Cigarette smoking and adult leukemia. A meta-analysis

Increasing evidence suggests that certain forms of adult leukemia may be related to cigarette smoking.

To evaluate the association between cigarette smoking and adult leukemia, we conducted a meta-analysis of available studies. Data were identified through an English-language MEDLINE search for the period 1970 through 1992 and through our knowledge of ongoing and unpublished studies. Among the studies identified, the meta-analysis included seven prospective studies and eight case-control studies. The US Surgeon General's criteria were used to assess the evidence for causality.

A positive association between smoking and certain histologic types of leukemia was found in both prospective and case-control studies. The summary smoking-related risk derived from prospective studies (relative risk, 1.3; 95% confidence interval, 1.3 to 1.4) was greater than that based on case-control data (relative risk, 1.1; 95% confidence interval, 1.0 to 1.2). Prospective data suggested an elevated risk of myeloid leukemia associated with cigarette smoking (relative risk, 1.4; 95% confidence interval, 1.2 to 1.6). Pooled case-control data showed increased smoking-associated risk for acute nonlymphocytic leukemia (relative risk, 1.3; 95% confidence interval, 1.1 to 1.5). Risk of leukemia increased according to the number of cigarettes smoked per day. Population-attributable risk calculations suggested that approximately 14% of all US leukemia cases (including 17% of myeloid and 14% of acute nonlymphocytic leukemias) may be due to cigarette smoking.

The consistency, temporality, and biologic plausibility of this relationship augment our findings, which support a causal relationship between cigarette smoking and certain forms of adult leukemia. Further studies are needed to examine risk among women, dose-response effects, and variation in risk by histologic type.