food groups were similar for cases and controls. Cases had a signifi- cantly lowerintakcoffish thancontrolsdid(oddsratio=0.5,confidence Interval = 0.2-I .O. p = 0.05). A protective effect for fish consumption in lungcanccrhasnotbwnprev~ouslyrcported.Thcdietary intakeoffoods contaming retmol and 3.carotcnc and the intake of alcohol were not sigmficandy different between cases and controls. For cases, smoking duration was longer and the hmc since cessation for exsmokers was shoncr, cigarette pack years were longer, and the number of cigarettes smoked per day was greater. The factors of occupation, dust exposure, and famdy history of cancer (mcludmg lung cancer) were similarly disulbutcd bctwcen cases and controls.

Study of lung cancer histologic types, occupation, and smoking in Missouri Hoar Zahm S, Brownson RC, Chang JC. Davis JR. Occupaliona[ Sfudw Section. Enwnnmen~al Epprdrmmlogy Branch, National Can- rer /n.wute, Rockulle, MD 2ORY2. Am J Ind Med lY89;15:565-78.

A case-control study of lung cancer was conducted to evaluate the relauonshlp bctwccn lung cancer htslologic types and occupation, adjusted for smokmg. A total of 4,431 whnc malt cases and 11,326 cancer controls, dlagnoscd bclwccn I980 and 1985, were identified through the MIssour Cancer Rcglstry. For all histologic types com- bmcd, cxccss risk walr ohxrvcd among many a prlori suspected hrgh- risk occupauons. Lung cancer wa\ clcvatcd among men employed a~ msulxors (odds ratio [OR] = 6.0: Y5? confidcncc mterval [Cl] = 0.7, 137,X),carpcntcrs(OR= 1.3: 95% Cl = I .O,l.7),pamtcrs,plastcrcrs,and wallpaper hangers (OR = 7.0; 95% Cl = 1.2,3.3), structural metal workers (OR = 1.9; YS% Cl = 0.6,6.0), mechanics and rcparrers (OR = 1.3; 95% Cl = l.O,l.7), motor vehicle drivers (OR = 1.5; 95% CI = 1.2,l.X). pohce and flrcflghtcrs (OR = 1.6; 95% Cl = l.l,2,3),andfood scrwcc pcrsonncl (OR = I .8; 95% CI = I .0,3.S). A deficit of lung cancer was observed among farmers (OR = 0.9; 95% Cl = 0.7.1.0). Adenocx- cmoma of the lung was clcvatcd among carpcntcrs (OR = 1.6; 95% CI = l.O,?.S)andcabinctandfurn~~urcmakcrs(OR=2.O;Y5~Cl=O.4,8.~), which 1h mtcrcstmg bccausc ofthc prcvwuh reports ofcxccs~ adcnocar- cmoma of the nasal cavny assocwcd \rlth wood dust exposure. Aden- ocarcmoma, were alho clwatcd among plumbers (OR = 2.0; 95% CI = 1.0,3.X)andprmtcrs(OR = 1.X: Y5% Cl =0.7,4.2). Elcctrlc~answereat shghdy lncrcascdriskforadenocarcinoma(OR = 1.5:95%C1=0.7,2.8) and ‘other’ or mixed ccl1 t)pc< of lung cancer (OR = 1.5; 95% CI = 0.8.2.9) but atdwreawd risk for small ccl] (OR = 0.8; 95% Cl = 0.3,2.0) and squamous cell (OR = 0.8; 953 CI = 0.4,l.h) tumors. Among welders, adcnocarcinoma (OR = 1.7: YS% CI = 0.7.3.x) and squamous cell (OR = I .7: 95% CI = 0.9,3.3) cancer, wcrc elcvatcd, but small cell and ‘other’ lung cancers hcrc not. Dc\pltc the IimItatIons of the Cancer Rcgl\try data, some mtcrcslmg a\wclations ucrc obscrvcd that merit further study, panlcularl) the a\wc1atIon bstwecn lung adenocar- clnoma and occupatmnal cxposurc to wood And wood dust.

Lung cancer in women and type of dwelling in relation to radon exposure Svensson C, Pershagen G, Klommck J. Dep~tmeni ofCanr:rrEpidemi- #logy, Karolinska Institute. S-104 01 Stockholm. Cancer Res lYXY;49:1861-5.

