Investigating risk factors for postpartum psychosis in bipolar disorder

Risk of postpartum psychosis (PP) is high in bipolar disorder (BD), however, factors influencing
this risk are poorly understood. In particular, whether adverse childhood experiences (ACEs),
anxiety disorders and a range of within-pregnancy factors increase risk of PP in BD is yet to be
examined. Of within-pregnancy factors that have been investigated, most have been assessed
retrospectively. Prospective studies examining risk of PP in BD are scarce. The main aims of this
thesis were to a) examine the relationship between ACEs, anxiety disorders and the lifetime
occurrence of PP in BD and b) investigate potential within-pregnancy risk factors for PP in BD
using a prospective follow-up design.
ACEs, anxiety disorders and lifetime occurrence of PP were assessed via semi-structured
interview and self-report questionnaires in 504 parous women with DSM-IV BD-I. Withinpregnancy potential risk factors (including medication, sleep, obstetric and psychosocial factors)
were assessed prospectively in 103 pregnant women with DSM-5 BD (n=84 BD-I/SA-BD; n=19
BD-II/BD-NOS) via semi-structured interview in late pregnancy (baseline); perinatal
psychopathology was assessed via interview at 12-weeks postpartum. All data were
supplemented by clinician questionnaires and/or psychiatric case-notes.
Neither history of ACEs (p-values of 0.06-0.95) nor anxiety disorders (of any type; p=0.47,
phobias; p=0.23, panic; p=0.53) were significantly associated with the lifetime occurrence of PP.
21% women in the prospective sample experienced PP within 6 weeks of delivery. In unadjusted
analyses, an episode of mania/affective psychosis during pregnancy (p=0.001; OR 11.67, 95% CI
2.61-52.21) and loss of at least one complete night of sleep across labour/delivery (p=<0.01; OR
5.35, 95% CI 1.50-19.11) significantly increased risk of PP. After adjusting for known lifetime
correlates of PP, mania/affective psychosis in pregnancy remained a significant predictor of PP
(p=<0.01; OR 16.49, 95% CI 2.76-98.48). Further adjustment revealed prophylactic mood
stabilising medication in the postpartum period had little influence in moderating risk of PP
(p=0.61; OR 0.67 95% CI 0.14-3.16). No significant associations were found between withinpregnancy psychosocial or psychological factors and PP.
This prospective study demonstrated that risk of PP is high in women with BD, despite use of
prophylactic mood stabilising medication. History of ACEs or anxiety disorders did not influence
lifetime risk of PP. Mania/affective psychosis with onset during pregnancy significantly increased
risk of PP, over and above associations with known lifetime risk factors. Sleep loss associated
with labour may act as a final common pathway in the triggering of PP. If replicated, these
findings have clinical implications for managing risk of PP in women with BD. The role of
neurobiological and medication related factors in the pathophysiology of PP requires further
investigation in larger samples.