Objective: The risks of major affective episodes during pregnancy and during the postpartum period have rarely been compared in large samples across diagnoses. The authors hypothesized that perinatal episodes would mainly be depressive, would occur more in the postpartum than the prenatal period, and would be more prevalent with bipolar than unipolar depressive disorders.
Method: The authors pooled clinical information on 2,252 pregnancies of 1,162 women with clinically treated DSM-IV bipolar I disorder (479 pregnancies/283 women), bipolar II disorder (641/338), or recurrent major depressive disorder (1,132/541) to compare rates of affective episode types by diagnosis during pregnancy and the postpartum period and to identify risk factors.
Results: Among women with bipolar disorder, 23% had illness episodes during pregnancy and 52% during the postpartum period. Among women with unipolar depression, 4.6% had illness episodes during pregnancy and 30% during the postpartum period. Based on exposure-adjusted risk per pregnancy, episodes were 3.5 times more prevalent during the postpartum period than during pregnancy, and the risk was consistently higher with bipolar disorder. Depression was the most frequent morbidity during and following pregnancy. In multivariate modeling, factors associated with affective episodes in pregnancy, in descending order, were younger age at onset, previous postpartum episodes, fewer years of illness, bipolar disorder, fewer children, and not being married. Postpartum episodes were associated with younger age at onset, illness during pregnancy, bipolar disorder, fewer children, and more education. Moreover, pregnancy was less likely and perinatal episodes more likely if diagnosis preceded a first pregnancy. First lifetime episodes occurred in the perinatal period in 7.6% of cases.
Conclusions: Among women with major affective disorders, illness risk was much greater during the postpartum period than during pregnancy. Illness mainly involved depression and was strongly associated with younger age at illness onset, bipolar disorder, and high lifetime occurrence rates. The relative risk during pregnancy compared with nonpregnant periods remains uncertain.