Document 2Rq9noOLNX954dE7oQdjY8KML

British Journal of Industrial Medicine 1987;44:446-453 Occupation and male lung cancer: a case-control study in northern Sweden L A DAMBER, L G LARSSON From the Centre ofOncology, University ofUmea, 90J 85 Umea, Sweden j I j abstract Using a case-control study comprising about 600 men with lung cancer in northern Sweden the potential risk of different occupations and groups of occupations was studied. Longitudinal data concerning occupation, employment, and smoking habits were obtained by questionnaires. Some occupational groups (underground miners, copper smelter workers, elec tricians, and plumbers) exposed to previously known lung carcinogenic agents such as radon daugh ters, arsenic, and asbestos, had considerably increased odds ratios, which persisted after adjustment for smoking. A slightly raised odds ratio was observed in a group of blue collar workers potentially exposed to lung carcinogenic agents; this rise in the group as a whole mainly disappeared after adjustment for smoking. Fanners and foresters had strikingly low odds ratios, which could only partly be explained by their more moderate smoking habits. The population aetioiogical fraction attributable to occupation was estimated as 9%. ? There is at present little doubt that tobacco smoking is a dominant aetioiogical factor for lung cancer. In several Western countries such as the United States, United Kingdom, and Sweden it has been estimated that smoking accounts for 80-90% (population aetioiogical factor) of lung cancer in men.1 ~3 Several other chemicals and industrial processes, however, increase the risk for lung cancer and may be of considerable importance for limited populations. They include radon daughters in uranium and non uranium mines, asbestos, chromium, arsenic, nickel, mustard gas, bischloromethyl ether, and polycyclic hydrocarbons in soots, tars, and oils.3 There are also some other chemicals suspected to be carcinogenic to the lung but not yet adequately verified, such as beryl lium, vinyl chloride and others. Smoking and other carcinogenic exposures often seem to interact in a synergistic way. This type of interaction has been observed for smoking and asbes tos,4 smoking and arsenic5 and smoking and radon daughters.6"8 Some data also suggest a multiplicative type of interaction between smoking and exposure to polycyclic hydrocarbons.9 10 The background of the present study was the finding that the incidence of lung cancer in men in northern Sweden showed large geographic van- Accepted 26 August 1986 ations.11 Most communities in the inner, western part of the region had low incidence rates compared with the more urban coastal communities and two districts with extensive iron ore mining. It was thought that a large case-control study in this region including occupational and smoking data could yield informa tion of interest about occupational aetioiogical fac tors and interaction between such factors and tobacco smoking. Results from the study have previously been reported concerning underground miners and profes sional drivers.6 10 In another report the observed effects of smoking have been specifically described.2 Material and methods The study region included the three northernmost counties of Sweden, with a total male population of about 390 000. The region contains both urban municipalities with different industrial activities (mines, smelters, steel factories, paper mills, and mechanical workshops) and rural municipalities with forestry and agriculture as dominating industries. The more rural municipalities, however, contain only about 15-20% of the total population. For the present study all male cases of lung cancer reported to the Swedish Cancer Registry during the six year period 1972-7 and dead before the start of the study (May 1979) were selected. The study originally 446 1 ( i iJ ! HWBUI0007496 study j northern studied. ained by :rs, eiecn daughjustment Hentially red after tildjanly # <-* * j , I I . j stern part ared with a districts :ht that a including informaeical faci tobacco usly been d prol'esobserved icribed.2 lernmost lation of h urban activities ills, and ties with ries. The tin onlv gcancer ring the ,rt oLihe | I \ ' I 1 ' Occupation and male lung cancer: a case-control study in northern Sweden 447 contained 604 cases. Copies of the original reports to the cancer registry and of the cytological and histopathological reports were