Consistent with the original report for the SH2B1 locus [10], 5 of the affected subjects were from our cohort of severely obese children, a significant enrichment compared to our general population cohort (P = 1.4×10−3). Extending the analysis to include multiple other population cohorts (Table 2) unambiguously confirmed the association between this deletion and childhood obesity (P = 8.7×10− 7; OR = 38.4, [95% confidence interval = 10.4–120.6]). This finding was further strengthened (P = 3.8×10− 10, OR = 25.0 [9.9–60.6]) by inclusion of previously published data [10], [21].
