings In order to objectively evaluate the extent to which the pattern of findings to date may be due REVIEW OF EPIDEMIOLOGIC STUDIES 53 to chance factors, and to better assess the degree of interstudy consistency, a probability analysis was performed for each cancer site, which was examined in at least four independent studies. The studies of Levy et al. in 1976 (2), Harrington et al. in 1978 (4), Kanerek et al. in 1980 (8) and Severson in 1979 (12) were not considered inde pendent studies, since they provided no unique information in light of the subsequently updated and improved analyses of Sigurdson et al. in 1981 (3), Meigs et al. in 1981 (5), Conforti et al. in 1981 (9) and Polissar et al. in 1982 (13), respectively. In addition, the study of Thrter in 1981 (10) was not included in the probability analysis, since no can cer site-specific results were shown. For each cancer site, the probability analysis consisted of first casting the independent study results of Thbles 1 and 2 into a 2 x 2 contingency table of male-female results as shown in Table 5. The next step in the analysis was to calculate for each cancer site the probability ofjointly ob serving in n. independent studies, n,. or more positive associations in males and or more positive associations in females. This was done assuming that for males and females the proba bility of observing a positive association in a given independent study due to chance alone is p = 0.05, and the probability of observing no asso ciation is (1 - p) 0.95. Designating M and F to represent the events of observing a positive asso ciation in males and females, respectively, and assuming that outcomes in males and females are independent events, the probability of the joint event (known as a large deviation probability, Pa) can be calculated as the product of two individual cumulative binomial probabilities as follows: PD = Min,.) PCFfcn,) PD was also calculated by using the binomial parameter p - 0.10 assuming that a predeter mined significance level of p ~ 0.05 would have actually been higher for any individual observed positive association due to the very large number of statistical comparisons that were made in most of the independent studies. Very small values of Male association Table 5. (f) (0 or -) Ibtai Famale association ( + ) (Oor -) Tbtal nu n\a n * i nj i. t n P0 (less that 0.05, for example) for a given cancer site suggest that the number of observed positive associations in males and females across several independent studies was unlikely to have been generated by chance factors alone, and, therefore, may have a biological basis related to ingested asbestos. The PD value as calculated above does not, however, take into account the degree of association between male and female findings. Unfortunately, the very small numbers of inde pendent studies showing results for specific can cer sites precluded the calculation of any reliable measure of association. However, in order to pro vide at least a crude objective comparison of the level of agreement between male and female find ings, the well-known phi coefficient given as 4> - (x.1/* )w was computed where i the uncorrected chisquare statistic tabulated from the above 2x2 contingency table as XJ - n (nuna - nl2n,,)* Values of close to zero indicate little, if any, asso ciation, whereas values close to unity indicate almost perfect predictability. By definition, the phi coefficient cannot be determined whenever ni.> n.i> *1.. or rt.j is equal to zero. Finally, the strength of the association between male and female findings was assessed through the use of the Fisher-Irwin exact test (18). Thble 6 shows the results of the probability analysis for gastrointestinal and nongastrointestinal cancer sites, which were examined in at least four independent studies. Only five of the 14 sites shown in Table 6 (esophagus, stomach, pan creas, lungs, and prostate) are associated with PD values that range consistently below or near a probability level as low as 0.05, for example. However, as shown by the <f> value and corre sponding Fisher-Irwin probability, or by inspec tion ofthe outcome frequencies, the level of agree ment between male and female findings for these cancers is generally moderate to low. Specifically, positive associations were jointly observed in males and females in only one of six studies of esophageal cancer, two of eight studies of stomach cancer, and one of eight studies of pancreatic cancer. It was not possible to quantify the level of male-female agreement for lung or several other cancers due to the presence of one or more zero marginal totals. Two additional neoplasms (small intestine, and leukemia/aleukemia) are associated with Pa val- ST0096995 M C. M. MARSH Table 1 Summary of male-female aaeociatione in Independent studies by cancar site. Cancer nu No. of independent studies in ) 1 + *1 <!!