collected and. in question able cases, copies of the hospital records also. Only five cases lacked microscopic verification and in a fur ther five cases the diagnosis of primary lung cancer was doubtful. These cases were nevertheless included in the study. From information in the reports, the cases were classified as small cell carcinoma, squa mous cell carcinoma, adenocarcinoma, and other types (or unclassified). To each case, a control who had died was drawn from the National Registry for Causes of Death, matched by sex, year of death, age, and municipality. Cases of lung cancer and suicides were not accepted as controls. For each case a living control was also drawn from the National Population Registry and matched against the cases by sex, year of birth, and municipality. Only living controls with an age not exceeding 80 (467 controls) were drawn in order not to disturb the very old by the questionnaire procedure. Longitudinal data concerning munic ipalities, type of residence, occupation, employment, and smoking habits were collected from postal ques tionnaires sent to close relatives of the cases and the dead controls and to the living controls themselves. For the living controls, data were registered up to the year the lung cancer was diagnosed in the respective case. Answers were obtained in 589 cases (98%), 582 dead controls (96%), and 453 living controls (97%). The questionnaires were designed to yield informa tion about each occupation or employment, or both, held for at least one year concerning type of industry, company name, task, and duration of employment. A five digit code was used to classify occupations12 and special codes were used for specific companies. One question concerned occupational exposure to asbes tos irrespective of specific occupation or employment. If occupational data were lacking for any period between age 20 and the time of diagnosis (or corre sponding time for the controls) a supplementary tele phone interview was performed. The information concerning smoking habits included approximate year of start of smoking, daily number of cigarettes, other type of smoking, and year of possible cessation of smoking. Pipe smoking was as common as cigarette smoking and a combination of pipe and cigarettes was also frequent, whereas cigar smokers were rare.2 For estimation of life time tobacco consumption and for stratification according to smoking intensity, 1 g of pipe or cigar tobacco was considered equivalent to one cigarette. Incomplete smoking data were always supplemented by tele phone interviews. Individuals who had smoked at least one cigarette or the equivalent amount of tobacco daily for one year or more at any time were classified as smokers. statistical methods The presented results are based on comparison between cases and controls with dissolved matching. All the essential results, however, were also checked with preserved individual matching which gave simi lar estimates. For the calculation of confidence inter vals for the odds ratio (OR), the "exact" method based on the hypergeometnc distribution was used.13 To study the relation between risk occupation and lung cancer, taking into account potential con founding by smoking habits, a linear logistic regression model was used in the analysis.14 The anal yses were performed with three discrete levels of occupational exposure and four levels of life time tobacco consumption. The population aetiological fraction (EF pop) for "risk occupation" was calcu lated according to the formula: EF pop = CFe x (RR-1)/RR, where CFE is case fraction (proportion of cases exposed) and RR relative risk.15 Results It was possible to obtain information about occupation and employment from age 20 for an aver age of 38-9 years in the cases, 38-4 years in the dead controls, and 38T years in the living controls. A per son who had been active for at least one year in a specific occupation was in the analyses assigned to this occupation. One person could therefore be coun ted for more than one occupation. Some occupations were pooled in the analyses to form larger groups. One of these was "white collar workers" (including teachers, clerks, and salesmen). Another group called "mechanics" included employees in mechanical industry, workshops, metal industry, garages, and machine shops, including engineers in such places of work. Plumbers and electricians, who are often exposed to asbestos, were combined to form one