Document MwLz3Qp1db3kNGx33DRm975M

.-. `3 i related to a p r e n a t a l y evelopment. '.ad Occupation Stewart in the grain industry. M and M. Mastem (NCI. VCaJaing industry am 2 -cor rbb aM0ci.d y found in other IMiculusl past control. Tbar ~lphorphib, ethylene dit, malathion, and methyl. bration of Grain Miller'# recorda of 40,247 wortcomponenta of the grain d as the basis for a cao. nal mortality analysin to imecticides. In additioa d c i d e use waa obtained survey of 6,000 current of the industry and &ing companies. Mortality for non-Hodgkin's I y m grain milling and a t o w o US white male pop&iponenta of this ind3s risk. In the nested camIodgkin'elymphoma c a m id storage facilities (& ? interval = 1.6-46.0.p < ?need their greatest CXCCI `s lymphoma during Lb nalysea are evaluating tb )f individual insecticider xrpal tunnel syndrome in .Morgenstern,' J. K r a u . I (UCLASchool of Publir 024). is a compression neuropa. :he wrist. The syndrome in pain, discomfort, and sen&&. The purposes of this to determine whether thr ry checkers increases the inre and to identify specific were mailed in 1986 union in California. Oft h r 055 were women aged 18ible questionnaires. In thr linical or electrodiagnostic mpared several carpal tun- ased on four eelf-reported f several testa of construct ie of these measures-the ~ ~ B - W M judged to be thr me. The prevalence of thir c all female subjects. With iysis and multiple logistic md that carpal tunnel syn. I independently related lo ~genumber of work houn UI a checker. There waa no SOCIETY FOR EPIDEMlOLOClC RESEARCH. ABSTRACTS 901 effect, however, of specific work-related behaviors, was observed in any cane ICD rubric w including the use of laser scanners. with their matched cancer controls. Occupational and environmental lead exposure and the risk of gout. B. Richter,' M. Hochberg, and M. Monk (The Johns Hopkins U.School of Hygiene and Public Health. Baltimore, MD 21205). An apaociation between lead poisoning and gout has been suggested. A cane-control study was designed to investigate the relation between lead exposure and gout in a population of subjects who did not have lead poisoning. Cases ( n = 100) and controls (n = 100) were selected from patients seen in the Johns Hopkins Hospital emergency mom. Baltimore, Maryland dur- ing 1983-1986. Cases met the American Rheumatism A case-control study of non-Hodgkin`s 1 agricultural factors in emtern Neb PLAINTIFF`S EXHIBIT Zahm* (NCI. Beth& MD 20892). ger. P.A. Babbitt, R. C. Saal, K. P.. Blair. - A recent study conducted in sixfold excesa risk of non-Hodgkin'r lymphoma.mow fanners expooad to agricultural herbicidm20 01man days per year. To further inwatigate the amociation between non-Hodgki's lymphoma and facton, a population-bawd cam-contd rsuQ w u conducted in -tern Nebraska. Telephona intarvim Association'srecommended criteria for gout. Men who worked 20 or more years in jobs classified by industrial hygienists Mhaving high lead exposure were a t highest rink for gout (odds ratio =6.0). The odds ratio WM 2.5 for jobs clasaified as having medium exposure to lead, regardleas of duration. Plumbers, painters, and shipbuildem or shipbreakers had statistically elevated rbka for gout. Men with gout were more likely to pump leaded gasoline than were the controls, although the risk was not statistically significant. There was a statistically significant association between "moon- shine" alcohol consumption and gout, although beer, were conducted with 385 histologically c0-d cases (201 men, l&-women) aad 1,432 controk (725 men, 707 women). Among men, use of the hm&i& 2,4-D was aurociatcd with 8 nonaignifiwt 60% increased risk of non-HodgWa lymphoma (& mtio (OR) = 1.5,95% confidence interval = 0.9-2.4). Bt- - -p u r e to 2.4-D more than 20 dam/year incmmdrirk threefold (OR 0 3.3,95% CI = 0.6-22.1). Urn of thc herbicide 2,4,5-T had M odda ratio of 1.6 (95% CI 0.7-3.2). Risk among atrazine w n (OR 1.4. 95% CI 5 0.8-2.2) increased with duration ( O bof 0.9,0.8, 2.0, and 2.0 for use for 1-5.6-15.lb20, and 21+ y w , wine, and liquor were not associated with the disease. Lead intoxication secondary to moonshine consumption han been documented. Adjusting for obesity, subjects with gout ate lead contaminated foods more frequently than did controls. Vietnam military service and risk of selected cancers. B. GOUP (Argonne National Laboratory, Argonne, IL 60439)and L. H. Kuller. Public concern about adverse health effects follow- respectively). Several c l ~ ofs hecticidea were associated with increased risk c h l ~ b t hd w m (OR = 1.4,95%CI = 0.8-2.31, carbamatcr (OR = 1.8, 95%CI = 1.0-3.21,and organophosphates (OR = 1.9, - -95% CI 5 1.1-3.1). specific inSeCticide8 auaciated with significant excess non-Hodgkin's lymphoma included chlordane (OR = 2.11, diazinon (OR ZO), dyfonate (OR = 2.4). and malathion (OR 2.2). Risk rose with increasing duration of use of lindane and sevin. Multiple exposures are evaluated ing exposure to 2.3,7,8-tetra-chlorodibenzo-p-dioxin contaminated defoliants used by U S Armed Forces in Environmental exposures in cytogenetically &tined the Vietnam War has not abated since the 1970s. subsets of acute nonlymphocytic leukemia. M. M. While Swedish studies support the association of soft- Crane,' M. J. Keating, J. M. Trujillo, D. R. Labarthe, tissue sarcomas, malignant lymphomas and phenoxy and R. F. Frankowski (U. of Texas, School of Public s herbicide exposure among forestry and agricultural Health, Houston, TX 77225). workers, studies of US Vietnam veterans have yielded Most newly diagnosed, untreated acute nonlympho- -controversial and conflicting results. Vietnam military cytic leukemia patients have detectable bone marrow service and lifetime overall military service of 349 chromosomal abnormalities. which define prognosti- Pennsylvania men who died during 1968-1953 at ages cally distinct groups and may be related to etiology. 18 to 50 years of soft-tissue sarcomas (International Environmental exposure profiles were obtained for Classilicationof Diseases,Ninth Revision (ICD-9)code 127 acute nonlymphocytic leukemia patients with 171). non-Hodgkin's lymphomas (1CD-9 codes 200. marrow abnormalities and 109 acute nonlymphocytic 202). and other rare cancers (ICD-9 codes 158. 164, leukemia patients with cytogenetically normal mar- 203) were compared with two matched control row. These patients represent 73% of newly diagnosed groups-349 males deceased from other cancers and 349 living males with driver's licenses. Vietnam and overall military service were ascertained from death certificate, next-of-kin questionnaire, and official U S military personnel record data. No excess Vietnam military service was seen amonR cases conipared with either control group: 20 cases (5.7961, 26 cancer controls (7.4%) and 18 driver's license controls ( 5 . 2 % ) had served in Vietnam. Overall lifetime US military seervice was established for 147 cases (42.1%) and 171 (49.0%)deceased cancer controls. Matched pair analyria failed to show an association for Vietnam military service or lifetime total military service. NO statistically significant excess of Vietnam military service acute nonlymphocytic leukemia patients treated at M. D. Anderson Hospital and Tumor Institute between January 1976 and May 1983. Odds ratios adjuakd for age, sex. and other exposures by logistic regression were computed for both sets of patients. Patienta with marrow abnormalities were more likely to report cy- totoxic treatment for a prior medical condition (oddn ratio (OR) (95% confidence interval) = 4.25 (1.38. 13.17);smoke cigarettes, OR = 1.82 (1.02-3.26); &ink alcoholic beverages, OR = 1.91 (1.05-3.48); and work at occupations with possible exposure to mutagens, OR 0 2.55 (0.51-12.83).Amwiationa between apacific abnormalities and prior cytotoxic therapy, smoking, and current occupation wen found in the ma 902 SOClEfI'Y FOR EPIDEMIOLOGIC RESEARCH: ABSTRACTS whole. Associationsrbetween specific abnormalities and crash severity, circumstances of the crash includand occupational exposure to benzene, pesticides, and ing involvement with a motor vehicle, income, and paints were found in 98 patients. These results are education. ! - SOC (RR 1.6, p 0.05). Alcoha daily drinking yes VI. no) WE ciatedwith risk in men orwon Consistent with earlier studies and suggest that tumor- the relative risk waa elevated specific markers may be present for some exposures The role of behavioral patterns