Mendelian randomization analyses identify bidirectional causal relationships of obesity with psychiatric disorders

The causal role of adiposity in mental illness: A systematic review and meta-analysis of Mendelian randomization studies

“…Study characteristics and reported associations of the included studies are given in Table 1 , and additional details are presented in Appendix . Twenty-eight studies reported 98 estimates of the association between adiposity and depression 25,32–35,37,40,42–45,47,48,50,53,55–59,62–64,66–70 , three studies reported 9 estimates for anxiety 25,41,67 , three studies reported 8 estimates for anorexia nervosa 61,67,68 , five studies reported 13 estimates for bipolar disorder 31,53,55,65,67 , three studies reported 7 estimates for OCD 61,67,68 , three studies reported 23 estimates for PTSD 49,60,67 , and twelve studies reported 26 estimates for schizophrenia 36,38,39,46,51,52,54,55,61,65,67,68 . One-sample MR design was conducted in six studies 33,34,43,56,57,62 , two-sample MR design was conducted in thirty-two studies 31,32,35–42,44–55,58,60,61,63–69 , both one-sample MR and two-sample MR were conducted in three studies 25,59,70 .…”

“…For the assessment of exposure and outcome, 86% of studies used adiposity data from the UK Biobank and mental illness data from Psychiatric Genomics Consortium. For the selection of genetic variants, most studies selected a genetic variant p-value threshold below 5E-08 (n=32 studies) 25,31,32,35,37–40,43–49,51–61,64–70 , but an F-statistic ≥10 (n=27 studies) 25,32–36,38,44,47–49,51,53,55–60,62–65,67,68,70 , and had a high percentage (≥3%) of explained variance (n=11 studies) 32,35,38,51,55,58,60,62,64,66 , providing good evidence that the chosen instrument was robust.…”

“…One study with 6 estimates on the associations of adipose tissue and bipolar disorder and schizophrenia was not included for meta-analyses because estimates were derived from a single cohort ( Appendix ) 65 . For general adiposity, five studies examined the relationship with bipolar disorder, but their study cohorts overlapped, with the largest study reporting a risk of bipolar disorder of 0.96 (95% CI 0.89-1.03) 67 . Similar overlaps were observed for the association between general adiposity and anorexia nervosa, with the largest study reporting a risk of 1.43 (95% CI 1.30-1.58) 67 , and for associations between general adiposity and OCD with the largest study reporting a risk of 0.68 (95% CI 0.54-0.85) 67 .…”

“…Associations for PTSD were not included in the meta-analysis because this study only reported significant results from European ancestry, results from African ancestry were eligible but not included in the meta-analysis due to overlapped cohorts. For abdominal adiposity, one study reported the associations of abdominal adiposity with anxiety (1.23 [95% CI 0.98-1.56]), anorexia nervosa (1.29 [95% CI 1.15-1.45]), and OCD (0.59 [95% CI 0.44-0.78]) 67 .…”

“…For general adiposity, five studies examined the relationship with bipolar disorder, but their study cohorts overlapped, with the largest study reporting a risk of bipolar disorder of 0.96 (95% CI 0.89-1.03) 67 . Similar overlaps were observed for the association between general adiposity and anorexia nervosa, with the largest study reporting a risk of 1.43 (95% CI 1.30-1.58) 67 , and for associations between general adiposity and OCD with the largest study reporting a risk of 0.68 (95% CI 0.54-0.85) 67 . For peripheral adiposity, two studies with 17 estimates reported the casual relationship between peripheral adiposity and mental illnesses 60,66 .…”