A meta-analysis of alcohol drinking and oral and pharyngeal cancers. Part 2 : results by subsites

Oral and pharyngeal cancers are strongly related to alcohol drinking. We combined findings from all case-control and cohort studies published up to September 2009 and presented analyses by subsites, using a meta-analytic approach. Summary measures were obtained using random-effects models, and taking into account the correlation between estimates from the same study. We also performed a dose-risk analysis, using a random-effects meta-regression model. Compared to non- or occasional drinkers, the overall relative risks (RR) for light drinkers were 1.17 (95% confidence interval, CI, 1.01-1.35) for oral (nine studies) and 1.23 (95% CI, 0.87-1.73) for pharyngeal (five studies) cancer, with no significant heterogeneity between the two sites (p = 0.793). RRs for heavy drinkers were 4.64 (95% CI, 3.78-5.70) for oral (17 studies) and 6.62 (95% CI, 4.72-9.29) for pharyngeal (17 studies) cancer (p of heterogeneity between the two sites = 0.075). The summary RRs for heavy drinkers were 4.11 (95% CI, 2.46-6.87) for tongue (five studies), 7.76 (95% CI, 4.77-12.62) for oropharyngeal (four studies), and 9.03 (95% CI, 4.46-18.27) for hypopharyngeal (four studies) cancer. In conclusion, the alcohol-related RRs are higher for pharyngeal than for oral cancer, particularly at higher doses, while the association with cancer of the tongue was similar to that for oral cancer.