in mediating the ef- = 1.7) and ounces of alcohol in thin disease. fects of tamily psychosocial factors on the probability 2.7 oz (79.8 ml)). Dietary < of pediatric injuries: analysis of the 1981 Child Health against hip fracture. Injury Supplement. Y. Hare]' and E. S. Ford (CDC, Atlanta. Chairman: C. Falvo C A :W:KM). Serum cholesterol and risk Injuries are the largest single cause of mortality and death. J. Pekkanen,. k Nisa Risk factors for traumatic injuries among professional disability among children in the western world. Each J. Karvonen (National Public fm fighters. E. Heineman,. C. Shy, and H. Checkoway year, four out of 10 children in the United States Helsinki, Finland). (U. of North Carolina, Chapel Hill, NC 27599). experience an injury severe enough to require medical The two largest cholesterol Thia case-control study within a metropolitan fire attention. This study investigates the role of childhood department was designed to evaluate the effect of self- behavioral characteristics ("host") in mediating the 4 reported increased mortalit! c violence in treated hypetch contained breathing apparatus and other risk factors effects of family psychosocial factors ("environment") brain semtonine turnsver f on three types of injury a t the scene of a fire (smoke on childhood home and school injuries. The study is possible explanation for the inhalation, burns, and falls). Data on 75 injured fire based on a combination of two theoretical frameworks: Berum cholesterol level and in fighters and 144 controls came from telephone inter- 1) the epidemiology Host-Agent-Environment model, death. This association w a d views and department records. The two sets of unin- and 2) a psychosocial approach focusing on a Stress- =-year follow-up study of m jured fire fighter controls were matched to cases on Behavior-Injury process. Data from the 1981 Child east Finland (n = 788)and: incident (n= 72) or on job position and fire type and Health Supplement on 15,416 children living in the During the follow-up, there w size (n= 72). Smoke inhalation cases were not signif- United States during 1981 were analyzed. Of these death in east Finland and 2 icantly different from controls in self-contained children, 210 reported medically attended injuries that eaet, casea had noneignifican breathing apparatus use, cigarette smoking, previous occurred in the 14-day period before the interview. A l i e serum cholesterol mean fires in the shift, or injury history. Jobs with high risk 32-item Behavioral Problem Index was used to mea- cases (7.00 mmol/l). In the 7 of bums included: nozzle operator, engine officer, and sure behavioral patterns. Path models using a case- reversed (7.01vs. 6.60 mmol/ forcible entry person in first due companies (odds ratio control design and exploratory logistic regression survival analysis including (OR) = 20.1). Other risk factors for burns were base- analysis were applied to the data. Indicators of family and BMI showed a signific ment origin of fire (OR = 10.2). prior fire fighting disadvantage nnd stress affect the probability of inju- association between serum training outside the department (same fire, OR = 11.2; ries, and child's antisocial orientation and lack of traumatic death in the west similar fire, OR = 3.9), and on-duty injury in the prior emotional control are significantly associated with the ciation in the east. T h e ac 12 months (same fire, OR = 4.3; similar fire, OR = probability of injuries. Path analysis showed that host increase in traumatic death 3.5), adjusting for other risk factors. When other risk behavior does not explain away the significant effects lowering trials is probably 6 factors were considered, consistent self-contained of family disadvantage and stress, suggesting that the breathing apparatus use was associated with falls social environment as well as host behavioral factorn --ently unknown, factors orb: . (same fire, OR = 11.8; similar fire, OR = 4.3), but not play significant roles in the pediatric injury process. Self-protection and the risk with burns. Risk of falls was elevated among members ciated with rape. P. March of truck companies (OR = 17.7), and fire fighters Dietary calcium, caffeine, and alcohol and risk of hip