W e read with interest the study by Hassan and colleagues 1 that evaluated a possible association between passive smoking and pancreatic cancer. The authors report that passive smoking is not a significant risk factor for pancreatic cancer, although it may increase the risk in smokers who are also exposed to passive smoking.
Although we congratulate the authors for their original research, we think that lack of data on the body mass index (BMI) among their considered confounding variables may somehow limit their conclusions.
Indeed, although these authors claim that results were adjusted for ''known risk factors,'' no data on body weight are discussed. Although older articles, especially retrospective ones, have reached inconsistent results, obesity is a well recognized risk factor for pancreatic cancer, with a 12% increased risk per 5 kg/m 2 BMI increase shown in a recent meta-analysis. 2 The mechanisms by which obesity increases pancreatic cancer risk are unclear, but a role for the insulin-IGF axis is likely.
Notably, in a cohort study on risk factors for pancreatic cancer, 3 obese smokers had a relative risk of 5.07 compared with 2.39 for nonobese smokers and with 1.87 for obese nonsmokers, suggesting a possible additional effect of BMI, which may be important for passive smoking also.
However, whereas the incidence of cancers mainly related to smoking, such as lung cancer, is overall decreasing in the US, possibly because of a reduction in smoking rates, that of pancreatic cancer is not.
In conclusion, we believe that greater attention should be paid to avoiding possible biases when analyzing the role of potentially preventable risk factors for pancreatic cancer. Therefore, we believe that health policies aimed at reducing the increasingly high prevalence of obesity in western countries may be very important. Recognition of obesity as a risk factor provides further rationale for such policies, and, thus, it should be included in evaluation of risk factors for pancreatic cancer in the future.