We read the review by Holm et al.1 on severe postpartum haemorrhage in relation to mode of delivery with interest. Based on a lower number of red blood cell (RBC) transfusions after primary caesarean section, the authors conclude that primary caesarean section reduces the risk of severe postpartum haemorrhage compared with planned vaginal delivery. The authors state that using RBC transfusion as indicator for blood loss is better than imprecise clinical estimates of blood loss and avoids problems with different definitions of severe postpartum haemorrhage (500 or 1000 ml). Unfortunately, the authors do not provide any information about the criteria used for RBC transfusion in the participating maternity units, nor do they report their average transfusion rates. The clinical decision to transfuse is likely to vary from one doctor to another (and may vary from one unit to another because of different guideline thresholds in each unit), and therefore the rate of RBC transfusion as a precise measurement of postpartum haemorrhage is questionable.
Austin Bradford Hill’s2 criteria form the basis of modern epidemiological research into causation. Hill’s criteria outline the minimal conditions needed to establish such a relationship. In our opinion, the decision for RBC transfusion is dependent not just on whether there is a postpartum haemorrhage, but also on the clinical context, which is variable. For example, the decision for RBC transfusion during emergency caesarean section after intended vaginal birth may be made at a lower threshold than during planned caesarean sections, so invalidating the conclusions of the study. Therefore, the results should in our opinion not be used when counselling women on the intended mode of delivery.