without children (same fire, OR = 8.4; similar fire, OR fracture: a 15-year prospective study. T. L. Holbrook,' = 7.4). On-duty injury in the past 12 months was D. L. Wingard, and E. Barrett-Connor (U. of Califor- Mercy (CDC, Atlanta, GA 3 T o explore whether the associated with rape varied. associated with falls when cases were compared with nia, San Diego, La Jolla. CA 92093). protection, the authors exe similar fire controls (OR = 5.5). Neither age, experi- T o assess the effect of dietary calcium, caffeine tional CrimaSurvey, an one ence, nor ethnicity was related to injury. intake, and alcohol use on risk of hip fracture, the victimizations throughout authors performed a retrospective cohort study in a analysis included data fr Field efficacy of bicycle safety helmets. R. S. Thomp- defined population surveyed between 1972 and 1974. women who experienced son, F. Rivara, and D. Thompson' (Croup Health A 24-hour recall of dietary intake was obtained by a women who experienced ntt Cooperative of Puget Sound, Seattle, WA 98121). certified dietitian, and data were coded in 1985 for 1982. Logistic regression w Head injuries comprise 25% of all bicycle injuries nutrient intake by the Nutrition Coordinating Center neously for age. race, and D but result in 65-7055 of such deaths. Helmets are an a t the University of Minnesota. Of the 957 men and age, race, and weapons of an yet unproven preventative measure. The authors women aged 50-79 vears with complete dietary data, persons other than the vie undertook a case-control study to determine the lield 15 iiieii ;rnd 18 woiiieii with hip lructure were identified nature of the victim-offeno efficacy of bicycle helrneta in reducing head injuries. from inortaliLy records and by two follow-up inter- with women who used no 8 Caaes were defined as persons seen in the emergency mom of five major Seattle, Washington state area hospitals for head injury as a result of a bicycle crash. The control group consisted of non-head injured cyclists seen at the same emergency rooms for treatment -'fa ix,~+-%:k=i q & %+A=-- 2 : r :~r _ :L.+ views conducted between 1984 and 1986. All women used self-protection were c were postmenopausal. The risk of hip fracture asso- pleted versus attempted m -ciated with each dietary variahle was adjusted for age, odds ratio (OR) 0.2 (95 -body iiiuss index (weight/height'), smoking (yes vs. 0.14.5); non-forceful -_- -- - _ _n-o).._-..-.k-.ilo.c.a.lo_ri..-es...a-n.-d1, .i1n%w;rorm~exn-,xpcoLst.m=e.n%opaus-al es:- &i (95% CI a 0.1-03); both 0.3 (95% CI = 0.24.51 and controls were obtained by review ol emergency room records and mailed'questionnaire with telephone follow-up. Only 7.6% of the 223 cases wore helmets, -In men and women combined, risk of hip fracture was significantly nnd inversely associated with dietary calcium intake (relative risk (RR) = 0.5 per 406 mg, p tended to increase the ris tion to the rape or attemp protection, OR = 1.7 (957 compared with 24.9% of the 409 controls (crude odds 0.01). For men and women separately, the risk wm self-protection. OR = 01 ratio 0.25, 95% confidence interval = 0.16-0.39). similar (RRs = 0.5 and 0.6, respectively). but was no force and non-force. OR Further analysis is planned to adjust for major sources longer significant. In women but not men, dietary Rape is 8 terrifying and of confounding, e.g., age, sex, riding experience, injury caffeine significantly increased the risk of hip fracture can have devastating p q I .. Expoaun aver cytotoxiclheraw Any ChemotheraplTonb' -Ifradishono~nly .:qamnemoking, No.ofc~ereUesperda~t Nonsmokerat diagnosis 1-20 21-40 do Alcoholconsumphon. OrldS ~ ~atdniagnsosa S0.5 ~ 0 5-1.O 1.la0 >2 n Not exposed None rsporld ~pa~~expxuresll Asbestor Bsnzana kshcides Eyes. gtues. lacquers.varnishes Metals.mew fumes AAlAN (n=126) 14 ( 1 1 ) 5 (4) 5 (4) 70 (56) 23 (23) 17 (13) 10 (8) 33 (27) 40 (36) 18 (15) 19 (16) 84 (66) 12 (9) 31 (24) 17 (14) 6 151 10 (8) 16 (13) 18 (15) NN (n-109) - OddaRattlo (S53CContldona Interval) 5 (5) 1 (1) 2 (2) 75 (69) 20 (18) 10 (9) 2 (2) 49 (48) 36 (35) 7 (7) 9 (9) 2 121 2 6 (0.9-7.2) 4.6 (0.5-40.0) 2 3 (0.4-12.1) .1.0 (. .) 1.6 (0.8-3.0) 1.8 (0.84.2) 5.4 (1 3-21.9) 1.0 1.. .) 