> Outcome (+ 0) 10* l iOjjl 100) 1*22* Toul malt i) <n L > Gastrointestinal Esophagus Stomach Small intastina Colon Ractum Biliary passages liver Gallbladder Psncraas Ptritonaum Nongastrointestinal Bronchus, trachea, lungs Kidneys Bladder Brain/CNS Lcukemiraleukemia Prostata 'males only! 6 8 4 8 8 4 4 8 4 7 8 S 5 8 4 1 1 04 2 2 204 4 1 003 1 0 l07 1 1 007 l 0 004 0 0 0 13 0 1 133 2 0 0t30 0 304 3 0 1 0S l 0 0 03 0 0 l04 1 0 204 2 - --- 2 *NC " not calculated dua to presence of ona or mort taro marginal frequencies. Tbul fern*It (m (ntl 1 .2 1 1 1 0 l 4 1 0 0 0 0 0 - Large deviation probability <PDl <p 0.061 Ip 0.10) Index of assoo.-tion Fishtr-fmnn * probability 0.0087 <0.0001 0.0344 0.3386 0.1132 1.00 0.1856 <0.0001 0.186S 0.0038 0.2649 1.00 02282 0.0328 0.0140 0.0533 0.0009 0.1183 0.3693 0.3243 1.00 0.3439 0.0009 0.3439 0.0257 0.4686 1.00 0.4093 0.1143 0.0523 0.63 0.55 1.00 NC* 1.00 NC NC 0.0 NC NC NC NC NC NC - 0.33 0.21 0.25 NC 0.12 NC NC 0.78 NC NC NC NC NC NC - ues below 0.05 when based on the binomial pa them cancer of the colon and rectum. This same rameter p * 2.05, but exceed P0 0.05 when feature is suggested in Table 6, where upper gas based on the more conservative p 0.10. While trointestinal cancers are among the strongest still based on very small numbers of independent positive results, whereas positive associations for studies, the P0 values for the remaining cancer colon and rectal cancer are virtually nonexistent. sites examined suggest that the number of posi The relatively large number of independent posi tive male and female associations, if any, ob tive associations found for pancreatic cancer sug served for these cancers is more likely to repre gests a possible link with ingested asbestos, al ! sent chance phenomena. It should be recognized that, next to very small though most occupational studies have not implicated this cancer site. sample size, the most severe limitation of the With respect to nongastrointestinal neoplasms, above probability analysis was the necessity to an increased risk for cancer of the kidneys has assume that the n independent studies provided been found in a recent occupational study of insu qualitatively and quantitatively equivalent infor lation workers (23). A biological basis for this risk mation toward the integration of findings for any has been described by Cook and Olson (24). How given cancer site. Therefore, the results of the ever, as shown in Thble 6, kidney cancer was probability analysis should not be regarded as observed in excess among males in only one of the i conclusive, but rather should serve as a rough guide for the future direction and emphasis of six independent studies reviewed that examined this anatomic site. It is uncertain whether the research. marginally significant number of leukemia/aleu kemia and prostatic cancer findings are related to Relationship to Occupational Studies ingested asbestos, since these are generally not considered in occupational studies as sites where asbestos-induced cancers would occur. The pattern of integrated findings presented for gastrointestinal cancers is somewhat consistent Recommendations for Future with patterns observed among workers occupa tionally exposed to asbestos. Epidemiologic stud Research ies of several occupational groups exposed to as Although no individual study or aggregation of bestos have shown an increased incidence of studies exists that would establish risk levels i cancer of the esophagus, stomach, colon, and rec tum and of peritoneal mesotheliomas (19-22). from the ingestion of asbestos, the studies to date do provide extremely valuable information that Furthermore, as noted by Mason et al. in 1974 (1), should be carefully considered when developing certain studies of asbestos installation workers in the protocols of future research. the United States have shown cancer of the upper First, the integrated study findings can be used gastrointestinal tract to be in far greater excess to generate a rough priority of specific etiologic i i REVIEW OF EPIDEMIOLOGIC STUDIES 53 ST0096996 hypotheses that could be tested in the original settings or in independent study populations via more sensitive and reliable epidemiologic de signs. The foremost intensive efforts should be made to further study the relationship of in gested asbestos to the gastrointestinal neoplasms :hat displayed the most suggestive findings in the jcologic studies. In approximate order of impor tance, these would be stomach, pancreas, esopha gus, and small intestine. The outcomes of these endeavors could be used to determine whether additional studies of other gastrointestinal neo plasms were warranted. In addition, the integra ted findings for prostatic cancer, although less biologically plausible, were sufficiently disconerting to make the relationship of ingested as!