group. Several occupations were not separately anal ysed because of the low number of subjects. These occupations included insulators, chimney sweeps, painters, and workers in the chemical industry, groups that could have been of considerable interest. The odds ratio for a specific occupation was esti mated relative to all the remaining subjects. Separate analyses were performed with dead and living con trols. The analyses with dead controls included 589 cases and 582 controls, and the analyses with living controls 456 cases and 453 controls. Table 1 gives the number of cases and dead controls in different occupations with corresponding estimates of the odds ratios; the person-years in the occupations are also shown. Table 2 gives the corresponding data for cases and living controls. HWBUI0007497 448 Table 1 Number of cases and dead controls, odds ratios (OR), andperson-years in different occupations Damber, Larsson Occupation Years Farmers Foresters Carpenters Navvies Administrative workers. including teachers. clerks and salesmen (white collar workers) Plumbers and electricians Enginemen, boilermen. repairmen, welders, stokers, mechanics. turners, engineers (mechanics) Concrete and asphalt workers Professional drivers Pulp workers Underground miners Copper smelter workers Occupational asbestos exposure according to questionnaires 1 2s 20 1 20 1 20 1 20 1 20 1 20 1 20 ! 20 1 20 1 20 3* 1 20 1 20 l 20 No Cases 90 59 94 34 65 26 18 7 98 60 36 24 74 33 28 10 72 37 24 9 25 16 22 15 99 64 Controls 140 105 125 38 80 32 31 11 84 51 21 7 57 24 12 0 58 25 14 7 10 2 8 3 45 22 Off ( unadjusted) 95 V. a 06 (04-0-8) 0-5 (0-3-0-7) 07 (0 5-0-9) 0-8 (0 5-1-4) 0-8 (0 5-1-1) 08 (0 4-14) 0-6 (0-3-10) 06 (0 2-17) 1-2 (0-9-17) 1-2 (0-8-1 -8) 1-7 <1 *0-3*2) 3-5 (1-4-9-6) 1-3 (09-2-0) 1-4 (08-2 5) 2-4 (1-2-5-2) -- 1-3 (0-9-1-9) 1-5 (0-9-2-6) 1-7 (0-9-3-6) 1-3 (0-4-41) 2-5 (1-2-6*0) 8-1 (1-9-73-0) 2-8 (1-2-7-3) 5-1 (1-4-27-4) 2-4 (I-6-3-6) 3-2 (19-5-5) OR (adjusted for smoking) 95% Cl 01 (0 5-1-0) 06 (0-4-1 -0) 0? (0-5-0-9) 0-7 (0-4-1-4) 0-8 (0-5-1-3) 08 (0-4-1-5) 0-5 (0 2-1-0) 0-8 (0-2--2-8) 1-2 (0-8-1 -8) 1-3 (0 8-2 2) 1-5 (0-8-31) 2-6 (10-7-8) 1-0 (0 7-1-5) 1-1 (0 6-2-1) 2-2 (10-5 3) -- 1-0 (0 7-1-5) 1-2 (0-6--2-2) 1-7 (0-7--4-2) 0-9 (0 3-3*3) 2-7 (10-8-1) 9-8 (1-5-414) 2-8 (10-9-6) 9-7 (1-5-413) 2-6 (16-4 3) 3-6 (1-9-7-2) Person years Cases Controls 2759 2397 1797 1222 1163 861 280 225 4537 4219 2192 1405 1463 1072 494 377 2542 21S2 10! 1 899 2004 1694 375 242 1457 1084 460 324 1461 1172 462 310 645 552 540 498 265S 2232 1056 809 109 0 1125 871 285 247 136 77 120 97 974 722 Table 2 Number of cases and living controls, odds ratios (OR). andperson-years in different occupations Occupation Years Farmers Foresters Carpenters Navvies Administrative workers. including teachers. clerks and salesmen (white collar workers) Plumbers and electricians Enginemen, boilermen. repairmen, welders. stokers, mechanics. turners, engineers (mechanics) Concrete and asphalt workers Professional drivers Pulp workers Underground miners Copper smelter workers Occupational asbestos exposure according to questionnaires 20 20 20 20 20 20 20 20 20 20 20 20 20 No Cases 64 40 70 24 50 19 14 5 81 53 32 23 54 21 24 8 63 33 22 i *>*> ?5 20 32 77 46 Controls 134 90 101 41 60 22 19 12 60 42 20 10 48 17 27 7 44 20 15 3 10 4 9 -> 66 40 OR (unadjusted) 95% Cl 0-4 (0-3-0-5) 0-4 (0-2--0*5) 0-6 (04-0-9) 0-5 (0-3-0-9) 0-8 (0-5-1-2) 0-8 (0-4-1-6) 0-7 (0-3-1-5) 0-4 (01-1-3) 14 (10-2-1) 1-3 (0-8-21) 1-6 (0-9-3-1) 2-4 (11-5-6) 1-1 (0 7-1-7) 12 (0-6-2-5) 0-9 (0-5-1-6) 11 (0-4-3-7) 1-5 (10-2-3) 1-7 (09-3*2) 1-5 (0-7-3-1) 2-7 (0-6-15-8) 2-3 (10-5-4) 3 8 (1-2-16-0) 2-3 (1-0-5 7) 61 (1-4-56-4) 1-2 (0-8-1-7) 1-2 (0-7-1-9) OR for smoking) 95% a 0-5 (0-3-0-8) 0-5 (0-3-0-8) 0-7 (0-5-- 1 -1) 0-7 (0-3-13) 0-7 (0-5-I-2) 0-9 (0 4-1-9) 0-9 (0 3-2 4) 07 (0-2--2-9) 1-2 (0 8-1-9) 1-2 (0-7-20) 1-5 (0 7-3-2) 1-8 (07-4-7) 10 <0 6-1*6) 1-5 (0-6--3-7) 0-8 (0 7-2 5) 1-0 (0 3-4 1) 1-0 (0 7-1-6) l-l (0-6-2-2) 1-5 (06-38) 2-3 (0-4-23 2) 1-5 (06-3*8) 2-2 <07-9 2) T-5 (0-6-41) 3-7 (08-34-3) H (0 7-1-7) IS (0 7-2-0) Person- ears Cases Controls 1738 1462 1283 854 899 632 198 152 3955 3515 2017 1453 1016 722 409 355 2172 1908 925 852 1736 1603 431 333 923 654 402 279 1299 1050 422 281 57? 507 473 41! 2016 1680 772 556 354 185 810 607 206 137 165 107 112 63 161! 