Document pmOna7kmpyQ7MD1j5ZwKo7a6w

REVIEWS Scand J Work Eirifrnn Heolrli 1993;I9:369-74 G-- - .-1 Cytogenetics and occupational exposures in acute nonlymphocytic leukemia and myelodysplastic syndrome by Stefania Rodella, MD,' Giovanni Ciccone, Me,*, Giovanna Rege-Cambrin, MD,* Paolo Vineis, MQ2 Working Group on the Epidemiologyof Hematolymphopoietic Malignancies in Italy4 c 4 RODELLA S, CICCONE G, REGE-CAMBRIN G. VINEIS P, WORKING GROUP ON THE EPIDEMIOLOGY OF HEMATOLYMPHOPOIETICMALIGNANCIES IN ITALY. Cytogenetics and occupational exposures in acute nonlymphocytic leukemia and myelodysplastic syndrome. Scand J Work Eiwiron Health 1993;I9:369-74. Studies on the association between cytogenetic aberrations and environmental or occupational exposures of patients with acute nonlymphocytic leukemia and myelodysplastic syndrome were selected from the literature and reviewed. Chromosome abnonnalities as a whole and specific aberrations involving chromosomes 5 , 7, 8, and 21 were considered. Occupational exposures were also considered as a whole or, whenever possible, as specific substances. An association between all occupations and any chromosome aberration has been observed in earlier studies but was not confirmed in later studies. Similar results have been observed when specific aberrations, or specific substances(ie, solvents) have been considered. Methodological issues and the comparability of the reviewed studies are discussed, and suggestions are made for refining epidemiologic investigations. Key rernrs: chromosome aberrations, epidemiologic perspective, hematolymphopoietic malignancies, occupation, review. Results from several studies suggest that some leukemias, particularly acute nonlymphocytic leukemia, may be caused by environmental or occupational exposures. Except for benzene, radiation, and alkylating agents, however, conclusive studies have not been performed ( I ). Leukemias as a whole represent a heterogeneous group of diseases with different prognoses and responses to therapy, as well as different immunologic characteristics (1). Recent improvements in laboratory techniques for cytogenetic analysis have revealed nonrandom chromosome changes in the majority of patients with acute nonlymphocytic leukemia. Moreover, cytogenetic changes that are secondary to cytotoxic therapy for prior medical conditions have been described in almost all cases of this disease (2). These observations have suggested that environmental factors might be related to cytogenetic aberrations i n human leukemias. ' Institute of Pathology, Verona University, Italy. Department of Biomedical Sciences and Human Oncol- ' ogy, Turin University. Italy. Hospital Manageinelit Unit, Molinette Hospital. Local ' Health Unit No 8, Turin, Italy. Paolo Crosignani. Vittorio Demicheli, Ambella Fontana. Giovanna Masala, Lucia Miligi. Paola Pisani, Valerio Ramazzotti. Miranda Ricci. Stefania Rodella. Adele Seniori Costantini. Emanucle Stapnaro. Piero Tozi. Ko- snrio Tumino. Clotilde Vigano'. Paolo Vineis. Reprint requests to: Dr S Rodella. Istittito Anatomia Pato- logica, Policlinic0 Universitario Borgo R o m a . 1-37 I34 Verona, Italy. In myelodysplastic syndrome, representing clonal abnormalities of hematopoietic stem cells that may undergo leukemic proliferation (3). the proportion of patients with an abnormal karyotype can also be high, approaching 60% (4).As in acute nonlymphocytic leukemia. therapy-induced myelodysplastic syndromes have characteristic cytogenetic features; a possible association with exposure to environmental cytotoxic agents has been suggested for this category of disorders as well (3). From an epidemiologic point of view, classifying patients with leukemia or other hematolymphopoietic malignancies by the presence of specific chromosome abnormalities may strengthen the possibility of identifying risk factors in etiologic research. Recent studies have investigated the association between some occupational exposures and