Using genetic variation to explore the causal effect of maternal pregnancy adiposity on future offspring adiposity: a Mendelian randomization study

Background
It has been suggested that greater maternal adiposity during pregnancy affects lifelong risk of offspring fatness via intra-uterine mechanisms. Our aim was to use Mendelian randomizationto investigate the causal effect of intra-uterine exposure to greater maternal body mass index (BMI) on offspring BMI and fat mass from childhood to early adulthood.

Methods and Findings
We used maternal genetic variants as instrumental variables to test the causal effect of maternal BMI in pregnancy on offspring fatness (BMI and dual-energy X-ray absorptiometry (DXA) determined fat mass index (FMI)) in a Mendelian randomization approach. This was investigated, with repeat measurements, from age 7 to 18 in the Avon Longitudinal Study of Parents and Children (ALSPAC; N = 2,521 to 3,720 for different ages). We then sought to replicate findings with results for BMI at age 6 in Generation R (N = 2,337 for replication sample; N= 6,057 for total pooled sample).

In confounder adjusted multivariable regression in ALSPAC, a 1 SD (equivalent of 3.7 kg/m2) increase in maternal BMI was associated with a 0.25 SD (95% confidence interval (CI) = 0.21, 0.29) increase in offspring BMI at age 7, with similar results at later ages and when FMI was used as the outcome.

A weighted genetic risk score was generated from 32 genetic variants robustly associated with BMI (minimum F-statistic = 45 in ALSPAC). The Mendelian randomization results using this genetic risk score as an instrumental variable in ALSPAC were close to the null at all ages (e.g. 0.04 SD (95% CI -0.21, 0.30) at age 7 and 0.03 SD (95% CI -0.26, 0.32) at age 18 per SD increase in maternal BMI), which was similar when a 97 variant generic risk score was used in ALSPAC.

When findings from age 7 in A 52 LSPAC were meta-analysed with those from age 6 in Generation R, the pooled confounder adjusted multivariable regression association was 0.22 SD (95%CI: 0.19, 0.25) per SD increase in maternal BMI and the pooled Mendelian randomization effect (pooling the 97 variant score results from ALSPAC with the 32 variant score results from Generation R) was 0.05 SD (95%CI: -0.11, 0.21) per SD increase in maternal BMI (p-value for difference between the two results = 0.05). A number of sensitivity analyses exploring violation of the Mendelian randomization results supported our main findings. However, power was limited for some of the sensitivity tests and further studies with relevant data on maternal, offspring (and paternal) genotype are required to obtain more precise (and unbiased) causal estimates.

Conclusions
Our findings provide little evidence to support a strong causal intrauterine effect of incrementally greater maternal BMI resulting in greater offspring adiposity.