(SLC-019) Relapse risk after antipsychotic discontinuation during pregnancy in women with bipolar and psychotic disorder: A propensity-score matched cohort study

Background: Antipsychotic medications are widely used in treating women with severe mental illness including bipolar disorder and non-affective psychotic disorders. Many women stop these medications around the time of conception, however, there is limited evidence regarding relapse risks after discontinuation.

Objectives: To investigate whether antipsychotic discontinuation before or during pregnancy compared to continuation is associated with an increased risk of relapse.

Methods: We conducted a propensity score-matched cohort study using nationwide registers. We included women with bipolar or psychotic disorders who gave birth to liveborn singletons between 1998 and 2021 in Denmark, and 2007 and 2017 in Sweden, and redeemed at least two antipsychotic prescriptions within one year before pregnancy. Disorders were identified using ICD 9 and ICD 10 codes. Women who discontinued antipsychotics before (N=708) or during pregnancy (N=237) were matched to equal numbers of women who continued treatment. Propensity scores based on maternal and disease severity characteristics were estimated via logistic regression. Outcome measures were hazard ratios (HRs) for severe psychiatric relapse during pregnancy and up to three months postpartum, estimated using stratified Cox regression, with country-specific results pooled via random-effects models.

Results: In our combined sample of women with bipolar or psychotic disorders from Sweden and Denmark, discontinuing antipsychotic treatment was not associated with a significantly increased risk of severe psychiatric relapse compared to continuation of treatment (HR = 0.93, 95% CI: 0.61–1.43, I² = 66.9% for discontinuation before pregnancy; HR = 1.43, 95% CI: 0.68–3.00, I² = 67.9% for discontinuation during pregnancy). However, among women with psychotic disorders (N=529 before pregnancy discontinuers and 234 pregnancy discontinuers), the HRs were higher after discontinuation before pregnancy (HR=1.34,95% CI 0.94-1.90) and during pregnancy (HR=1.60, 95% CI 0.82-3.11) compared to women with bipolar disorders (N=175 for before pregnancy discontinuers and 69 for during pregnancy discontinuers), HR=0.93,95% CI 0.38-1.90 and HR =1.22, 95% CI:0.51-2.95 respectively.

Conclusions: For women with bipolar disorder, we did not find an increased risk of severe psychiatric relapse after antipsychotic discontinuation before or during pregnancy. For women with psychotic disorders, data trended towards an increased relapse risk after antipsychotic discontinuation before and during pregnancy. Although the confidence intervals indicate a non-significant finding, this trend suggests that replication in other samples may reveal differences in relapse risk.