Post-partum psychosis and its association with bipolar disorder in the UK: a case-control study using polygenic risk scores

Background: For more than 150 years, controversy over the status of postpartum psychosis has hindered research and caused considerable confusion for clinicians and women, with potentially negative consequences. We aimed to explore the hypothesis that genetic liability differs between women with first onset postpartum psychosis and those with bipolar disorder more generally.
Methods: Participants were 203 women with first onset postpartum psychosis (defined as a manic, mixed or psychotic depression episode within 6 weeks of delivery without previous psychiatric history) and 1,225 parous women with history of bipolar disorder. All were assessed using a semi-structured psychiatric interview and psychiatric case note review. 2,809 women from the general population were included as controls. All self-reported their ethnicity as White and were recruited across the UK.
Outcomes: In this case-control study, women with first onset postpartum psychosis (median age at interview: 46; range: 21-79) and women with bipolar disorder (49; 19-83) were ascertained between September 1991 and May 2013. Women with first onset postpartum psychosis had similar bipolar disorder and schizophrenia polygenic risk scores (PRS) to women with bipolar disorder, significantly higher than controls. When compared to controls, women with first onset postpartum psychosis had an adjusted relative risk ratio (adjRR) for bipolar disorder PRS of 1·71 (95% Confidence Intervals: 1·56-1·86, p<0.0001) and for schizophrenia PRS of 1·81 (1·66-1·97, p<0.0001). The effect sizes were similar when comparing women with bipolar disorder to controls (adjRR 1·77 (1·69, 1·84), p<0·0001 for bipolar disorder PRS; adjRR 2·00 (1·92, 2·08), p<0·0001 for schizophrenia PRS). While women with bipolar disorder also had higher major depression PRS than controls (adjRR 1·24 (1·17, 1·31), p<0·0001), women with first onset postpartum psychosis did not differ from controls in their polygenic liability to major depression (adjRR 0·97 (0·82-1·11),p=0·63).
Interpretation: Our study supports the recognition of first onset postpartum psychosis as a separate nosological entity within the bipolar disorder spectrum both in research and clinical settings.