cytogenetic aberrations in patients with acute leukemia, particularly acute nonlymphocytic leukemia, and other malignancies of the hematolymphopoietic system, with different outcomes. The purpose of this paper is to review such studies and discuss their results, especially focusing on methodological aspects and comparability issues and suggesting some priorities for future research. Methods All of'the studies on the association between cytogenetic aberrations and environmental or occupational exposures among patients with acute leukemias or other hematolymphopoietic neoplasms were selected l'roni thc published literature. The literature search 369 Scrim/ J Work-E ~ i v i r mHrw/r/t 1993. vol 19. n o 6 was based on data banks (Medline) and personal- clusion was based on poor quality of clinical records Table 1. Desigl files. The chromosome aberrations were indicated in three studies (6-8) and on refusal or other im- nonlymphocyti with AA (clonal abnormalities i n all analyzed cells) pediments to interview in two studies ( 1 1, 14). The and AN (clonal abnormalities in a portion of the proportion of cases excluded was rather important in metaphases). Among the specific cytogenetic aber- some of the studies (6, 8, 9). In one study (7) 22% rations, we singled out those analyzed in all (or most) of the patients were excluded because the chromostudies, namely, complete or partial loss of chromo- some examination had failed or there was insufficient some 5 (-5/5q-), monosomy 7 or deletion of its long information about occupation, but the number of pa- Miielman ANLL et al. 1978 (sl arm (-7/7q-), trisomy 8 and 21. and translocation tients i n each category was not specified'in the pa- t(8;21). This choice was made to compare the fre- per. Finally, the proportion of patients excluded be- quency of specific abnormalities in differentStudies cause of a missing cytogenetic test (44 of 367, 12%) and their association with exposure. As the loss of was reported in only one study ( 1 0). sequences belonging to chromosomes 5 and 7 is the The type of information collected was rather poorhallmark of treatment-induced hematological malig- ly defined in four studies (6-9) ("previous or nancies ( 5 ) , these chromosome abnormalities were present occupation"). In two studies ( I I , 14) the pa- Mitelman ANLL I el al, 1981 (7) analyzed separately to disclose a possible preferen- tients were interviewed about their lifelong occupa- tial involvement of these rearrangements also i n en- vironmental-induced acute leukemias. Eight studies were selected, published between 1978 and 1990. Seven of them (6-1 I ) included cases of acute nonlymphocytic leukemia, and only one concerned cases of non-Hodgkin's lymphoma ( 12). A paper published i n 1979 ( 13) was not com- tional history. Finally, i n one study (IO). the occupational history of the patients was classified according to three categories (usual, current, exposure to specific agents). Allegedly, i n all of the studies data on occupation and exposure were collected and categorized by one or more experts who had no knowledge of the cy- I t Golomb et al. 1982 (8) \ ANLL I 1 r r I t r pletely comparable with the other studies because it togenetic status of the subjects. t also included patients with nonclonal chromosome The number of exposed and unexposed patients Y aberrations. Therefore, this study was not reviewed. with presence or absence, respectively, of abnormal Mitelman ANLL I In one of the seven studies regarding cases of acute karyotype was reported or computed from the origi- et al, I nonlyniphocytic leukemia, cases of myelodysplastic nal data, whenever possible. syndrome were also considered ( I I ). In another study The results of each study were expressed using the myelodysplastic syndrome, chro!iic myeloid leuke- odds ratio as a measure of the association between mia, chronic lymphatic leukemia. and acute lym- chromosome abnormalities and occupational expo- phocytic leukemia