ALCHOHOL-CONTAINING MOUTHWASH LINK TO CANCER IS "VERY WEAK," NDMA MAINTAINS

ALCHOHOL-CONTAINING MOUTHWASH LINK TO CANCER IS "VERY WEAK," NDMA MAINTAINS in July 29 comments to FDA. "The preponderance of data suggests no association or a very weak association at best between mouthwash use and oral cancer and no indication of a consistent dose-response effect with mouthwash," the Nonprescription Drug Manufacturers Association contends. In a January letter to industry, FDA requested data on high- alcohol mouthwashes and their possible connection to cancer ("The Pink Sheet" Feb. 3, T&G-13). The agency raised the concern following published reports that individuals using mouthwashes with 25% or higher alcohol content have a slightly increased risk of oral and pharangeal cancers, which appeared to increase with longer and more frequent mouthwash use. "If there were an association between mouthwash use and oral cancer, more consistent results from the various studies would be expected, much like the convincing evidence for an association between tobacco or alcohol and oral cancer," the association suggests. NDMA pointed to what it called "significant inconsistencies" in the outcomes of the four studies referred to by FDA in its letter regarding high-alcohol mouthwashes and cancer. For example, NDMA notes that one of the first studies on mouthwash use and cancer (Wyander et al.) found that the relative risk for developing oral or pharangeal cancer was 1.1 for men (a 10% increase in risk) and 2.8 for women (a nearly three-fold increase in risk). But while the odds ratio for men who were nondrinkers and nonsmokers dropped to .2, NDMA states, the risk to women who did not smoke or drink increased to 3.6 (a near four- fold increase in risk) despite the obvious expectation that risk would diminish. In addition, NDMA conducted a reanalysis of the most recent study on the issue -- a National Cancer Institute study conducted by Winn et al. and reported in a 1991 issue of the Journal of Cancer Research. The original analysis of the data showed that men's risk for developing cancer increased to 1.4 and women's to 1.6 with the use of mouthwash. Maintaining that "mouthwash is not used in the pharynx," the association included data on 601 of the original 866 patients in the study who had oral cancer in the reanalysis and excluded data on those with pharangeal cancer. "There is no significant effect of mouthwash use or any other variables related to mouthwash use in males, [and] the observed effect of mouthwash among females is greatly diminished after adjusting for the interaction between smoking and mouthwash use," NDMA concluded from its reanalysis. "A significant effect of mouthwash is only found in women who smoke." The risk of developing oral cancer was 1.6 for all women compared to 1.1 for women who neither smoked nor drank alcohol. NDMA suggested that "misclassification of exposure -- particularly on the primary confounders, alcohol and tobacco, which are known risk factors for oral cancer -- could account for the results in females." The association's comments include a review of six additional published studies on the link between mouthwash use and the risk of oral cancer. NDMA said the studies possess a "number of design and interpretive limitations, which include: the basic limitations of hospital-based case/control studies; exclusion of controls with tobacco-related diseases; use of surrogates or next-of-kin as sources of exposure information; lack of statistical power to detect an effect if one existed." The association also lists five "major deficiencies" in the studies. They include: "a lack of dose response based on frequency and/or duration...of mouthwash use"; "questionable inclusion of oral cavity and pharynx cases, in view of the fact that the oral cavity is the principal site of mouthwash exposure"; "a lack of consistent findings between studies; a criterion that is critical in attempting to demonstrate a relationship between mouthwash use and oral cavity cancer"; "a lack of biological basis to explain the inconsistent results observed between males and females"; and "failure to correct for residual confounding due to alcohol consumption and tobacco use."