1383 Occupation and male lung cancer: a case-control study in northern Sweden Table 3 Cell types oflung cancer in different occupations Occupation Squamous cell carcinoma Total No No % Small cell carcinoma No % Adenocarcinoma No % Farmers, foresters, carpenters, navvies White collar workers Mechanics, concrete and asphalt workers, professional drivers, pulp workers Underground miners Copper smelter workers, plumbers, electricians, occupational asbestos exposure according to questionnaires 267 98 198 25 157 122 46 45 46 104 53 n 44 98 62 68 25 25 26 39 20 11 44 26 17 36 13 16 16 2! n 3 12 15 10 449 Other types or unclassified No % 41 16 12 12 34 17 00 18 n On the whole, the odds ratios obtained with dead and living controls were in good agreement. Farmers, foresters, carpenters, and navvies (railroad workers) thus had low odds ratios and underground miners, copper smelter workers, and electricians and plumb ers had high odds ratios. An intermediate group included "white collar workers," "mechanics," pro fessional drivers, and pulp workers who had a ten dency towards slightly, but not significantly, increased odds ratios. Adjustment for smoking (smokers/non-smokers) did not change the odds ratios substantially for most occupations; exceptions were professional drivers and mechanics for which adjustment reduced the odds ratios considerably (tables 1 and 2). A more detailed adjustment with consideration of life time tobacco consumption gave similar results. Table 3 shows the distribution between different cell types of carcinoma; this is commented on further in the discussion. A linear logistic regression model was used to study the possible relation between "risk occupations" and lung cancer with adjustment for smoking. This anal ysis included adjustment for smoking and was per formed for three different groups. (1) Workers with more than five years employment in occupations known or definitely suspected from previous studies to increase the risk of lung cancer. This group included miners, copper smelter workers, plumbers, electricians, insulators, and chimney sweeps. (2) Workers with more than five years employment in occupations suspected of increased risk for lung cancer either due to the character of the work--for instance, exposure to polycyclic aromatic hydro carbons or asbestos--or from epidemiological stud- Years Relative risks estimated by a linear logistic regression model: la) known risk occupations, dead controls: (b) suspected risk occupations, dead controls; (cj known risk occupations, living controls; (d) suspected risk occupations, living controls.---------------------- without adjustment for smoking; with adjustmentfor smoking. 450 ies: enginemen, boilermen, welders, mechanics, stokers, turners, professional drivers, painters, build ing workers, stevedores, printers, workers in the chemical industry (including pulp workers), and con crete and asphalt workers, (3) All remaining subjects. In the analyses group 3 was used as the reference from which the relative risks (RR) were estimated. In group 1 those with more than 25 years in the occupation ran a substantially increased risk of lung cancer. The excess risk persisted after adjustment for smoking (figure). In group 2 also those with more than 25 years in the occupation seemed to run a slightly increased risk; no excess risk was observed, however, after adjustment for smoking. Quantitatively, group 2 was strongly dominated by mechanics and professional drivers. The choice of boundaries for time in an occupation in the figure was not critical and similar results were obtained if the same boundaries were used as in tables 1 and 2. If workers with more than five years employment in any of the occupations in group 1 (figure) were regarded as employed in risk occupations the popu lation aetiological fraction was 9% regardless of whether dead or living controls were used. This figure, which included adjustment for smoking, did not increase if group 2 was also regarded as a risk group. Without adjustment for smoking, however, the esti mated population aetiological fraction became 17% with dead controls and 18% with living controls if both groups 1 and 2 were included as risk groups. Discussion Several types of epidemiological investigations may be used to find or elucidate occupational cancer risks. Studies of occupational mortality or morbidity in registers, achieved by reports of occupation in con nection with cause of death or incidence of cancer reporting16 or by linkage of different registers,17 may give crude indications. This type of study is often based on extensive material but has shortcomings because of uncertainties in the occupational data and a lack of information on smoking habits. Important information about risks of occupational lung cancer derives from cohort studies, often ini tiated by clinical or experimental observations. Cohort studies are mainly useful for the study of specific groups. Examples from northern Sweden are cohort studies on miners and copper smelter work ers.18 19 Individual smoking data are usually lacking in this type of study. Case-control studies have often been used to explore specific factors of suspected aetiological importance. Large case-control studies may also be performed with a more general purpose such as a Damber. Larsson screening procedure in the search for possible risk factors, analysis of interaction between different exposures, and estimation of population aetiological fractions. The latter type of study has been advocated by Doll and Peto.3 A few case-control studies of this type have recently been reported.20'22 The study reported here may be regarded in some ways as a systematic screening for possible occupational risk factors for lung cancer. Looked on from this point of view it was of interest to study how efficiently occupational exposure was registered. One group could be specifically studied from this aspect-- namely, employees at a copper smelter (Ronnskarsverken). The risk of lung cancer among these employees has previously been analysed in a cohort study.18 The original cohort comprised about 4000 male workers, employed for at least three months at this copper smelter at any time between 1928 and 1966. Of the total number of 76 cases of lung cancer found in the cohort, 26 were included in the present study. In 25 of these cases the employment at the smelter was obvious from the questionnaires. Three of these 25 cases, however, were employed for less than one year and therefore were not classified as cop per smelter workers in the present study. The ques tionnaire information thus had high validity in this specific group. In the present study obvious excess risks for lung cancer were found in some occupations previously shown (or highly suspected) to give such risks; these included underground miners, copper smelter work ers, and electricians and plumbers (who are often exposed to asbestos). In underground miners (especially iron ore miners) the risk is in all proba bility caused by exposure to radon daughters in poorly ventilated mines. Data from the present casecontrol study concerning underground miners have been presented in more detail elsewhere.6 This study, and an extended case-control study conducted in two municipalities with large iron ore mines,7 suggested a multiplicative interaction between smoking and underground mining. The copper smelter workers, who were concentrated in only one factory (Ronnskasverken), have previously been analysed in a large cohort study which showed an increased risk of lung cancer associated with working at sites with heavy exposure to arsenic.18 A case-control study within this cohort study suggested a multiplicative effect of this exposure and smoking.5 Electricians and plumb ers in the present study showed similar excess risks and were combined in the analyses to form one group; both types of workers are often exposed to asbestos, which is a probable explanation for the observed increase in risk. According to many studies smoking is especially associated with small cell and squamous cell cur- HWBUI0007500 ursson - risk Terent 3gical cated 'f this some ssible :d on ; how One ect-- itnnsthese >hort 4000 hs at and tncer esent t the ~hree ' less cop- iuesi 'L lung" >usly hese ork)ften ners obasm tave lid> . two ed a and nnsarge ung :a\y thin i of mbisks tup; tos. red 4 ally :ar- Occupation and male lung cancer: a case-control study in northern Sweden 451 cinoma and. to a much lesser degree, ade nocarcinoma. Exposure to radon daughters probably increases the risk of all types of lung cancer but has the strongest association with small cell carcinoma and squamous cell carcinoma; furthermore, small cell carcinoma seems to develop after a shorter latency and at a younger age and is therefore overrepresented in some series of underground miners (cf ref 6). There are some indications that other lung carcinogens such as asbestos and aromatic hydrocarbons are most strongly associated with squamous