were considerd in addition to sures considered as a whole and, whenever possible, acute nonlyniphocytic leukemia ( 1'4). This review mainly refers to acute nonlymphocytic leukemia and myelodysplastic syndrome. The only study regarding non-Hodgkin's lymphoma has been considered separately in the text. as specific substances. The odds ratio was reported in the original paper in a few studies; when not available it was computed by us from the published raw data. Confidence intervals for the odds ratios were also computed by us, according to Miettinen (IS), Crane 01 a1 1989 (10) ANLL I C a Y ri it All of the studies included patients with a diag- when not present in the original paper. nosis of acute nonlymphocytic leukemia or myelo- dysplastic syndrome and an available cytogenetic analysis. Cases with a previous diagnosis of myelo- Results N a r d 6 AHLL'C Dube MDS n i! proliferative disorder. cases receiving cytotoxic The most relevant aspects of each study were syn- 198gr11) C c therapy before cytogenetic analysis. and cases of thesized, and the results are given in two tables, to a secondary leukemia were always excluded or ana- make possible a comparative interpretation of study 5 lyzed separately. Therefore, it is reasonable to as- designs (table I ) and the results (table 2). The fol- C n sume that all of the cases considered in these stud- lowing comments are presented according to the layies were incident (ie. newly diagnosed) cases, even out of tables I and 2. s a 5 though this definition was clearly used in only four Each study had eligible cases that were not in- papers. cluded in the subsequent analyses either because a Vineis et a1 ANLL lr MDS c The recruitment of cases was prospective in three cytogenetic test was not available (see footnotes to lW 'l" +; studies (9. IO. 14). In four a large proportion of the cases was identified retrospectively (6-8. I I ). All of the reviewed studies can be considered analyses of case scries. but in one the term "case-control study" was used ( I I ), because the comparisori group (acute nonlyniphocytic leukemia patients without chromosome aberrations J was selected to resemble the case group as nearly as possibie. In each study sonic cases w r e excluded because of' insufficicnr inlbriiiation on occupation. The ex- table I) or insufficient data on occupational history were collected. The hypothesis of an association between a missing cytogenetic test and probability of exposure, although rather unlikely, cannot be definitely excludcd. This might be a potential source of bias. The exclusion of all of the prevalent cases ahd of previously treated patients is very important when one focuses on environmental exposures. i n order to exclude all secondary chromosome rearrangements Unexposed categ workers and othE and others, refere . ment among exp The 162 patients cShIXrohmunodsroemdeaenxda a Four hundred ant and one because and 9 nonrespon e out of 270 respo out Qf 97 nontes : FiYeen cases an " Cases with ChrOr v.!th MDS were I 370 Scarid J Work Environ Health 1993, vol 19, no 6 AI records other im14). The portant in (7) 22% chromosufficient x r of pan the paluded be67, 12%) her poorvious or i)the pa!occupahe occud accordmsure to .cupation .d by one f the cy- patients tbnormal he origi- using the between la1 expopossible, reported lot avail,bed raw ios were ien (15), ere synables, to of study The folI the lay- - not in- :cause a notes to I history Ition bebility of be defiource of s and of nt when order to gements Table 1. Design of studies on cytogenetics and occupationalexposures in hematolymphopoieticma1ignancies.a (ANLL= acute nonlymphocytic leukemia, MDS = myelodysplastic syndrome, CT =chemotherapy, RT = radiotherapy) Path- Selection Recruitment ology criteria Eligible Analyzed Excluded subjects subjects subjects (NI (N) N(%) Sources ~ y p eof information Occupational exposures Non. occupa- lional exoosures Exposure attribution Mitelman et ab, 1975 (61 ANLL Incident