A case-control study based on mtcrwcws with 210 incident fcmalc lung cancer patrents, 209 age-matched population controls, and I91 hospital controls was carried out in Stockholm county, Swcdcn. Radon measurements made in a sample of 303 dwellings, in which the study subjects had lived, showed that dwellings with ground contact had an avcragc concentration of approximately 160 Bqm 3, twrcc the avcragc conccnuatton of other dwellings. A cumulated radon cxposurc index was calculated for each subject based on dala from the intcrvicws and thcmcasurcmcnts.Forthc totalgroupoflungcanccrarclatlvcr~sk(RR), adjusted for smokmg, age, and degree of urbanwation, of I .X (95% confidence mtcrval: 1.2-2.9) and 1.7 (0.9-3.3) assoclatcd wnh ‘mtcrmc-

diatc’ and ‘high’ exposure to radon was found. There was also a significant trend to a positive dose-response relationship (P(trend) = 0.03). For small ccl] cancer the corresponding figures for RRs wcrc I .9 (0.64.5) and 4.7 (1.5-14.2), respectively (P(trcnd) = 0.01). There seemed to be a positive interaction between randon cxposurc and smoking in relation to lung cancer. The findings indicate that domestic radon may bc of importance for the induction of lung cancer, particularly for some histological types.

Amosite mesothelioma in a cohort of asbestos workers Ribak J, Seidman H, Selikoff IJ. Environmental Research Laboratory, Mount Sinai School ofMedicine, New York, NY. Stand J Work Environ Health 1989;15:10&10.

A cohort of 280 asbestos workers with a short duration of cxposurc to amositc between 1941 and 1945 was followed. These men were alive five years afier starting work and were observed until 1988. Seventeen cases of malignant mesothelioma (eight pleural, nine peritoneal) were found. The mean age at the onset of exposure was 33 years for men with pleural mesothelioma and 30 years of those with peritoneal mesotheli- oma. Chest pain was the main symptom in pleural mesothelioma and abdominal pain in peritoneal mesothelioma. Open lung biopsy was the most useful diagnostic approach for pleural mesothelioma, whereas for pcritoncal mesothclioma it was exploratory laparotomy. Pleural pa- tients died of pulmonary insufficxncy, and peritoneal patients of wastmg and Inanition. In both groups the death certificate diagnosis was less accurate than the climcal diagnosis at death. The mean survival was 12.5 months from first symptom to death for the pleural group and 5.4 months for the perrtoneal group.

Time trends in occupational risks of long cancer among Swedish men from 1961-1979 Carstensen J.M. Pershagen G. Eklund G. Department of Oncology, Uni- versity Ilospital, S-581-85 Linkopmg. Am J Ind Med 1989;15:441-8.

Using data from the Swedish Cancer-Enwronmeni Register, time trends in the standardized Incidence ratio (SIR) for lung cancer 1961- 1979 were studrcd in different occupations. After adjustments for differences in the SIR with respect to year of birth, only the decreasing trends for blacksmiths and members of the armed forces as well as increasing trends for foundry workers and constructron machine opera- tors were significant (p < 0.05). For bakers and pastry cooks, there was a significant interaction between year of birth and observation period in relation to SIR. Data on smoking habits, obtained from a sample of the population, did not stem to explain the trends.

Risk estimates for lung tomours from inhaled 2’pPu0,, 2’8Pu02, and “9Po(N0,), in beagle dogs Dagle GE, Park JF, Gilbert ES, Weller RE. Pacific Northwest Labora- tory, Richland, WA 99352.. Radiat Prot Dosim 1989;26: 173-6.

Lung cancer risks arc bemg studied m beagle dogs given smgle exposures to aerosols of L39P~0,, 238Pu0,, or 239P~(N0,),. A major objective of these studies is to examine the risk of lung cancer relative tothespecificactivityofthcradionuclidc,rateofdoscaccumulationdue to diffcrcnccs in solubilities of the radionuclides, and the presence of competing risk from extrapulmonary lesions. Dose-response relation- slups were studled for the three groups of dogs, with analysts specifi- cally designed to evaluate differences in response. Based on estimated cumulative dose to the lung, risks were found to differ sigmficantly among the radionuclidcs; they were highest for 239P~(N0,), and lowest forz3*Pu0,.Amodc1 in which therisk wasassumcd to beapurcquadratic function of dose fitted the data much better than a pure lmcar model. Currently,allthrccgroupsofdogscanbecomparcdonlytolOyearsaftcr exposure. Howcvcr, it is apparent that the average cumulative dose to the lung may not be an adequate predictor of lung cancer risk for different isotopic and physicochemical forms of plutonium.