1.a (1.0-3.4) 3.8 (15-9.8) 3.1 (1 3-7.6) 5.9 fl.4-25.61 72 (66) 9 (8) 28 (26) 1.0 (...) 1.2 (0.52.9) .(. .) 7 (7) 5 151 15 (14) 17 (16) 12 (11) 2.6 (1.04.5) 1.1 10.3-3.81 0.6 (0.3-1.4) 0.8 (0.4-1.8) 1.5 (0.7-3.2) w:'TOWS of psrcsntngesarslessthan1WYeboauseof mlssingdare.W A N indicates ab~wmal and NN. -l.lt03.1;P- 04). ~ ~ k M r s , ~ g l a S S O s d m n s , o r N v o m i x d d n n k s c ~1l toazionfalcdml MentelTrsndTestforw~spwday. 2 12.44(P<.Wl):odds rat~Oforelchohou(se is2.4 (95%mfidenca inlelvBI. 1 4 Io 4.1;Pc.01). ~ o n a p n o n e x p o s u m~ l e . S ~ t e ~ l f o r ~ ~ p l a ~ t ~ ~ n . 1a-j on cswondentswho kosw whst the sutstanmWBS and w r e suret! mpabonal exposure dld or did no1 m r . Only 54%and 75%of all respondems had aasptable informatton for benzene and asbestos. respectnely: howewr, between88%and 93%rnntnbubdmfomumonfor maotherexposurss. netic studies of both bone marrow and peripherai blood samples were required for a test to be acceptable. The most common reasoils for IMswere an inade- quate quantity of material or low cellularity in the material that was submitted. Demographic, clinical, and exposure information abstracted from the medical record for the 270 respondents and the 97nonrespondentswith cytogenetic studies revealed few differences (Appendix Table 1 [first page of National Auxiliary Publications Service (NAPS) document]) except that nonrespondents were less likely to have had a complete remission (x'=6.38, Pe.011, to have been alive at the last follow-up (x'=22.96, P<.Ol), or to have infonnation on education (29%vs 59%). No statistically sigrufkant differencesin the distribution of demographic or clinical variables among M A N , ", and IM patients are present in Ta- ble l with the exception that more W A N patients were male, more IM patients had had a prior diagnosis of myeloplastic syndrome (leukemic infil- trate cZ%), and the median white blood cell count at diagnosiswas higher in NN patients. Odds ratios for selected exposures in M A N vs NN patients are shown in Table 2. Of 19patients who had received cytotoxic therapy for a prior disease, 14 had clonalabnormalities (OR=2.6; 95% confidence interval [CI], 0.9 to 7.22 Point estimates for chemotherapy and radiation were 4.6 and 2.3, respectively. Cigarette smokingand alcohol use at the time of diagnosis were statistically significant. and there was evidence of a dose-response effect for both (P<.OlX When smoking data were arrayed by pack-years, there was the same doseresponse pattern, but the trend was no longer statistically significant( P=.14). Odds ratios for comparisons involving "ever vs never" w e n of alcohol were also statistically s i m c m t . There was no association between abnormalities and chemically exposed hobbies. Odds 836 JAMA August 4,198~-Vol262. N O 5 ratios for occupationalexposures to spe- eific agentsrangedfrom2.6 (95%Cl,1.0 to 6.5) for asbestosto0.6 (95% CI, 0.3 to 1.4)for pesticides. Therewere no significant differencesin exposuresto animal species (cats, cattle, dogs, sheep) h o w n to be susceptibleto leukemicvimses (data not shownr ; Part 1of Table 3 shows the percent- age of MAN and NN patients in gen- eraloccupational categories, m y e d by likelihoodof exposuretochemicals.Just under one fourth of M A N patients were employed at the time ofdiagnosis in pssibly exposed occupations vs 16% ofNN patients (OR=1.7;95%CI, 0.9 to 3.2; D.10). The OR for usual occupa- tions was 1.5 (95% CI, 0.8 to 2.7; P>.lO). Part 2 of Table 3 shows the exposed occupations based on the more restrictive exposure algorithm de- scribed under "Exposure Data" By these more stringent criteria, the pro- portions of exposed in the M A N and NN groups were 8% and 2%, respectively, for occupation at diagnosis (OR=4.6; 95% CI, 0.9 to 18.8;P = .07) and 9% and 496, respectively, for the usual occupation (OR=27; 95% CI, 0.9 to 8.3; P=.13). No evidence of an in- creased prevalence of abnormalities with increased duration of employment was present for any ofthe various occupational classifications. l badjust for the simultaneouseffects of age and sex, and to examine possible interactions, ORs for each exposure variable were computed using a logistic model, first with age and sex as covariates and then with age, sex, and