>estos to this male neoplasm the subject of an>ther more intensive study. Second, the existing studies have produced a virtual checklist of methodologic limitations and uncertainties that should be avoided or controlled :o the fullest extent possible in all future research i fforts. Many of the aforementioned weaknesses ; nd limitations can be avoided by simply choos ing more suitable geographic areas for study. The " ideal" study area would be one associated with a 1 mg history of a wide range of asbestos exposures cf known and well-documented magnitude. This would allow a sufficient latency period for the development of disease and would permit a more sensitive and accurate assessment to be made of dose-response relationships. While none of the reas' studied to date can be necessarily consid ered as ideal, the Bay Area and Puget Sound are relatively the most suitable areas for future res -arch. Further studies in new independent areas should also be considered since this will improve the ability to evaluate the strength and consiste \cy of findings statistically. Many of the other methodologic limitations are f>atures of the underlying ecologic study designs that were employed. The ability to make a causal iterance from ecologic data often can be en hanced using more sophisticated analytical tech niques. There will always remain an element of uncertainty, however, until the etiologic hypothe ses generated from ecologic studies are tested more definitively at the individual rather than group level. The diseases implicated in the ecologic studies to date are relatively rare in the general populat on and are associated with long incubation perods. Thus, the retrospective approach is apropriete using, for example, either an unmatched or matched individual case-control design. Basi cally, a case-control study would compare the ingested asbestos exposures of individual sitespecific cases of cancer (incidence or mortality) with unmatched or matched controls. This ap proach would enable a much more precise mea surement of confounding factors such as occupa tion, socioeconomic status, tobacco and alcohol consumption, dietary habits, and migration his tory through personal interviews with each case (or next of kin) and control. While the level of asbestos exposure would probably still be deter mined by geographic residence, the duration of exposure could be much more accurately mea sured and controlled by determining length of residence. In addition, individual differences in water ingestion habits due to daily mobility and other personal factors could be assessed during the interviews. It is very important that the casecontrol protocol include procedures for checking the reliability and validity of the methods used to ascertain historical ingested asbestos exposures. The number of subjects to be selected for a study of a specific disease-exposure relationship will be a fundamental consideration in planning future studies. Basically, an answer to the ques tion of how many subjects should be selected for a case-control study, for example, depends on the specification of four values: the relative frequency of exposure among controls in the target popula tion p0; a hypothesized relative risk associated with exposure that would have sufficient biologic or public health importance to warrant its detec tion R; the desired level of significance a; and the desired study power, (1 - P) (25). As an illustra tive example, Thble 7 shows for a standard un matched case-control design the required sample size n (per group) under the conventional a * 0.05 (one-sided), fi = 0.20, and for selected values of R and p0. In the study areas recommended for individual case-control analysis (the Bay Area and Puget Sound), relative risk levels R for gas trointestinal cancer were generally found by eco logic analysis to be only moderately elevated (J? = 1.1-2.0), if elevated at all. This is likely to be the case in most areas unless levels of asbestos in the drinking water are inordinately high. In order to detect these putative moderate elevations in rela tive risk at acceptable statistical error levels (a, 0), it will be necessary, as shown in Ihble 7, to study literally hundreds of cases and controls. This may be a serious drawback when studying the rarer forms of gastrointestinal cancer (e.g., small intestine), for it may be difficult to observe and locate the required number of cases during a reasonable period of time. For some cancers, therefore, it may be necessary to accept some what higher levels of statistical errors in order to sas-r- ST0096997 96 G. M. MARSH Table 7. Unmatched case-control sample sizes needed in each group for a 0.06 (one-sided) and 9 * 0.20. Relative risk# 1-2 1.5 1.7 2.0 2.5 3.0 5.0 Proportion of controls exposed, p. 0.25 0.50 0.75 0.90 1890 365 208 119 66 45 20 1484 303 179 107 63 46 24 2071 447 272 168 104 78 45 4431 987 610 386 248 187 114 test the null hypothesis of no risk with the avail able number of cases. In conclusion, there is no question that studies designed at the individual level, such as casecontrol studies, are now needed to establish firmly risk levels to ingested asbestos. However, as illustrated