Author reply

We thank Dr Yavuz and Prof. Karaca for their interest in our article and their comments.¹

This letter gives us the opportunity to clarify the two issues raised. The first concerns the definition of postpartum haemorrhage (PPH). The validity and utility of the various definitions of PPH and major PPH is a crucial issue.² PPH is a multifaceted complication that requires a variety of definitions to capture accurately the spectrum of its pathophysiological evolution. A recent study comparing PPH prevention and management guidelines from eight high-income countries highlighted that only half of these guidelines define severe PPH.² In our study, major PPH was defined as the combination of a PPH diagnosis associated with blood transfusion and/or radiological techniques (transcatheter embolisation of the uterine artery) and/or surgical interventions such as hysterectomy, hypogastric artery ligation or uterine vascular pedicle ligation for postpartum haemorrhage via laparotomy. The clinical diagnosis of PPH was thus an essential criterion for defining the outcome. Undergoing a blood transfusion within 42 days of delivery outside the context of PPH was not sufficient to determine PPH. In our population, 12% of women received a blood transfusion unrelated to PPH. The decision to include only PPH requiring blood transfusion and/or emergency surgery and/or interventional radiology is mainly because the definition of PPH based solely on the volume of blood loss remains highly unreliable in many cases. In addition, as mentioned in the article,³ we previously demonstrated that linkage with the blood transfusion database significantly improves the reliability of the measurement of PPH. Finally, while we fully agree that international guidelines provide no precise criteria for the optimal timing of blood transfusion, French national guidelines specify that transfusion is recommended to maintain a haemoglobin concentration ≥8 g/dl. Consequently, blood transfusion is reserved for the most severe PPH and is not widely used, as evidenced by the results of a large cohort study.⁴

The second question concerns the number of confounding factors introduced into the models. The general rule of at least 10 events per variable is subject to much debate and criticism.⁵ To ensure the robustness and reliability of our results, we first checked that the fit of the model was good for the modelling performed. For example, the Hosmer–Lemeshow test was not significant (P = 0.09) for the main results comparing the association between the mode of conception and major PPH. In addition, we performed several sensitivity analyses that yielded very similar results. To provide a complete answer, we also conducted an additional analysis including only the five main confounding factors, which confirmed the robustness of our findings (Table 1). The results obtained suggest that incorporating more covariates into the initial analyses did not lead to an overestimation of the association between mode of conception and major PPH.

We agree that further examination of placental pathology status in women with PPH would be beneficial, as long as these analyses are systematically conducted using a reproducible method – which is not currently the case.

This research project was conceived by AA, PS and TD. They were involved in study design, analysis and interpretation of the results and wrote and revised the original draft. JC and MA were involved in analysis and interpretation of data, and reviewed the manuscript. CQ reviewed the paper. All authors approved the final article as submitted and agree to be accountable for all aspects of the work.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

PS received funding from the following commercial companies: Merck Serono, Finox Biotech, Ferring, MSD France SAS, Teva Santé SAS, Allergan France, Gedeon Richter France, Effik S.A., Karl Storz Endoscopie France, GE Medical Systems SCS, Laboratoires Genevrier, H.A.C. Pharma, and Ipsen. The authors confirm that none of this funding was used to support the research in this study. The other authors report no conflict of interest.