Retrospective cases?, 1 July 7972- no Drevious 31 Oct 1977 myeloproliferative disorders, no previous treatment, age >15 years 93 56 Insufficient Clinical data on records occupation 37 (39%) Previous or present occupation Solvenls. Not Blindly insecticides. investigated (1 person) petroleum products Mltelman et al, 1981 (7) ANLL Incident Retrospective? cases?, 1972-1980 no previous 1974-1980 myeloproliferative disorders, no previous treatmenl, age >15 years 208?b 162 Insufficient dinical Previous or Solvents, NO^ Blindly data on records present insecticides, investigated (1person) occupation ( t interview occupation petroleum ?? of cases and products relatives) Golomb et 81, 1982 (81 ANLL Incident Retrospective cases?, Feb 1970- no previous ADrii 1981 myelopro- liferalive disorders, no previous treatment, age >17 years 128 74 Insufficient Clinical Previous or Insecticides, Hobbies Blindly + +dala on records presenl solvents t (in not (3 persons) occupation ( interview occupation chemicals exposed) 54 (42.2%1 of relatives) metals, petroleum products Mitelman ANLL Incident et al, cases?, 1984 (9) age >16 years Prospective 1 Jan 198031 March 1982 595c 361 Insufficient Personal Previous or Insecticides, Not Blindly data on interview present chemicals, investigated (? number occupation (standardized occupation radiation. of persons) 234 (39.3%) questionnaire) petroleum products, metals/ minerals, biological agents, undefined exposure Crane et al 1989 (10) ANLL Incident cases, age 218 years. resident in Texas Prospective 1 Jan 197630 June 1983 3236 2359 (270 re spondents1 Refusals Personal usual or not located: interview (N = 891, occupation. current 80 (27%) telephone occupation. (97 nonre interview specific spondents)' (N =57), exposure postal questionnaire (N= 124) (standardized questionnaire) Asbestos, benzene, peslicides, dyes, glues, varnishes, metals, paints, radialion CT or RT, smoking, alcohol, hobbies A priori malrix Narod h DUM, 1989 (111 ANLL Cases: Case-referent, MDS no previous retrospective CT or RT chromo. some aberration 5 or 7; controls: no chromo. some aberration 5 or 7 Vineis et al, 1990 (14) ANLL Inci&nt MDS cases, aoe > I 5 years Prospective Oct 19WJune 1989 44 ca5es.g 44 cases, lnsufficienl Telephone 38 re. 38 re. data on interview ferents lerents occupation of subjects ?? and their relatives Occupalional Metals, history pesticides. radiation. chemicals, petroleum products 22 22 - 6n 6 Persona Occupational Organic interview history solvents (standardized auestionnairel Not Blindly investigated (3 spe. cialists in occupational medicine) Smoking Blindly (1 person). 3 exposure categories a Unexposed categories 01 the sludies reference 6 housewtves. students and while collar workers reference 7 housewives students whqte collar woruers and others reference 8 housewfves students white collar worker5 and others, reference 9 housewwes students white collar workers and others. reference 10 housewives. managers whlte collar workers and others reference 11 (hose exposea < 1year in 30 last years or disagree ment among experts, reference 14 - 1 The 162 patients analyzed represent 78% 01 the total nLmber of patients identified by the abtnors 22% was excluded because of laiiure 01 the cnromosome examination or insufficient data on occupation (the number in each category 01 exclusion was not reported) Six hundred and sixty patients were initially identified but 65 oflhem were excluded because of their age I C 16 years1 11 Four hundred and forty patients were initlaiiy identified O L I 70 were excluded because they were loreign two because of lheir age I< 18 years) and one because 01 a previous diagnosis Moreover n e have not included i n the niimber of eligible cases (N = 3671the 44 patients (35respondents and 9 nonrespondents) with a cylogenetc test insbfficient or not performed L Out of 270 respondents 31 nad an insufl clent lest and lour did not have a test perlorrred % Out of 97 nonrespondents seven