cell carcinoma.23 In the present study small cell carcinoma was obvi ously overrepresented among underground miners (table 3). There was also a tendency towards over representation of squamous cell carcinoma in the sus pected or proved risk occupations compared with farmers/foresters and white collar workers. Farmers and foresters had significantly low odds ratios (even after adjustment for smoking). A low risk of lung cancer in these rural groups has been reported in several studies24 25 but the reasons for this are not fully known. It is tempting to compare farmers and foresters with the group "white collar workers" (administrative workers, clerks, and teachers). In the latter group no specific occupational risk exposures were suspected; nevertheless, this group had almost twice the relative risk of lung cancer as farmers and foresters. There were definite quantitative and qual itative differences in the smoking habits of the two groups (table 4). There were more non-smokers and pipe smokers among the farmers/foresters than the white collar workers. Pipe smoking in the present population gave about the same relative risk as ciga rette smoking.2 Detailed adjustment for smoking, taking into account different levels of lifetime tobacco consumption, however, only slightly reduced the difference among the odds ratios. It is still possible that even a detailed adjustment for smoking was inadequate and that the smoking habits mainly explained the disparity observed between the two groups. Other more speculative explanations may be environmental tobacco smoke, indoor radon, and general air pollution. As in many reported cohort studies and occu pational mortality (or morbidity) studies slight but statistically not significantly increased odds ratios were found in several other "blue collar jobs" such as mechanics, concrete and asphalt workers, and pulp workers. In the logistic regression analysis, however, no increase in risk persisted in this group as a whole after adjustment for smoking. This indicates that the occupational exposures in this group had minor quantitative importance. A real excess risk cannot be excluded, however, in small specific subgroups--for instance, due to polycyclic aromatic hydrocarbons or other chemical pollutants. Two such groups were pulp workers (commented on below) and asphalt and concrete workers. For the latter group, high odds ratios were obtained only after comparison with dead controls; it was striking that 10 cases had had this exposure for more than 20 years compared with none among the dead controls. No increase of the odds ratios was observed, however, if the ratios were esti mated with living controls. This might have been a "healthy worker effect" explained by the fact that asphalt and concrete work is heavy and therefore overrepresented among living controls, who might have been positively selected concerning health. Asbestos exposure may illustrate some of the prob lems connected with a questionnaire study, when an exposure is not exclusively associated with some specific occupations. Occupational asbestos exposure (according to questionnaires) gave a high odds ratio with use of dead controls but not with use of living controls. A possible explanation might be that the liv ing controls more adequately reported asbestos exposure than the surrogate respondents for cases and dead controls. It seems natural that the exposed workers themselves should have better knowledge about this type of exposure than close relatives. Simi- Tj ble 4 Smoking habits amongfarmerslforesters and white collar workers Pipe and OR after OR after Non- Pipe Cigarette cigarette simple detailed smokers smokers smokers smokers < 20 cigjd 20 rigid adjustment adjustment So % % % % % % for smoking for smoking Farmers/foresters: Cases 184 14 46 18 22 34 52 065 0 71 Dead controls 265 43 31 12 12 33 23 White collar workers: Cases 98 4 21 52 20 31 65 1 22 M3 Dead controls 84 36 13 35 15 31 33 "Smoking non-smoking. +Threc levels of life lime tobacco consumption. i i HWBUI0007501 452 lar experience has previously been reported.26 The odds ratios presented in tables 1 and 2 are esti mated over all ages. If an occupation is hetero geneously distributed over age due, for instance, to changing access to different jobs, the odds ratio is not constant over age groups. In professional drivers this was shown to be the case10 and a more detailed anal ysis of the data suggested an association