Search

The purpose of this article was to perform a critical analysis of publications having estimated the utility of X-ray pelvimetry, in order to allow tangible and useful conclusions for the clinical practice. X-ray pelvimetry was proposed in 3 indications: trial of labour among patients with a history of caesarean section, breech presentation, suspicion of cephalopelvic disproportion. The large majority of these publications are retrospective studies, studying a low number of patients and especially without control groups or randomisation. Their contradictory results and their methodological weaknesses do not allow any conclusion. Published randomised trials are exceptional. Among patients with a history of caesarean section, there is only one randomised trial; it demonstrates that ante-partum X-ray pelvimetry is not necessary prior to a trial labour in women with one previous caesarean section. It increases the caesarean section rate and is a poor predictor of the outcome of labour. There is also only one randomised trial which evaluated the interest of X-ray pelvimetry in patients with a breech presentation: the use of pelvimetry in breech presentation at term does not significantly reduce the overall caesarean-section rate, and does not improve the neonatal issues. However, it allows better selection of the delivery route, with a significantly lower emergency Caesarean-section rate. Finally, the only one randomised trial having studied the utility of X-ray pelvimetry for the prediction of cephalopelvic disproportion shows that pelvimetry is a poor predictor of the outcome of labour, has no influence on the neonatal issues and increases the caesarean sections rate. Furthermore, although radiation exposure during a X-ray pelvimetry is very weak, diagnostic X-ray studies during any stage of gestation have been shown to increase the risk of childhood cancer in the irradiated fetus. In the rare cases where pelvimetry is useful (trial of labour with a breech presentation), it is thus careful to perform a MRI pelvimetry.