ha0 an Insulflcient lest and lwo dto not have a test performed 11 Fifteen cases and one referelt with secondary IeLremia are not consioered 'I Cases with chronic myeloid leukem a chronic lyinDnat8c leuuemia and acute Ikmphatic leukemia are not considered qn the revieu. I t pallenls with MDS were initially tdentifieo bLt '<e01 lhem d 0 1 0 1 nave a test perforrneo 37 I I j ! Scand J Work Environ Health 1993, vol 19. no 6 Table 2. Results of s t u d i e s on c y t o g e n e t i c s and o c c u p a t i o n a l exposures in h e m a t o l y m p h o p o i e t i c malignancies (OR = od( ratio, 95% CI = 95 O h confidence interval, AA = clonal a b n o r m a l i t i e s in all analyzed cells, AN =clonal a b n o r m a l i t i e s in a portic of t h e metaphases) Authors Mitelman et ai, 1978 (6) Miteiman et al, 1981 (7) Golomb et al, 1982 (8) Mitelman et al. 1984 (9) Crane et al 1989 (10) Narod B Dubd. 1989 (11) Vineis et al. 1990 (14) Patients with chfomosome aberrations1 eligible patients ~~$~~~~ analyzed patients Type of exposure 27/56 (48.2 %a) 741162 (45.7 nq 37174 (50.0%'I 165/299 (55.2 O h c ) 178/323 (55.1 %) 44182 (53.6 %) 10128 (35.7 %) 23156 (41.1 O h ) All 521162 (32.1 O h ) Solvents Insecticides Petroleum products All 16/74 (21.6 O h ) 103299 (34.0 %) Solvents Insecticides Petroleum products All Schoelvmenictasls++ metabs Insecticides Petroleum products All 621235 (26.4 O h ) 461235 (19.6 %) Chemicals Insecticides Petroleum products Minerals All (usual) All (current) 17 cases144 (38.6 %) 14 relerentsl38 (36.8010) 9/28 (32.1 %) Benzene Pesticides h:etals Asbestos Dyes. glues Paints Radiation All Chemicals Pesticides Petroleum products Metals Radiation Solvents Type of chromosome aberration AN + AA -5/5q-. -7/7q. + 8,A+N2+1 . t(8;21) AA AN+AA AN + AA AN + AA -515q. -717g +8,A+N2+1,AtA(8;21) AN+AA AN + A A AN + AA +- 515q.. -7/7q+ 8, 21, t(621) Exposed, chromosome aberration Yes No 19 4 04 10 13 3 --4 34 39 13 17 13 19 13 -24 3 11 4 10 12 4 84 14 AN+AA AN c A A AN + AA AN + AA -5/5q.717q+8. +21. t(8211 AN +AA AN + AA AN AN + + AA AA AN + AA AN+AA -515q 717q. ++8, +21 t(8,21) AN AN AN + + AA AA AA AN+AA AN + AA AN + AA AN +AA Abnormality 5 or 7 Abnormality 5 or 7 Abnormalily 5 or 7 Abnormaiity 5 or 7 Abnormality 5 or 7 Abnormality 5 or 7 AN+AA 5 1 -2 62 69 34 14 24 15 34 24 15 20 11 17 7 81 38 24 2-9 6(?) 6 10 18 17 16 16 6 17 11 2 17 2W) %(?I 5 15 12 7 17 6 5 14 10 1 45 63 02 36 Unexposed, chromosome abarration Yes No 8 25 1 25 1 25 8 25 8 25 8 25 35 75 13 75 8 75 35 75 35 75 35 75 25 33 7 33 7 33 25 33 25 33 25 33 96 1Do 22 100 20 100 96 100 96 100 96 100 96 100 88 84 97 91 97 91 97 91 ?? ?? ?? ?? ?? ?? ?? 27 24 27 24 27 24 27 24 27 24 27 24 7 12 OR 95% CI 14.80 56.25 62.50 (81.251D I18.75lb 2.34 6.43 7.54 13.70 17.14 (47.141b 0.86 3.96 9.43 1.18 4.29-51.34 9.15-345.6 10.35-377.3, 11.62-568.01.63-216.1' 0.43-12.67 3.16-13.08 3.15-18.06 5.49-34.17 6.04-48.59 7.90-281.2~ 0.25-2.94 1.19-13.18 2.45-36.26 0.11-12.52 3.30 [2.64Ib 3.96 2.11 1.87 2.21 1.67 1.89 2.53 8.33 1.w 1.70a 0.16a 1.1oa 0.60d 1.504 2.60a 0.806 2.604 1.106 1.18' 0.91' 1.78' 0.79' 2.12' -i 1.oo' 0.63-17.37 0.09-74.59 0.80-19 70 1.30-3.46 0.87-4.04 1.02-4.75 0.83-3.36 0.87-4.13 1.03-6.23 141-49.38 0.84-2 730 0.82-3.106 0.30-3.W 0.30-1.4lY 0.70- 3.m 1.00-6.W 0.40-l.@Oe 0.390-36.87@@ 0.40-3.31e 0.29-2.84'' 0.15-21 .loc 0.17-3.59 0.42-10.6b" -e 0.20-4 7 0 ~ This percentage was computed only for analyzed cases because the number of chromosome abnormalities in patients excluded from the analysi: was not mentioned by the authors (5ee the text ) :I Since the value in one of the cells IS 0.the OR is equal to inlinity. The value between brackets has been computed using correction to 0 5 Only palients 30 years and older are considered (separately analyzed i n the original article). ': OR reported in the original article. * 95% CI reported in Ihe original article. ' Age-adjusted OR reported in the Original article or abnormalities induced by antineoplastic