between the occupation and the risk of lung cancer in the older age groups. In table 5 odds ratios are given for some other occupational groups with stratification according to age (< 70 and J 70). For farmers and foresters the OR for the two age groups were similar. For plumb ers and electricians and for mechanics an increased OR was found in the older age group but not in the younger, whereas in white collar workers and pulp workers the opposite was observed. Deviations obtained after this subgrouping must be evaluated with great caution due to random errors and prob lems of mass significance. Some of these findings might, however, have been caused by a changing risk over time due to environmental changes in the work places. The rather high OR obtained for pulp workers in the younger age group may indicate the need for further studies, especially as an increased prevalence of respiratory disease (chronic bronchitis) has pre viously been described among this type of worker.21 In published reports varying proportions of male lung cancer have been attributed to occupation. Per centages between 5 and 35 have been reported.2 28 29 In the present material the population aetiological Table 5 Odds ratiosfor lung cancer in some occupations after stratification according to age and estimated with dead controls Occupation Age at diagnosis years OR Farmers Foresters White collar workers Plumbers and electncians Mechanics Pulp workers <70 70 <70 70 <70 70 <70 70 <70 70 <70 70 Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls 36 63 54 77 48 73 46 52 65 49 33 35 20 16 16 5 32 35 36 19 !7 6 7 8 05 0-6 0-6 0-9 1-4 0-9 1-2 3-3 0-9 2-0 2-8 0-9 Damber. Larsson fraction attributed to occupation was. with adjust ment for smoking, about 9% regardless of whether a more limited definition (group 1) or a wider definition (groups 1 and 2) for risk occupations was used. With out adjustment for smoking, however, the estimated population aetiological fraction became 17-18% if both groups 1 and 2 were included as risk occupations. Several investigations lack adjustment for smoking which may explain the higher population aetiological fractions reported. Another explanation may be different occupational patterns within the studied populations. This work was supported by the Swedish Cancer Society and by the Swedish Work Environment Fund. We are deeply indebted to Miss Ulla Jonsson and Mrs Monica Johnsson for their skilful help with the questionnaires and Mr Claes Olsson for his valuable help with the telephone interviews. References 1 Higginson J, Muir CS. Environmental carcinogenesis: miscon ceptions and limitations to cancer. JNCI 1979:63:1291-8. 2 Damber LA. Larsson LG. Smoking and lung cancer with special regard to type of smoking and type of cancer. A case-control study in north Sweden. Br J Cancer (in press). 3 Doll R. Peto R. The causes of cancer. Quantitative estimates of avoidable risks of cancer in the United States todav. JNCI 1981;66:1193-308. 4 Seiikoff 1J. Hammond EC, Churg J. Asbestos exposure, smoking and neoplasia. JAMA 1968:204:104-10. 5 Pershagen G. Wall S. Taube A, Linnman L. On the interaction between occupational arsenic exposure and smoking and its relationship to lung cancer. Scand J Work Environ Health 1981;7:302-9. 6 Damber L. Larsson LG. Combined effects of mining and smok ing in the causation of lung carcinoma. Acta Radiol Oncol 1982;21:305-13. 7 Damber L, Larsson LG. Underground mining, smoking, and lung cancer: a case-control study in the iron ore municipalities in northern Sweden. JNCI 1985;74:1207-13. 8 Whitlemore AS. McMillan A. Lung cancer mortality among US uranium miners: a reappraisal. JNCI 1983;71:489-99. 9 Vena JE. Air pollution as a risk factor in lung cancer. .4m J Epidemiol 1982;116:42-56. 10 Damber L. Larsson LG. Professional driving, smoking, and lung cancer: a case referent study. Br J Ind Med 1985:42:246-52. ! I Larsson LG. Sandstrom A. Demographic studies on breast and lung cancer in northern Sweden. Nordic Council for Arctic Medical Research Reports 1978; No 22:15-22. !2 Nordisk yrkesklassificering (Nordic classification ofoccupations). Stockholm: Arbetsmarknadsstyrelsen (National labour market board). 1978. (In Swedish.) 13 Thomas DG. Exact confidence limits for the odds ratio in a 2 x 2 table. Applied Statistics 1971;20:105-10. 14 Dixon W[. BMDP statistical software. Berkely: University of California Press. 1981. 