Genetic analysis of selection bias in a natural experiment: Investigating in-utero famine effects on elevated body mass index in the Dutch Hunger Winter Families Study

Jiayi Zhou,Claire E Indik,Thomas B Kuipers,Chihua Li,Michel G Nivard,Calen P Ryan,Elliot M Tucker-Drob,M Jazmin Taeubert,Shuang Wang,Tian Wang,Dalton Conley,Bastiaan T Heijmans,L H Lumey,Daniel W Belsky
DOI: https://doi.org/10.1101/2023.10.23.23297381
2023-10-24
MedRxiv
Abstract:Natural-experiment designs that compare survivors of in-utero famine exposure to unaffected controls suggest that in-utero undernutrition predisposes to development of obesity. However, birth rates drop dramatically during famines. Selection bias could arise if factors that contribute to obesity also protect fertility and/or fetal survival under famine conditions. We investigated this hypothesis using genetic analysis of a famine-exposed birth cohort. We genotyped participants in the Dutch Hunger Winter Families Study (DHWFS, N=950; 45% male), of whom 51% were exposed to the 1944-1945 Dutch Famine during gestation and 49% were their unexposed same-sex siblings or "time controls" born before or after the famine in the same hospitals. We computed body-mass index (BMI) polygenic indices (PGIs) in DHWFS participants and compared BMI PGIs between famine-exposed and control groups. Participants with higher polygenic risk had higher BMIs (Pearson r=0.42, p<0.001). However, differences between BMI PGIs of famine-exposed participants and controls were small and not statistically different from zero across specifications (Cohen's d=0.10, p>0.092). Our findings did not indicate selection bias, supporting the validity of the natural-experiment design within DHWFS. In summary, our study outlines a novel approach to explore the presence of selection bias in famine and other natural experiment studies.
What problem does this paper attempt to address?