therapies. This criterion of selection was correctly adopted by all of the authors. In the earlier studies with a retrospective recruitment of patients (6-8) niost of the inforniation about occ11pa t ionit I ex posu res \vas drawn from cI i n ical records with occasional integration from other sourcc\ o f data (ie. telephonc interview of patients or their next-0I'-ki t i ) , Moreover t he type oI i II 1.0 rmat i on collected on occupation was not clearly described. and the exptession "previous o r picsent occupation" w a s 371 used. On the contrary, a better quality of inforination would be expected from prospective studies (9. I O , 14) in which standardized questionnaires were used for personal interviews and a lifelong occupation history was collected. The percentage of patients showing chroniosoqie ahnormalities over the total of eligible patients was similar among the studies (table 2) and ranged. i n niost of them. between 45.7 and 55.2%. I n the study of Vineis et a1 (14) a proportion of 35.7% was oh~ r v e dh.ut the small nunihef of cases miphi explain .. .. . 1 Scand J Work Ensirwi Health 1993. vol 19. no 6 ic malignancies. (OR =odds iai abnormalities in a portion 1, le - OR 1 95% CI 14.80 1.25 62.50 tei.25p [18.75]b 2.34 6.43 7.54 13.70 17.14 (47.141b 0.86 3.96 9.48 1.18 4.29-51.34 9.15-345.61 10.35-377.39 11.82--568.04 1.63-216.19 0.43-12.67 3.16-13.08 3.15- 18.08 5.49-34.17 6.04-48.59 7.90-281.26 0.25-2.94 1.19-13.18 -2.45-36.26 0.11 12.52 3.30 [2.641b 3.96 2.1 1 1.87 2.21 1.67 1.89 2.53 8.33 1.506 1.706 0.16d 1.104 0.606 1.506 2.606 0.806 2.606 1.106 i.iw 0.91` 1.78` 0.79` 2.12` -1 1.00' 0.63-17.37 0.09-74.59 0.80-19.70 1.30-3.46 0.87-4.04 1.02-4.75 0.83-3.36 0.87-4.13 1.03-6.23 1.41-49.38 0.84-2.736 0.82-3.106 0.30-3.804 0.30-1.400 0.70-3.20. 1.00-6.m 0.40- 1.Bee 0.90-6.7W 0.30-3.60. 0.40-3.31` 0.29-2.64` 0.15-21.1P 0.17-3.591 0.42-lo.@ -e 0.20-4.7Oe ents excluded from t h e analysis x l e d using correction to 0.5. er quality of informaprospective studies (9, d questionnaires were .ind a lifelong occupa- showing chromosome t` eligible patients was ihle 2 ) and ranged, i n ild 55.2%. In the study ioii of 35.7% was obot'citscs might explain this difference. In three studies (6-8) the results of the cytogenetic test were not reported for patients excluded from the analyses because of poor information on occupation. The blind procedure for exposure assessment should have avoided a selection bias with regard to the chromosome findings. The reported prevalences of aberrations closely approached data reported by other authors. In fact, in the eartier literature, approximately half of the patients with acute nonlymphocytic leukemia were found by different authors to have clonal aburrations and, more recently, at least two-thirds of these patients have bwn estimated to be affected by recognizable clonal chromosome abnormalities at diagnosis (16). The proportion of exposed patients showed some variations among the studies, ranging between 4 1.1% (6) and 19.6% (10) for all exposures as a whole. These discrepancies could reflect either a different prevalence of exposure among the populations under study or the application of different criteria in the exposure assessment. A decreasing prevalence of exposure among patients with chromosome aberrations and, conversely, an increasing prevalence among patients without chromosome aberrations were observed when earlier (ie, 6, 7 , 9 ) and later (ie, 11, 14) studies were compared. Selection bias in the earlier studies cannot be ruled out. The odds ratios observed in the earlier studies indicated an association, even though of different strength, between all of the occupations considered and any chromosome aberration, but these observations were not confirmed in the later studies (table 2). A similar difference between earlier and later studies was observed when specific aberrations involving chromosomes 5 and 7 or trisomy 8 and 21 were analyzed (table 2). The differences