Fig. 4: Genetic correlation (rg) estimates between eating disorders and psychiatric disorders, and between eating disorders and cardiometabolic diseases using two different familial relationship types.

All within-country (Denmark and Sweden) rg estimates reported per birth year were first meta-analysed using random effect inverse variance weighting, followed by a between-country meta-analysis using the same method. Top row panels rg estimates based on full-sibling comparison, bottom row panels rg based on parent-offspring comparisons. Bonferroni p = 1.28 × 10−3. Confidence intervals are reported as a meta-analysis of Danish and Swedish rg estimates ± 1.96 × standard error.