Chunk #44 — 6. The wider picture of replication efforts: consortia, data availability, and field synopses — 6.ii. Anthropometrics and the analysis of “secondary traits”
The meta-analyses of body mass index and height conducted by the Genetic Investigation of ANthropometric Traits (GIANT) consortium raised additional issues. [45, 47, 73] Specifically, unlike the diabetes consortia, where each participating study was designed with diabetes as its primary outcome, the studies involved in GIANT were not originally designed to study determinants of BMI and height, rather they were originally case-control studies of diabetes, prostate and breast cancers, and other diseases. In principle, if the studied trait is associated with disease risk, then conditioning on case-control status can create a spurious association between a marker and the trait. [74, 75] In practice, only a small number of markers will have an inflated Type I error rate—namely, those markers that are associated with disease risk but not directly with the secondary trait—and the magnitude of the inflation depends on both the strength of the association between the secondary trait and disease (which could be modest or controversial, as in the case of smoking and breast or prostate cancer, or quite strong, as in the case of BMI and T2D or