observed among the studies were even more impressive since some of them were conducted with very similar methods. However, the very large confidence intervals in the earlier studies motivate special caution when conclusions are drawn from these results. Similar caution must be applied when the association between solvent exposure and any chromosome abnormality is considered. Insecticides or pesticides were considered in six studies (6-1 I). The odds ratios were not significant (9-ll), or the numbers observed were too small (6-8). Exposure to petroleum products was analyzed in five studies (6-9. l l ). and the odds ratios ranged between 0.79 and 3.96. Other exposures (rnetals, minerals, dyes. paints, radiation) or some specific chromosome aberrations were unalyzed o n l y in a few studies, and it was not possible IO compare the results. Compareb with the several studies o n acute leukemias, only one study, to our knowledge. has been performed on non-Hodgkin's lymphoma ( 12). The paper discusses the relation betwecn chro111osome aberrations in 54 incident cases o f non-Hoclfkin's lymphoma and occupational exposure to organic solvents, assigning patients to three categories according to the number of cytogenetic events producing clonal aberrations. (This number was used as an indicator of karyotypic complexity.) The results suggest that chromosome aberrations, particularly a 14q+ abnormality, are more frequent in patients with occupational exposure to organic solvents. This finding seems to be similar to findings concerning acute nonlymphocytic leukemia. The number of patients in this study was very small, and more formal investigations are needed to confirm these results. Discussion The majority of the studies published so far, to our knowledge, concerning the association between chromosome aberrations and occupational exposures in hematolymphopoietic malignancies are limited to acute nonlymphocytic leukemia. There are probably different reasons for this choice. Its incidence has been increasing over time i n the United States in the last few decades ( I ) , especially among men. Gender and age-related differences in temporal trends suggest that some of the increased incidence may be real and possibly attributed to occupational or environmental exposures ( I ) . Many epidemiologic studies have suggested the role of environmental factors like benzene or radiation in the etiology of acute nonlymphocytic leukemia. Chromosome rearrangements are recognizable at diagnosis in at least two-thirds of the patients with acute nonlymphocytic leukemia (16). The results of our review deserve some important comments. Refinements in cytogenetic techniques over time are not sufficient to explain the discrepancies observed among the studies. First, the overall prevalence of chromosome abnormalities in acute nonlymphocytic leukemia is similar in all of the reports, and these data make the results largely comparable. Second, the specific abnormalities which have been analyzed, namely, -5/5q- -7/7q-, trisomy 8 and 21, and t(8;21) are rather gross rearrangements which can be detected also in poorly banded chromosome preparations. Conversely, methodological issues in study design, especially concerning the retrospective recruitment of patients, poor definitions of occupational exposures. the exclusion of` large proportions of eligible patients, and a lack of analysis of confounding factors. like age, gender and other exposures. are likely to influence strongly the results of the earlier studics. In addition, a chance cffcct cannot he excluded cornpleteI y I Finally, ii clear causative role in leukenlogenesis has not been established for broadly defined exposures like "sdvcnts" or "petroleum products." The awociation bciweeii chromo~onicaberrations and , chemical exposures should follow :I clear detiionstra- 373 &and J Work Environ Health 1993, vol 19, no 6 tion of a causative role of such exposures in the etiology of acute