Birth order and disease risk across the human phenome

Birth-order effects on disease risk have been studied for individual conditions but have not been systematically assessed at phenome-wide scale in large sibling claims cohorts. We apply two complementary designs, a between-family matched cohort (1.6 million pairs) as a high-powered phenome-wide scan, and a within-family sibling comparison (5.1 million families) as an internally controlled sibling contrast, to 569 diseases in Merative MarketScan claims data. Of 418 diseases with adequate case counts, 150 show Bonferroni-significant associations. First-borns carry excess risk for neurodevelopmental conditions (other/unspecified pervasive-developmental-disorder (PDD) code group odds ratio (OR) = 0.57, autism OR = 0.74, attention-deficit/hyperactivity disorder OR = 0.93) and immune-allergic diseases (food allergy OR = 0.80, allergic rhinitis OR = 0.91); second-borns for substance abuse (OR = 1.19) and gastrointestinal conditions (gastritis/duodenitis OR = 1.14). Across diseases analyzed in both designs, between-family and within-family estimates were positively correlated (r = 0.66; 74.2% directionally concordant); 84.7% of Bonferroni-significant between-family hits agreed in direction. Results are robust to state fixed effects (r > 0.99), full-sibling restriction and stricter clinical rematching (r = 0.93). These findings provide a comprehensive map of birth-order effects in the human disease phenome. In multiple study designs involving millions of matched individuals, either between or within families, birth order determined differential associations with disease risk across 569 diseases.