all. second-order interaction terms. Point estimates for cigarette smoking and alcohol use remained unchanged using this procedure. By contrast, the ORs increased for prior cytotoxic therapy (from2.6 to 3.6) and decreased for occu- pation a t diagnosis ( h m 4.5 to 3.61. There was no evidenceof interaction. A second set of logistic models WL. constructed to examine possible confounding effecta of age, sex, and t h r h o u s exposures on each of the above. reported univariate point estimate: ThisP ~ O C ~ SbeSgan witha logisticmocitwith age, sex, and the four expos~n. variables-prior therapy, smoking. ;,. coho1 use, and occupationsatdiagno.. . (classified by the exposum rule). E w exposure variable was removed fm.. the model, and the likelihood for L."., resulting model was computed. On:: one variable, occupation,was not stat:- tically significantby the likelihood m:.test. The finalfive-term model is show in Appendix Table 2 (second page oftt.*. .NAPSdocument). Point estimates associated P values for each of the exp. sure variables in thismodel were as f$SI i I I I I 1 1 i I I 1 1 I 1 ,.f . 1 mental exposures and cytogenetically defined subsets of patients with ANLL but need to be interpreted with some caution in light of possible biases that may have been present. First, it is likely that gost of the NN patients had abnormal karyotypes. If the true proportion of aneuploid pa- tients was 80%or higher, asimplied by recent high-resolutionstudies, the ORs for comparisons between M A N and NN patients can be shown to be art& d y high"because the misclaasification is unidirectional rather than random. Unfortunately, it is impossibleto quan- I tify the magnitude ofthispotentialbias Clod I Sem 10 8 8 12 withixithisstudy. Themagnitudeof this 4 5 4 5 bias is much lower for comparisonsia- nOusemf0 10 23 13 22 volvingthe specific abnormalities. MiscellcvwKuS 8 6 55 Second, the artificially high point es- others la EXW I 12 20 8 6 timates attributable to m i s c h*dcation ofcytogenetics t a t u may have been off- 23 16 30 22 set, to some extent, by the possibility UntrX& wds ram(95% Cl) - A pnondedsm fulOll1 bm Urn- OddSnmO (95%CI) 77 a4 i .r (0.9-3.2) a2 92 98 4.6 (0.418.8) 70 78 1.5 (0.6-2.7) 9 91 27 (O.W.3) 4 98 that smogaterespondents may not al- ways have known about patient$ exposues. The magnitute of thiaeffect waa estimated by cornparingresults based on the 97 patient informantswith those of the 138 surrogate respondents. In general, ORs resultingh m patientin- terviews were equal to or him than those based on surrogate reapondents (Appendix Table 3 [third page of the NAPS doeumentl). The ratios of self- reported to mgate-mporkd ORs were asfollowa: agarettesmoking, 1.1; alcohol use, 1.5; chemical exposm in current occupation, 1.5;chemical expo- lows: prior cytotoxic therapy, 4.4 (95% CI, 1.4 to 13.6; P= .Ol); agarettesmoking, 1.9 (95%CI, 1.1to 3.4; P = .03);and alcohol use, 1.8 (95% CI, 1.1 to 3.4; P= .04z Associations between eight common chromosomalabnormalitiesand fouren- normality was compared using the exposure OR,as distinct from ORs above comparingAA/AN and NN patients. Of the eight abnormalities, both -71-76 andKIshowedanassociationwithcyb toxic therapy for a prior condition, and inv16was strongly associatedwith aga- sure in hobbies, 1.0; and exposure to animala, 1.5. Third, interpretation of results for specificabnormalitiesiscomplicated by the large number of compari_sonsthat were made. In Table 4, the number of statistid tests performed WBB 32.If a vironmentalexposuresareshownin Ta- ble 4. Eleven patients were t(15;17),14 were t(8;21),6 were t(9;22x 22 were plus8, 7 were -6/-5Q, 16 were -7/-7Q, 14 were invl6, and 24 had evidence of karyotypic bbb%ty (KI). rette smoldng with a suggestion of a doseresponse trend (P=.O8!. Plus8 -waspositively aseociatedwith smoking, whereas 7/- 7Q was negatively 8880- ciated, with an exposure OR ofO.2 (95% CI, 0.1 to 0.9). None of the specific ab- Bonfemnicorrectionisused forTable 4, then a P value of ,002(.05c 32)would be required to maintainasignificancelevel d6%.By this rather conservative6 rion, no associations were present. On the other hand, at the 5% level, two A patient was ciassifiedashavingoneof these a b n o d t e s , regardless