above, the costs of reliably estab lishing these risk levels will be high, a fact that should be recognized by the sponsors and investi gators of future research in this area. Funding for thu work *u provided by the Center for Environmental Epidemiology, Gradual* School of Public Health, University of Pittsburgh. Pittsburgh, Pennsylvania, and the U.S. Environmental Protection Agency, Cincinnati, Ohio, under Cooperative Agreement No. 806615. The research described in this paper has been peer and administratively reviewed by the U.S. Environmental Protec tion Agency and approved for presentation and publication. Mention of trad* names or commercial products does not constitute endorsement or recommendation for use. REFERENCES 1. Mason, T. J., McKay, F. W., and Miller, R. W. Asbestoelike fibers in Duluth water supply, J. Am. Med. Assoc. 228: 1019-1020 (1974). 2. Levy, B. S., Sigurdson, J., Mandel, E., Landon, E., and Peanon, J. Investigating possible effects of asbestos in dty water Surveillance of gastrointestinal cancer inci dence in Duluth, MN. Am. J. Epidemiol. 108: 362-366 (1976). 3. Sigurdson, E. E.. Levy, B. S., Mandel, J., McHugh, R, Michienxi, L. J., Jagger, H., and Pearson, J. Cancer mor bidity investigations: Lessons from the Duluth study of possible effects of asbestos in drinking water. Environ. Res. 25: 50-61 (1981). 4. Harrington, J. M., Craun, G., Meigs, J. W,, Landrigan, P. J., Flannery, J. T., and Woodhull, R S. An investigation of the us* of asbestos cement pipe for public water supply and the incidence of gastrointestinal cancer in Connecti cut, 1935-1973. Am. J. Epidemiol. 107: 96-103 (1978). 5. Meigs, J. W., Walter, S., Heston, J., Millette, J. R, Craun, G. F., Woodhull, R S., and Flannery, J. T. Asbestoscement pipe and cancer in Connecticut, 1955-1974. J. Environ Res. 42:187-191 (1980). 6. Wigle, D. T. Cancer mortality in relation to asbestos municipal water supplies. Arch. Environ. Health 32:1' 189 (1977). 7. Ibft, P., Wigle, D. T., Meranger, J. C. and Mao, Y. Asbe? and drinking water in Canada. Sci. Tbtal Environ. 18:7 89(1981). 8. Kanarek, M. S., Confetti, P., Jackson, L., Cooper, R and Murchio, J. C. Asbestos in drinking water and car.< incidence in the San Francisco Bay Area. Am. J. Ep < miol. 112:54-72(1980). 9. Conforti, P. M., Kanarek, M., Jackson, L, Cooper, R. and Murchio, J. C. Aabeetoa in drinking water and car.c in tha San Francisco Bay Area: 1969-1974 incidence. Chronic Dia 34: 211-224 (1981). 10. Thrter, M. E. Pattern recognition in the context of ; asbestos cancer threshold study. Preprint, Proceeding > the 13th Annual Symposium on the Interface: Comput Sci. Statistic. (1981). 11. Sadler, T. D., Rom, W. N., Lyon, J. L, and Mason, J. The use of asbestoe-cement pipe for public water iuj.p and the incidence of cancer u selected communities Utah, 1967-1976. Graduate thesia. University of Uia 1981. 12. Severson, R A study of the effects of asbestos in drink ir water and cancer incidence in the Puget Sound rep :o M.S. Thesia, Univ. of Wash., Seattle, Washington (1979' 13. Polisaar, L, Severson, R K,, Bostman, E. S., and Thor-.a D. B. Cancer incidence in relation to asbestos in drinkm water in the Puget Sound Region. Am. J. Epidemiol, j 1' 314-32811982). 14. Marsh, G. M. Review of epidemiologic studies related ' ingested asbestos. Report submitted to the Center f. Environmental Epidemiology, Graduate School of Pub! Health, University of Pittsburgh and the U.S. Enviror mental Protection Agency, October 12,1982. 15. Lengbein, L. I., and Lichtman, A. J. Ecological Infereict Sage Publications, Beverly Hills, CA, 1978. 16. Duncan, O. D., Cuzzort, R P., and Duncan, B. Statist.ca Geography: Problems in Analyzing Areal Data. Green wood Pnss, Westpoint, CT, 1961. 17. Valkonen, T. Individual and structural effects in ecol'gi cal research. In: Social Ecology (M. Dogan and S Rokbran, Eds.), MIT Press, Cambridge, MA, 1969. 18. Hollander, M., and Wolfe, D. A. Nonparametric Statistics Methods. John Wiley and Sons, New York, 1973. 19. Selikoff I. J., Churg, J., and Hammond. E. C. Asbe.-to: exposure and neoplasia. J. Am. Med. Assoc. 188: 22--2t (1964). 20. Enterline, P. E., and Kendrick, M. A. Asbestos dust expo sures st various levels and mortality. Arch. Environ Health 15: 181-186(1967). 21. Newhouae, M.LA study of the mortality ofworkers ir an asbestos factory. Brit J. Ind. Med. 26: 294-301 (1969) 22. Elmet, P. C., and Simpson M. J. C. Insulation worker- in Belfast A further study of mortality due to asbe.-tos exposure. Brit J. Ind. Med. 34: 174-180 (1977). 23. Selikoff, I. J., Hammond, E. L., and Seidman, H. Mortality experience of insulation workers in the United States ?nd Canada, 1943-1977. Ann. N.Y. Acad. Sci. 330: 91-116 (1979). 24. Cook, P. M., and Olson, G. F. Ingested mineral fibers elimination in human urine. Science 204: 195-198 (19"9) 25. Schlesslman, J. J. Sample size requirements in cohort a ad case-control studies of disease. Am. J. Epidemiol. 99' 3t 1384(1974).