An individual scoring system for the prediction of postpartum anaemia

Abstract

Background: Postpartum anaemia (PPA) is a common postpartum complication. The goal of this study was to prospectively construct a predictive score for individual risk of PPA. PATIENTS ET METHOD: We prospectively analyzed factors associated with PPA (<10gdL(-1) at 48hours postpartum). Parameters analyzed were demographic data, pregnancy characteristics, delivery and postpartum characteristics. Univariate analysis was performed using Anova or X(2); the Cox model was used for multivariate analysis. The scoring system was validated using ROC curve.

Results: Analysis was performed in 475 patients and validation was carried using an additional 95 patients. Multivariate analysis found four factors independently associated with PPA: anaemia during the third trimester of the pregnancy, Southeast Asian ethnic origin, episiotomy and severe postpartum haemorrhage (PPH) identified by the use of sulprostone. According to the score derived from the Cox model, patients were classified as low (22%, score=0), medium (55%, score=2 or 3) and high (86%, score>3) probability of PPA. Using the AUC of the ROC curve for both the first and the validation cohorts (performed on 95 further patients), we recorded AUCs of 72% and 70% respectively.

Conclusions: This study allowed the derivation and validation of a predictive score of PPA. This score might be useful in targeting prophylactic strategies for PPA. Such strategies could include a more active treatment of iron deficiency (increasing oral iron treatment observance or intravenous iron therapy) especially in exposed population, improvement in the prevention and treatment of postpartum haemorrhage and decreasing the use of episiotomy. Future studies must focus on the external validation and generalisation of this scoring system.