of whether another WBB alsopresent. There were 18 patients with two or normalitiea waa associatedwith alcohol results in Table 4 (32x.Os) would be use. Employment at diagnosis in occu- expected to be statistidy sisnificaut pations involvingexposureto mutagen- because of random variation alone. In ic chemicals wasassociatedwith t(8;21), actual fact, six (five positive and one more abnormalities. The combinations -that were present were asfollows: plus8 and invl6 (3patients); -6/ 6Q and KI (3 patients); -7/-7q and KI (3 pa- - -tients); plus8 and KI (2 patients); 7/ 7Q and t(9$2) (2 patients): plua8 and 5(15;17) (1 patient); -6/-6Q and - --7/-7Q (1 patient); plus8, -6l-56, and KI(1patient); plus8, 7/ 76,and KI (1patient); snd -6/-6Q, -7/-7Q, and KI (1patientl In Table 4, exposure statu in patients with and without that specific ab- with an exposure odds ratio of 6.4 (95% CI, 1.8to 22.9). COMMENT Several f&ctols, including cytotoxic therapy for a prior disease, cigarptte smoking, and alcohol use, were &ated with clonal abnormalities in these patients. The point eatimate for occupation, although elevated, waa not etatiet i d y SiWcant. These results are consistent with and extend previous reports of BsBociBti011sbetween en-n- negative)s t a t i s t i d y sigmficantassociations were present in Table 4. This amugoge.ents.sth.at some were not chance Interpretation ofTable4 incomplicated by the fact that 18 afthe 93patienta with specific abnormalities had more thaaone abnormality,thun violatingthe assumption that each ofthese cmpari~ n iescompletely independent of the others. Moat of the overlap (61%) w88 &&butable to KI in combbation with oneormoreotherabnormalities(wthe JAM AUCJJS~4 . igag-voi ZEZ. NO.5 - . Table 4.-Patients. . . &tween Spatic Cytogenetic Abnumalies and Selected Exposums in 235 No previous studies have considered cigarette and alcohol use as exposures, Exporun No.fxpom&7otrl(%) OR (95%Cl) P(2-SW) and it is unclear whether these reflect a direct effect of these agents or an indi-. R l O f therBDY Tdal MAN 1(15;17) 1(8:21) @Use -51-5a ..-71-70- - invl6 KI Smokingat diagnosis TOO l 9 M 5 (e) 1U128 (11) i n 1 (9) 1114 (7) --9/79 1I"0, 1 l n (14) Y-.1.6- 1311 1- - I 0114 (0) 6/24 (25) 891234 (38) ... 2.6 (0.9-7.4) 1.1 (0.2-8.9) 0.9 (0.16.6) L..l (U.S 1Y.J) 7.2 (0.3-5.21 1.9 (0.2-16.4) A-.?. 1,-9..-7-.1-0.l.,\ .UNO (. .) 5.1 11.9-13.71 . ... ... .11 .66 .7t .82 .s3 M* ... .Wt ... rect effect ofother exposuresthatmay be correlated with them. Although the strength of the association for both is low (adjusted ORs of 1.8 and 1.9 for cigarette smoking and alcohol use, re- spectively), impressive dose-response lplationships were present. Interest- ingly, only cigarette smokingwas relat- ed to any ofthe specific abnonnalities, these being plus8 (OR=3.2) and invl6 (OR~ 4 . 5 1Cigarette smoke contains a variety of mutagens, has been implicat- WAN 5Bll27 (44) 1.8 (1.1-3.1) .04: ed in a Variety of epithelial tumors,and 1(15:17) 3 l l (18) 0.4 (0.1-1.5) 28 hasbeen associated in at least onestudy l(821) 8/14 (57) 2.3 (0.8-6.7) .22 with ANLL." Alcohol has been associ- 1(9;22) 3n (50) 1.7 (0.3-8.2) .85 atedwith cancer atseveralsites,includ- 14/22 (64) 3.2 (1.3-7.7) .02$ ingtheesophagus and liver,"but not, to -51-5a -71-70 - 3/7 (43) 2118 (13) 1.2 (0.3-5.6) 0.2 (0.14.9) .90 .OS our knowledge, with tumors of the hematopoietic system. These relation- invl8 10114 (71) 4.5 (1.5-13.5) .02$ ships need confirmationin other series. KI l l R 4 (48) 1.4 (0.8-3.3) .54 Inprevious reports, the proportion of Alcohol use eldinosio TOM MIAN 1(15;17) l(821) l(9221 -plus8 51- 5a -?I-7Q hl8 KI Occupaliona1diagnosis TOW ANAN ((15:17) l(821) W;22) plus 8 -51-5a -71-70 h18 KI 15OM2 (85) 92/125 (74) 8/11 (73) 9114 (64) 6a (100) 1922 (68) G' (86) 14/16 (88) 12/14 (eS) 1 5 2 3 (65) (5) 101126 (8) 0111 (0) 3/14 (21) rm (17) 3 ~ (214) w (0) 0116 (0) 1/14 (7) lR4 (4) ... 2.4 (1.4-4 1) 1.5 (0.4-5.7) 1.0 (0.5-2.1) .UNO (. .) 1.2 (0.5-3.0) 3.4 (0.5-25.3) 4.1 (0.9-16.7) 3 3 (0.8-14.6) 1.0 (0.5-2.3) ... 4.5 (0.9-18.8) .UNO (. .) 6.4 (1.8-22.9) 4.0 (0.5-31.3) 3.6 (0.9-14.3) UNO (. , .) .UNO (. .) 