Primary Outcome
The failure rates were similar in the fibrinogen and placebo groups,i.e. 40.0% versus 42.4%, respectively. Logistic regression analysis showed no significant between-group difference: OR=0.99 (95%CI: [0.66;1.47]; p=0.96). The planned sensitivity analyses yielded similar results.
Secondary Outcomes and Exploratory Analyses
Bleeding was controlled in all cases, with mean total blood losses in the fibrinogen and placebo groups of 1555 and 1723 mL, respectively (p=0.21). There were no maternal deaths (Table 2). The two components of the composite primary endpoint were analysed separately. Transfusion rates (proportion of patients who had more than 2 RBC units) were similar in the fibrinogen and placebo groups (23.4% versus25.0%; OR=1.01; p=0.98). Similarly, the proportion of patients with Hb drops > 4 g/dL on D2 compared with the reference value were identical in the fibrinogen and placebo groups, i.e. 19.1% and 19.5%, respectively (OR=1.02; p=0.95).
All patient characteristics, PPH management actions, transfusion outcomes, and haemostatic rescue procedures are shown in Supplement 2, Table S1 and Table S4. There were no between-group differences in secondary outcomes (Table 2) or recorded adverse events (Table S5) in either the ITT or PP populations.
Total blood loss > 1000 mL and Hb loss > 2 g/dL and plasma fibrinogen level ≤ 4 g/L were identified as significant baseline predictive factors of failure as per the composite primary endpoint (multivariate analysis). Conversely, the centre, a baseline heart rate > 100 beats per minute (bpm) and baseline TXA administration were not. Some post-randomization variables were also significantly associated with an excess risk of failure: fresh frozen plasma transfusion, the need for a rescue procedure and a heart rate > 100 bpm (Table S3). However, adjustment of the effect of IMP supplementation on those variables demonstrated no between-group differences either (not shown). The number of subjects with initial fibrinogen concentrations ≤ 2 g/L was too small (n=5) to assess a treatment effect in this subgroup of patients.
At baseline, the plasma fibrinogen concentration was 4.1 ± 0.9 g/L with no between-group difference. Around 11% of patients had fibrinogen levels ≤ 3 g/L, including 1% whose levels were ≤ 2 g/L (Table 1). Following administration of 3 g fibrinogen concentrate, mean plasma fibrinogen concentrations were 4.2 ± 0.8 g/L at H2, and 5.1 ± 0.9 g/L on D2. By contrast, the fibrinogen concentration decreased in the placebo group at H2 (3.5 ± 0.8 g/L) and remained lower than in the fibrinogen group on D2 (4.6 ± 0.8 g/L) (Figure 2). The mean change between baseline and H2 was +0.03 ± 0.69 g/L in the fibrinogen group and -0.56 ± 0.56 g/L in the placebo group. The MMRM showed a treatment-time interaction (p=0.023), indicating that the overall plasma fibrinogen level time-course was different between the two groups.
Longitudinal analysis of Hb concentration and haematocrit showed significant decreases during the first 48 hours compared with baseline, but no time-course differences between the two groups (not shown).