nonlymphocytic leukemia. In conclusion, further epidemiologic investigations are needed, that fulfill standard epidemiologic requirements of analytical studies such as the prospective enrollment of patients, control of selection and information biases, collection of accurate information on both exposures and confounders, and adequate statistical analysis. In addition, quality control of cytogenetic tests should be performed, and the source of interobserver variability should be studied. Internal comparison of case series for cytogenetic damage should be focused on known risk facto? for leukemias. In addition, little is known about the etiology of myelodysplastic syndrome, and an investigation on exposure to environmental cytotoxic agents was recommended in a recent report of a cooperative study group (3). Including this group of diseases in future research on the relationship between cytogenetics and chemical exposures in hematolymphopoietic malignancies would be of great interest. Acknowledgments This work was supported in part by the Italian Association for Research on Cancer (AIRC) and the American National Cancer Institute (grant R01-CA 5 1086-01A2). References 1. Sandler DP, Collman GW. Cytogenetic and environmental factors in the etiology of the acute leukemias in adults. Am J Epidemiol 1987;126:1017-32. 2. Rowley JD. Chromosomesand cancer.New York, NY: Academic Press, 1983:1-2. 3. Third MIC Cooperative Study Group. Recommendations for a morphologic, immunologicand cytogenetic (MIC) working classification of the primary and therapy-related myelodysplastic disorders. Cancer Genet Cytogenet 1988;32:1-10. 4. Jacobs RH, Cornbleet MA, Vardiman JW, Larson RA, Le Beau MM, Rowley JD. Prognostic implications of morphology and karyotype in primary myelodysplas- tic syndromes. Blood 1986;67:1765-72. 5. Pedersen-Bjergard J, Philip P, Larsen SO, Jensen G, Byrsting K. Chromosome aberrations and prognostic factors in therapy-related myelodysplasia and acute non-lymphocyticleukemia. Blood 1990;76:1083-91. 6. Mitelman F, Brandt L, Nilsson PG. Relation among occupationalexposure to potential mutageniclcarcino- . . .genic agents, clinical findings and bone marrow chro- mosomes in Acute Nonlvmohocvtic Leukemia. Blood 1978;52(6):1229-37. 7. Miteiman P. Nilsson PG. Brandt L.Alimena G. Gast- aldi R,Dallapiccola B. Chromosome pattern,-occupa- tion and clinical features in patients with acute non- lymphocytic leukemia. Cancer Genet Cytogenet 198I ; 4:197-2 14. 8. Golomb HM, Alimena C.Rowley ID. Vardiman JW, Testa JR, Sovik C. Correlation of occupation and karyotype in adults with acute nonlymphocytic leuke- mia. Blood 1982;60(2):404-1 I . 9. Mitelman et al. The correlation of karyotype and oc- cupational exposun to potential mutageniclcarcinogen- ic agents in acute nonlymphocytic leukemia. Cancer Genet Cytogenet 1984; 11:326-31. Crane MM, Keating MJ, Trujillo JM, Labarthe DR, Frankowski RF. Environmental exposures in cytoge- netically defined subsets of acute nonlymphocytic leukemia. JAMA 1989;262:634-9. 11. Narod SA, Dub15ID. Occupationalhistory and involve- ment of chromosomes 5 and 7 in acute nonlymphocytic leukemia. Cancer Genet Cytogenet 1989;38:261-9. 12. Brandt L, Kristofferson U,Olsson H, Mitelman F. Re- J lation between occupational exposure to organic solvents and chromosome aberrations in non-Hodgkin lymphoma.Eur J Hematol 1989;42:298-302. 13. Lawler SD. Summersaill BM. Clink MDH.McElwain TJ. Chromosomes, i&kemia and occupational expo- sure to leukemogenic a ents Lancet 1979;2:8534. 14. Vineis P, Avanzi GC, 8iovinazzo B, Ponzio G, Rege- J Cambrin G,Ciccone G. Cytogeneticsand occupational exposure to solvents: a pilot study on leukemias and myelodysplastic disorders. Tumori 1990;76:350-2. 15. Miettinen OS. Estimability and estimation in case-ref- erent studies. Am J Epidemiol 1976;103:226-35. 16. Heim S. Mitelman F. Cancer cytogenetics. New York, NY: Alan R Liss. l987:65-7 1. Received for publication: 1 I August 1992 374