1.5 (0.2-12.0) 0.8 (0.16.2) ... .mt .eo .eo ... .90 .a .09 .l6 .86 ... .07 ... .a: .72 .81 ... ... .79 .79 exposed occupations ranged from 22% to 37%,which issimilar to thatfor gen- eral occupational categories in this study (24%). The ORsbased on the gen- eral occupationalcategories were lower (1.7 for occupation a t diagnosis and 1.5 for usual occupation) in the current study than those for the threepositive studies (dculated by us to range from 2.1 to6.4). When amorerestrictive classification of exposurewas used,only 5% of the patients were considered ex- posed, althoughthe ORs were closer to those in previous studies (4.5 a n d 2 7for current and usual occupation, respec---- tively). The issue of where in the pathogene- sis of leukemiaU chromosomal abnormalities occur is unclear.'* The discov- ery of oncogenes near break points in common abnormalities and the rapid progress in understanding the role of 'EOR indmaa tho exposumodds ram computedagainst the resduel that remuned aner sublrac~~hwor~nIhe these genes in various tumors present l ~ : C l , c o n t i d e n c a i n I e ~ l : W A N . a k r a n n c l l k ~ : I ( 1 5 : 1 7 ~ t r a ~ b c a ~ o l E h r o m o1~7:l15(8a2n1d). translocabon of chromosomes 8 and 21: lW?Zl. tnnrlocationd c h m r o m e r 9 and 22: D ~ U S a~a.in of an m r a areasfor future research.=' ' chmmtome: - 5 1 - ~ ~ . d e * b o n o f a u o r p ~ o 1 e h r o m o ~-751;-7a. d e I e t i o n o f s l l o r ~ n o f s ~ s o r n e 7 j We conclude the following: invl8. imrsnlonof chromosome 18; KI, kuVotVplcinslability:and UNO. undefined. t.01<P<.Ool. i.OS<P< .01. 1. Patients with clonal abnormalities differed from patients with normal karyotypes in reporting a higher fre- "ResulWsection, 11thparagfaph). It is sibility that these persons were differ- quency of exposure to cytotoxic agents unclear to what extent, if at all, the ent in ways that would affect the out- and use of both cigarettes and alcohol. presence of one abnormality influences come of the study also cannot be ruled These factors remainedstatistidly sig- the likelihood of a second occurring. As out. nificant after adjustment for age, sex, presented, the data imply that these Despite the potential problems dis- and the effectof other exposures. lesions occur independently of one an- cussed above, we believe certain associ- 2. A nonsignificant but suggestive other, but future research may prove ations may reflect an underlying biolog- association between clonal abnormali- thisassumptionis incomct. ic mechanism. Ofthe variousexposures ties and occupationw 8 p~resent but was Last, this study lacked information that were investigated, cytotoxicthem- infiuenced by the definitionof exposure on one fourth of eligible subjects. Al- py for a prior condition showed a stlong employed and by the confounding ef- though the exposure infomation from ajsociation with abnormal karyotypes fects of demographic variables and oth- the medical record suggests that the (adjusted OR -4.3). This relationship er exposures. -nomspondents were no different than has been reported previously,Dand the 3. Certain specific abnormalities Pwicipahts (Appendix Table 1 [tirst specificitywith 71- 7Q (OR =7.4) and were more likely to occur in patients Page of the NAPS document]),the pos- KI (OR=5.1) is noteworthy. with reported exposure to environmen- EnvironmentalExposureIn Leukemia-Crane e* a. ... .ed a, 8e t :di- UY :he lis for reIse .stat- es, 116 sa attnd 'dY )C1- Id- .to heon- ?S. 1 of 2% onhis enrer 1.5 ent ive .om as3% ex-to !or WX- neor- )V- in id of ?nt .es .ld :.ents l0l. ig- tX, ive :1li- '*as Ire efth- ies ?ts :n- t al tal agenta. In particular, -71-7Q and KI were associated with exposure to @toxic agenta, plus8 and invl6 with c i p t t esmoking, andt(8;21)~ 8as8so- ~tedwithoccupationainvolvingchem- i d exposrtl.es, w h a OCCXI~ tions Were defined UslW relatively ~tsictiveeriteriaC~d~~det- fectmodificationcouldnot be evaluated for any of these specific abnormalities due tosmall numbers. 4. C- s p e c ~ c abnonnalitiei showed no association with any of the expos-, hdudhg -S/-SQ, t(9;22), and t(15;17). 6. Although associated with pres- ence of abnormalities, alcohol use waa not related to any particular ab- normality. 6. Exposure to animaiS and partici- pation inhobbiesinvolvingchemicalex- posurea were not associated with abnormalities.