Fluoridation of water supplies and cancer mortality. III: A re-examination of mortality in cities in the USA.

Differences in trends in crude mortality rates from cancer between 10 fluoridated and 10 non-fluoridated cities in the USA were claimed by Yiamouyiannis and Burk to indicate that cancer mortality had risen to an unusual extent in the fluoridated cities.' 2 These claims, which were widely publicised, prompted us, among others, to examine all relevant published data for the cities in question. When, however, account was taken of the demographic differences between the two groups of cities and the way they changed, there was no suggestion of a greater rise in cancer mortality rates (standardised by the indirect method) in the fluoridated than in the non-fluoridated cities.3 4 Later Yiamouyiannis and Burk reported that their earlier claim was supported by unpublished data obtained from the vital statistics offices and health departments of the cities concerned when the direct method was used to standardise the mortality rates.5 In this analysis, however, men and women and whites and non-whites were classed together in broad age groups (0-24, 25-44, etc.) and this analysis did not take account of the changes in the sex and ethnic structure of the populations under comparison or adequate account of changes in their age structure. Although these shortcomings were repeatedly pointed out," the claims of Yiamouyiannis and Burk have continued to be published in many parts of the world and extreme statements made about the merits of the direct over the indirect method of standardisation-even when the former is used, as in this case, in a crude and inappropriate manner. Indeed, official inquiries have been instigated in several places including Victoria (Australia) and Minnesota and Kansas (USA).