点击显示 收起
1 From the International Agency for Research on Cancer, Lyon, France
2 Reprints not available. Address correspondence to P Boffetta, Gene-Environment Epidemiology Group, International Agency for Research on Cancer, 150 cours Albert-Thomas, 69008 Lyon, France. E-mail: boffetta{at}iarc.fr.
See corresponding article on page 657.
The risk of lung cancer as a result of alcohol consumption has been the subject of epidemiologic investigation for >40 y [the first report was a case-control study from France published in 1962 (1)], but, as is the case with many other important questions addressed by observational research, the evidence has not yet firmly allowed a conclusion on the presence or absence of a causal relation (2). There is little doubt that heavy drinkers experience a greater incidence of lung cancer than do abstainers or light drinkers; in addition, a lower incidence of lung cancer has been observed among moderate drinkers than among abstainers in several populations. A causal interpretation of these observations requires that both bias and confounding be excluded as explanations. Residual confounding by tobacco smoking is one of the main concerns in the interpretation of the increased risk among heavy drinkers, as is misclassification as nondrinkers of drinkers who are at high risk of lung cancer, in light of alcohol’s apparent protective effect among light drinkers.
The report by Freudenheim et al (3) in this issue of the Journal represents the most ambitious attempt to address this issue from an epidemiologic perspective. The analysis is based on studies in 7 cohorts from North America and Europe with a range of drinking patterns (eg, the average consumption in the Alpha-Tocopherol Beta-Carotene Cancer Prevention Study cohort of men from Finland was more than 4 times that in the Iowa Women’s Health Study cohort) and with each cohort contributing >120 cases of lung cancer. This study therefore leads to very precise overall effect estimates and allows an exploration of consistency among populations and sources of heterogeneity. We can still ask: Does it provide the final answers to the questions mentioned above? If not, what else do we need?
Let us consider first whether tobacco smoking explains the greater risk of lung cancer in heavy drinkers than in lighter drinkers or nondrinkers. During the past couple of decades, there has been vigorous debate about whether a residual confounding effect can be ruled out (4, 5). Recent simulations have suggested that even a moderate misclassification of smoking status might explain most of the results reported in the literature (6). The results of the pooled analysis by Freudenheim et al (3) can be interpreted as offering arguments for both sides: on the one hand, the strong relation observed in male nonsmokers would exclude confounding as an explanation; on the other hand, the lack of an effect in smokers—the group in which most of the previous positive results were reported—would suggest that careful control for smoking eliminates a spurious association (the results in female nonsmokers are hampered by the small number of heavy drinkers in this group). One should consider, however, that any effect, if real, is likely to appear only in drinkers of 60 g alcohol/d (6), a group for which Freudenheim et al did not report results.
Because most drinkers in the pooled analysis had a low consumption, the study may be particularly helpful in addressing the apparent protective effect in light drinkers that was reported in previous investigations (7, 8). The results show such an effect in women (and particularly in smoking women) but not in men, at least not in drinkers of >5 g alcohol/d, which corresponds to one drink every 2 or 3 d. Whereas it might be possible to provide a mechanistic interpretation for these findings [eg, by postulating a stronger reduction in C-reactive protein concentrations among female than male drinkers, as suggested in a few studies (9, 10)], I prefer to interpret these results with caution while waiting for replication.
What are the conclusions, then, that we can draw from the study by Freudenheim et al? All questions dealing with the etiology and the mechanisms of human cancers are more complex than they look superficially, and, the more we learn, the more complexity emerges. The large study by Freudenheim et al is very useful in clarifying and sharpening our questions, but it does not answer them completely. Contributions toward a full elucidation of these questions will come from studies encompassing large populations with a high proportion of heavy drinkers and a validation of self-reported alcohol intake, coupled with research on the mechanisms of alcohol carcinogenesis in human and animal models. Each year, alcohol causes an estimated 355 000 deaths worldwide from cancer, of which 17 000 occur in the United States and Canada (11), and yet the resources dedicated to this area of research are scanty.
ACKNOWLEDGMENTS
The author had no personal or financial conflict of interest with the article by Freudenheim et al or the study it concerned.
REFERENCES
Related articles in AJCN: