Interpretation
The temporal proximity of the traumatic event appears to increase the risk of the onset of depression and of PTSD at 2 months and thus suggests a response to the acute event (here the PPH).
The prevalence of postpartum depression at 2 months after delivery was 24.4% among the women with PPH, a prevalence higher than those reported by Ricbourg et al. (15% and 23.5% at 1 and 3 months)20, Eckerdal et al. (13.8% at 6 weeks)19, and Thompson et al. (11% and 13% at 2 and 4 months).16 The different thresholds used for the EPDS may explain these differences. A threshold ≥11 is recommended for the French version, while the other studies used higher thresholds: Thompson et al. >12 and Eckerdal et al. ≥12.16,19 The definition for PPH also varies with the study.16,17,19,20,29
The prevalence of PTSD was 12.8% (IES-R≥30) among women with PPH at M2, compared with the 45% at 1 month and 23.5% at 3 months reported by Ricbourg et al., and the 5% at 2 months reported by Thompson et al.16,20 These differences in prevalence may be explained by the choice of questionnaire used. Thompson et al. used the 17-item PTSD Checklist (PCL).16 Ricbourg et al. used the same scale and the same endpoint as we did: IES-R≥ 30.20 Again, the definitions of PPH differed between studies16,20 and the number of individuals was sometimes low.20
At 2 months after delivery, the prevalence of generalised anxiety disorder was 15.9% (GAD-7>7) among women with PPH and 11.7% in the control ground. As no study has used the GAD-7 to evaluate generalised anxiety disorder among women who had a PPH, comparison was difficult. The meta-analysis by Goodman et al. found an overall mean prevalence of GAD in a population of women during the postpartum period was 3.59% (95%CI, 1.85-6.66), lower than our results.6 Our results for anxiety were close to those of Thompson et al., who used a 6-item short form of the Spielberger scale.16 At 2 months, the women in Thompson’s PPH group had a median score of 10 [IQR, 9-11], indicative of low anxiety (median <12); the mean score in our study was 33.0 (95%CI, 31.9-34.1) on Spielberger’s STAI Y-A scale — very low anxiety (score<35).16
As expected, our 2 groups differed for the known risk factors for PPH, that is, mode of delivery (more women with PPH had either an instrumental vaginal or caesarean delivery), geographical origin and smoking, etc. We therefore adjusted the analyses for these factors. We did not observe any statistically significant differences at M2 or M6 for any of the disorders studied (table 1). At M12, the women with PPH had higher adjusted mean scores for PTSD than the control group: 7.6, 95%CI, 6.3-9.1 vs 5.8%, 95%CI, 4.9-6.8, p=0.02. Sentilhes et al.22 found that women with vaginal deliveries only had an unadjusted median score for PTSD less than 5 (95%CI, 0-11) at one year, via the unrevised version of the IES; we used the revised IES.22 We did not find a statistically significant difference in the risk of any one of these psychological disorders at any assessment point between women with and without PPH (Table S8). This lack of association could be linked not only to the exposure studied but also to other potential confounding factors or clinical prognostic factors might modify this risk (wanted or unplanned pregnancy, the women’s investment in the pregnancy, etc). The only others to measure this risk were Eckerdal et al., who found that PPH did not increase the risk of postpartum depression (OR=1.81, 95%CI, 0.91-3.57) (unadjusted/crude).19
Our work focused on 2 current major public health issues. Among maternal deaths up to one year after giving birth, suicide is the second leading cause in France and PPH the fifth leading cause.10 Few publications have considered the psychological impact of immediate PPH, and as discussed in the introduction, all have them have methodological problems.16–20 Most of these studies are limited to postpartum depression.17,19,20