15 Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol 1974;99:325-32. 16 Registrar General. Occupational mortality. London: HMSO, 1978. 17 Wiklund K., Einhorn J. Wennstrom G, Rapaport E. A Swedish Oo IS 19 20 21 23 t4 i HWBUI0007502 mber. Larsson with adjusis of w hether a ider definition as used. With- the estimated ne 17-18% if ided aS risk -'k adjustment ter population r explanation as within the ! i , j edish Cancer Environment Vila Jonsson Iful help with isson (or his lews. ' Occupation and male lung cancer: a case-control study in northern Sweden 453 cancer-environment register available for research. Scanihnavian Journal of Environmental Health 1981;7:64-7. ]S Wall S. Survival and mortality pattern among Swedish smelter workers, hu J Epidemiol 1980:9:73-87. |9 Radford EP. Lung cancer in Swedish iron miners exposed to low' doses of radon daughters. .V Engl J Med 1984:310:1485-94. 20 Lcrchen ML. Occupational causes of lung cancer in New Mexico's Anglos anil Hispanics. Houston: University of Texas. School of Public Health. 1983. (Thesis.) 21 Pastorino O. Berrino F. Gervasio A. Peslenti V. Riboli E. Crosignani P. Proportion of lung cancers due to occupational exposure. Jm J Cancer 1984;33:231-7. 22 Buiatti E. Kriebc! D. Gcddes M. Santucci M. Pucci N. A case control study of lung cancer in Florence. Italy. I. Occupational risk factors. J Epidemiol Community Health. 1985:39:244-50. 23 Vena JE. Byers TE. Cookfair D. Swanson M. Occupation and lung cancer risk. Cancer 1985:56:910-7. 24 Wikiurtd K. Swedish agriculture workers. A group with a decreased risk of cancer. Cancer 1983:51:566-8. 25 Occupational mortality 1970-1973. Oslo: Central Bureau of Statistics of Norway. 1976. 26 Blot WJ. Davies JE. Morris Brown L. et al. Occupation and high risk of lunc cancer in northeast Florida. Cancer 1982:50:364-71. 27 Stjernberg N. Nystrom L. Linden G, Roscnhall L. Mikacisson B. Chronic bronchitis in sulphite pulp factory workers (: a crosssectional study. Br J ind Med (in press). 28 Pike MC. Jing JS. Rosano IP. et al. Occupation: "explanation" of an apparent air pollution related to localized excess of lung cancer in Los Angeles county. In: Whiuemorc AS. Brcslow NE.eds. Energy and health. Philadelphia: Society for industrial and Applied Mathematics. 1979:3-15. 29 Hammond EC. Garnnkel L. General air pollution and cancer in the United States. Prev Med 1980:9:206-11. Jgencsiis: miscon'9:63:1291-8. mcer with special r. A^iic-coniro! Jes of i .JNCl cposure, smoking ti the interaction smoking and its Environ Health lining and smok.7a Radiol Oncol g. smoking, and ire municipalities lalit\ among US :489-9<3 le cancer Am J nokinc. and lung 985:42.246-52. es on breast and ->unat *"r Arctic of occupations/ il labour market s ratio in a 2 x 2 >: University of prevented by a J Epidemiol ondon: HMSO, E. A Swedish Vancouver style All manuscripts submitted to the Br J Ind Med should conform to the uniform requirements for manuscripts submitted to biomedical journals (known as the Vancouver style) The Br J Ind Med. together with many other international biomedical journals, has agreed to accept articles prepared in accordance with the Vancouver style. The style (described in full in Br Med J, 24 February 1979, p 532) is intended to standardise requirements for authors. References should be numbered consecutively in the order in which they are first mentioned in the text by Arabic numerals above the line on each occasion the reference is cited (Manson1 confirmed other reports2-5...). In future refer ences to papers submitted to the Br J Ind Med should include: the names of all authors if there are six or less or. if there are more, the first three followed by et a!; the title of journal articles or book chapters; the titles of journals abbreviated according to the style of Index Medicus: and the first and final page numbers of the article or chapter. Examples of common forms of references are: 1 international Steering Committee of Medical Editors. Uniform requirements for manuscripts submitted to biomedical jour nals. Br Med J 1979.1:532-5. 2 Soter NA. Wasserman SI. Austen KF. Cold urticaria: release into the circulation of histamine and eosino-phil chemotactic factor of anaphvlaxis during cold challenge. N Engl J Med 1976;294:687-90. 3 Weinstein L. Swartz MN. Pathogenic properties of invading micro-organisms. In: Sodeman WA Jr. Sodeman WA. eds. Pathologic physiology: mechanisms of disease. Philadelphia: W B Saunders. 1974:457-72.