Abstract

When a risk factor influences the intermediary but not progression to the endpoint, it has been shown that the relative risk estimate for the causal intermediate is identical to that for the endpoint under a single pathway framework. When there are multiple pathways, the relative risk estimate for the endpoint is reduced. The authors examine how the reduction of the effect size from a risk factor's association with the causal intermediate to that with the endpoint relates to the proportion of endpoint cases arising through other pathways, and the measure of effect used. For multiple pathways, all measures of effect are reduced and the reduction increases as the proportion of endpoint cases arising through other pathways increases. For single pathways, the relative rate ratio and odds ratios are reduced. In particular, the reduction in the odds ratio may be dramatic because of the commonness of causal intermediates relative to the endpoint. Comparisons of causal intermediate studies with those for the endpoint should consider the influences of multiple pathways, the prevalence of the causal intermediate, the measure of effect used, and the multiple effects a risk factor may have along the pathway when interpreting the differences observed across the causal chain. Am J Epidemiol 2000;151:339–45.