Reasoning is too narrowly focused.

The assertion by Takayama et al. (1) that their study ‘‘strongly supports’’ saccharin’s lack of carcinogenicity in nonhuman primates is incorrect. The study actually presents results from four small groups of rhesus, cynomolgus, African green, and rhesuscynomolgus hybrid monkeys (total, n 4 20 animals), the latter two groups having no controls. Study insensitivity from small numbers of animals (at a minimum, six of the 20 animals would have had to have been positive for the results to be significant) is compounded by the low dose of saccharin that was employed—17.9 mg/kg body weight per day, averaged over 1 week. That level of saccharin approximates what many Americans consumed. For instance, the 1977–78 90th-percentiles of saccharin consumption in adults and 3to 5-yearold children were 10.5 mg/kg body weight per day and 19.7 mg/kg body weight per day, respectively(2). Per capita consumption has increased slightly since then(3). Thus, the study’s results provide no assurance of safety. While Takayama et al. assert their study indicates that saccharin is not carcinogenic, we note that three of the treated monkeys, but none of the controls, had tumors. Only better-designed studies could determine if saccharin can cause tumors in primates. Takayama et al., and the accompanying editorial by Zurlo and Squire(4), present as established fact a theory as to how saccharin might cause bladder tumors in male rats, but not in humans. The theory proposes a concatenation of events triggered by high doses of sodium saccharin that lead to amorphous precipitates that irritate epithelial cells and cause tumors. That mechanism is far from proven. Furthermore, saccharin causes bladder tumors not only in male rats but also in females, albeit less frequently, but the mechanism by which saccharin causes tumors in females has been poorly investigated. The mechanism also does not explain saccharin’s promoter activity. Even if that theory were proven for bladder tumors in male rats, saccharin may cause tumors by more than one mechanism. Moreover, it has caused tumors in other organs and in other species [reviewed in(5)]. In rats, saccharin has caused tumors of the ovaries, uterus, forestomach, skin, and at all sites (other than bladder). In mice, saccharin has caused tumors of the vascular system, lung, uterus, and other sites. The authors fail to report that saccharin causes dominant-lethal mutations in mice, strongly suggesting potential carcinogenicity(6). Takayama et al. state that epidemiologic studies ‘‘failed to show any effect’’ on bladder tumor incidences, and Zurlo and Squire note the ‘‘absence of any conclusive epidemiologic data’’ that saccharin is associated with bladder tumors. Those statements ignore findings from the most sensitive studies. By far the largest study(7), conducted by the National Cancer Institute, found associations between consumption of artificial sweetener and bladder cancer in high-risk males, low-risk females, and males and females combined (heavier consumers of artificial sweeteners). While epidemiology can never provide ‘‘conclusive’’ proof of cause and effect, the best studies are consistent with animal studies demonstrating saccharin’s carcinogenicity. Lastly, we are troubled by the lack of full disclosure of potential conflicts of interest of several authors. Cohen and his colleagues(1) acknowledge funding from the International Life Sciences Institute (ILSI), but readers may not know that the ILSI is funded by food manufacturers, including Cumberland Packing Corp., a marketer of saccharin. Squire has been affiliated with the ILSI and served as a paid consultant to the Calorie Control Council, the trade association of makers and users of artificial sweeteners. It is unfortunate that the Journal published a study having little scientific merit and lacking full disclosure of affiliations of the authors and invited an editorial to be coauthored by someone who has the same potential conflicts as the authors of the study.