One in three women with a previous diagnosis of bipolar disorder or postpartum psychosis will relapse during the postpartum period, a meta-analysis has found.¹

Thirty seven studies from a number of countries including the UK were included in the analysis.

There were 4,023 patients with bipolar disorder or previous postpartum psychosis and 5,700 deliveries.

The overall postpartum relapse rate was 35% (95% CI: 29-41%). Patients with bipolar disorder had an overall relapse risk of 37% (95% CI: 29-45%) compared with 31% (95% CI: 22-42%) in patients with a history of postpartum psychosis.

Patients with bipolar disorder were significantly less likely to experience a severe relapse (psychosis or mania) than those with a history of isolated postpartum psychosis with relapse rates of 17% (95% CI: 13-21%) and 29% (95% CI: 20-41%) respectively.

Women with bipolar disorder who took prophylactic medication during pregnancy were significantly less likely to experience a relapse than those who did not with relapse rates of 23% (95% CI: 14-37%) and 66% (95% CI: 57-75%) respectively. There was also evidence that women with a history of postpartum psychosis benefited from prophylaxis commenced immediately after childbirth.

Heterogeneity, both clinical (e.g. population characteristics) and methodological (e.g. retrospective² or prospective data collection), is to be expected with meta-analysis and this study is no exception. Therefore, the reported relapse rates should only be taken as a very approximate guide for any given population. A large UK study reported a 20% risk of postpartum mania or psychotic depression for women with bipolar I but a risk of only 2% for women with bipolar II.² Surprisingly, the meta-analysis does not report different risks for bipolar I and II.

Other risk factors for postpartum psychosis include primiparity, recurrent major depression and a family history of bipolar disorder or postpartum psychosis.3 Women at high risk i.e. personal or family (first-degree relative)⁴ history of bipolar disorder or postpartum psychosis should be referred for specialist preconception advice. Valproate should not be given to women of childbearing potential and NICE recommends that lithium should only be used if antipsychotic medication is ineffective because of the risk of fetal heart defects.⁵

In the case of unplanned pregnancy it can be very difficult to decide whether or not to continue mood stabilising medication and urgent specialist advice should be sought.

The authors of the present study recommend that women with bipolar disorder should continue their prophylactic medication during pregnancy. However, the findings of the UK study suggest that this advice should only apply to those with bipolar I.²

NICE recommends that antipsychotic medication can be continued but that we should consider gradually stopping lithium or switching to an antipsychotic.⁵ If the decision is made to continue treatment, a Swedish study provides some reassurance in that, compared with untreated women, the incidence of congenital malformations was only slightly higher among treated women (3.4% vs 2.0%).⁶

The main lesson from this meta-analysis is that women with bipolar disorder have a high risk of relapse during the postpartum period. Approximately half of those who relapse will have severe illness and are likely to need urgent admission to hospital.

We should identify women of childbearing age with a diagnosis of bipolar disorder who are registered with us and ensure that they have been adequately counselled regarding the risks associated with pregnancy and the importance of adequate contraception and preconception counselling.