Genetic liability to bipolar disorder and body mass index: a bidirectional two-sample mendelian randomization study

Abstract Objectives Bipolar disorder is associated with increased body mass index (BMI), but it remains undetermined if this association is causal and, if so, in which direction it goes. Here, we sought to answer these questions using bidirectional two-sample Mendelian Randomization, a method from genetic epidemiology that uses data from genome-wide association studies (GWAS) to examine whether a risk factor is causal for an outcome. Methods We used summary statistics from GWAS of bipolar disorder and BMI conducted using data collected by the Psychiatric Genomics Consortium and the UK Biobank, respectively. The genetic instrument for bipolar disorder contained 53 SNPs and explained 0.5% of phenotypic variance, while the genetic instrument for BMI contained 517 SNPs and explained 7.1% of phenotypic variance. Results We found that a two-fold increase in the genetic liability to bipolar disorder causes a 0.6 (kg/m 2 ) reduction in BMI, which is statistically significant (P=0.009), and predominantly driven by reduced fat mass. Conversely, we found no evidence that BMI causes changes in the risk of developing bipolar disorder (P=0.59). Conclusion The findings of this study indicate that the increased BMI observed among individuals with bipolar disorder is not a direct consequence of genetic liability to bipolar disorder, but more likely represents the sum of downstream correlates of manifest bipolar disorder, such as side effects of pharmacological treatment, poor diet and sedentary lifestyle, that are all modifiable and could potentially be targeted as part of clinical management.