Document NEg0aqYevrX54yJwempv9EVmg

ASBESTOS EXPOSURE--SELIKOFF ET AL 109 that these subjects were primarily insulation work Table 4.--Estimated Number of Lung Cancer Deaths Expected to Occur During a Period of 52 Months per 10.000 Men Living at the Start of Period; by Age and by Smoking Habits* ers. While all of them were occupationally ex posed to asbestos dust, their degree of exposure was light as compared with the degree of ex posure of asbestos miners, processors, and weavers Age. Yr (Jan 1. 1963) 35*39 40-44 45-49 50*54 55-59 60-64 65*69 Never Smoked Regularly 0 2 2 7 6 16 14 Pipe, Cigar Only ...t 2 12 15 31 32 Excigarette Smokers) 5 75 10 22 53 71 97 . 1-9 a .D.a.y ... ... 28 .8.0. Current Cigarette Smokers) 10-19 a Day 20-39 a Day ... ... 44 ~ 91 " 157 15 39 -t 69 117 190 305 404a Oay 15 46 90 185 256 350* in earlier times. Bronchogenic Carci noma.--Bronchogenic car cinoma accounted for 24 deaths while only about 2.3 were expected on the 70*74 75-79 80-84 12 28 25 52 100 103 206 288 450* S3 100 32 148 ...185 341 329 'Based upon data from a prospective study with adjustment for US mortality experience, tEllipses indicate rates omitted for categories with no subjects in this study. See Table 2. (Includes cigarette smokers who also smoked pipe or cigar. Men with a history of only cigarette smoking have higher lung cancer rates than snown here. Rates obtained by smoothing the data. basis of general United States mortality data for divided into many five-year age groups, some of the white males. However, as previously mentioned, subgroups contained only a small number of men. evidence at hand suggests that there were propor In consequence, the lung cancer death rate was tionally somewhat more cigarette smokers among statistically unstable in some of the very small sub the 370 subjects than among white males in the groups. In three instances where the observed rate United States as a whole, age being taken into in a small subgroup appeared to be badly out of consideration. This might have partially accounted line, we arbitrarily made an adjustment to bring it for the high bronchogenic carcinoma death rate of more into line with adjacent figures in the table. the subjects. For this and other reasons we made These adjusted figures which are indicated with estimates of the expected number of bronchogenic symbols in Table 4 carry very little weight in the carcinoma deaths, the smoking habits of the men final calculation. All of the rates were then ad being taken into consideration. This was done as justed as follows: follows: Lung cancer death rates in the United States Data are available on lung cancer deaths in re have risen steadily year by year and were higher lation to the smoking habits of 440,000 men en in 1964 than during the period 1960 to 1964 as a rolled by American Cancer Society volunteers in whole. Furthermore, in the study described above, a prospective epidemiological study between Oc we avoided enrolling seriously ill people and, as of tober 1959 and March 1960 and traced through the cut-off date for preparing the computer tape, Sept 30,1964. Causes of death were ascertained from we had not yet received death certificates for all death certificates, but whenever cancer was men of the men now known to have died during the tioned on a death certificate inquiry was made of specified period of time. For these reasons, lung the physician who signed the certificate. In case cancer death rates in the study population were of disagreement between the two sources of infor appreciably lower than those reported for white mation, the physician's statement was accepted. males in the United States in 1964. To compensate For the purpose at hand, we only made use of data for this, we raised the rate of each individual smok covering the 52-month period beginning on June 1, ing category so that the total lung cancer death 1960, and ending on Sept 30, 1964. The number of rate (disregarding smoking habits) in each five- lung cancer deaths occurring during the 52-month year age group would be the same as that of all period was divided by the number of men alive at United States white males (based upon the 1964 the beginning of the period. This was done by five- life table and the 1964 distribution of deaths by year age groups for men in each of the following causes of death). The results of these computations smoking categories; (1) never smoked regularly are shown in Table 4. (It should be noted that (nonsmokers and occasional smokers being com Table 4 shows only such rates as were required for bined); (2) history of regular pipe or cigar smok further calculations.) ing, past or present, but never smoked cigarettes The rates shown in Table 4 were then applied to regularly; (3) excigarette smokers (including those the number of asbestos workers shown in each of who had smoked or currently smoked pipes or the corresponding internal cells of Table 2. This cigars); and (4) current regular cigarette smokers yielded an estimate of the number of lung cancer (including those who also had smoked or currently deaths expected to occur during a 52-month period smoked pipes or cigars). The last of these cate among the 370 asbestos workers classified by their gories was further divided by current number of smoking habits. By "expected" number, we here cigarettes smoked per day: (4a) one to nine ciga mean an estimate of the number of lung cancer rettes a day; (4b) 10 to 19 cigarettes a day; (4c) deaths which would have occurred under the null 20 to 39 cigarettes a day; and (4d) 40 or more hypothesis that asbestos workers do not differ from cigarettes a day. Since the men were divided into other men in respect to their lung cancer death seven groups by smoking habits and further sub- rates, both age and smoking habits being taken JAMA. April 8, 1968 Vol 204. No 2 P . 107 FOS 0009342