December 11, 1995

Are Lean Smokers at Increased Risk of Lung Cancer?The Israel Civil Servant Cancer Study

Author Affiliations

From the Hadassah University Hospital (Drs Kark and Rasooly); Hebrew University—Hadassah School of Public Health and Community Medicine (Dr Kark), Jerusalem, Israel; Computer Center, Bar Ilan University, Ramat Gan, Israel (Ms Yaari); and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel (Dr Goldbourt).

Arch Intern Med. 1995;155(22):2409-2416. doi:10.1001/archinte.1995.00430220065007

Background:  Whether leanness is related to an increased risk of lung cancer is controversial.

Objective:  To examine the association of leanness with lung cancer incidence in a sample of Israeli men.

Methods:  The 23-year lung cancer incidence (1963 through 1986) was determined by linkage to the Israel Cancer Registry in 9975 male civil servants aged 40 through 69 years at initial examination in 1963. In 198 298 person-years of follow-up, 153 cases of lung cancer were identified. In 1963, body mass index (BMI) and cigarette smoking status were determined; in the 1968 reexamination, lung function tests were performed and BMI was reassessed.

Results:  Adjusted for age, smoking, and city by Cox regression, BMI was exponentially inversely related to lung cancer incidence, with a relative risk of 2.3 (95% confidence interval [CI], 1.4 to 3.8) comparing the lowest fifth of BMI (<22.93 kg/m2) with the highest. The association was evident in light, moderate, and heavy smokers. Among smokers, the adjusted relative risk was 3.7 (95% CI, 1.9 to 7.3) for the lowest fifth of BMI. The associations were stronger for men in the lowest 10th of the BMI distribution (<21.38 kg/m2). Controlling for lung function did not materially change the results. The adjusted population-attributable fraction associated with the lowest fifth of BMI among smokers was 20.4% (95% CI, 10.1% to 29.9%). Survival analysis showed that the association of BMI with lung cancer persisted throughout follow-up.

Conclusions:  The association shown between thinness and lung cancer incidence, particularly in smokers, was not attributable to the confounding factors studied, preclinical weight loss, or competing risks. Thinness in smokers may lead to, or may reflect, enhanced host susceptibility.(Arch Intern Med. 1995;155:2409-2416)