BREAST-CANCER RISK IS POSITIVELY ASSOCIATED WITH HEIGHT

Inter- and intra-national epidemiological studies point to an associat ion between socio-economic status and breast cancer risk. Although the re is no direct evidence, the most favoured reason for this relationsh ip is nutritional. An enhanced dietary status, especially during child hood, would be reflected in adult body build. It is, therefore, surpri sing that there is uncertainty in the literature concerning the associ ation between height and breast cancer risk. In reviewing the publicat ions on this topic it became apparent that case-control studies which found no association between height and risk tended to use self-report ed height. In contrast reports claiming a significant, and positive, c orrelation tended to use heights which were measured by the investigat ors. In a prospective study we found in a cohort of 2731 ostensibly no rmal women that, although there was a highly significant linear correl ation between self-reported and measured height, the shortest women ov erestimated their height whilst the tallest volunteers under-estimated theirs. The significance of crude relative risk and height in this co hort was markedly attenuated when self-reported height was used compar ed to measured height. Such a systematic error could have a profound e ffect on the conclusions of studies in this field which relied on self -reporting and could explain the conflicting reports in the literature .