I have read with interest the paper by Teo et al.1 My first comment is that the authors have failed to include one-step surgery as a potential conservative treatment for placenta percreta.2,3 We have previously reported this technique and consider that it has a number of advantages, and it should be considered an alternative to other techniques such as embolisation. In the case reported by Teo et al., I support their diagnosis and initial approach but consider their conclusion that bilateral occlusion of the uterine arteries was a sufficient protective measure that was flawed. Magnetic resonance imaging can allow the assessment of the likely maternal vessels related to placental invasion.4 This can therefore allow an assessment to be made, before surgery, of the likely efficacy of procedures such as uterine artery embolisation. In the case described, such embolisation did not prevent torrential bleeding on the third postoperative day, which I consider likely to have been due to arterial inflow from subperitoneal blood vessels in the pelvis originating from the internal pudendal artery.5 In this situation, uterine artery embolisation is generally not sufficient as the only haemostatic control method. Teo et al. thought that the vaginal removal of prolapsed placental material 3 days postembolisation would be safe and therefore the torrential haemorrhage this provoked was completely unexpected. I would argue that during the angiography leading up to the embolisation, the blood flow from the lower anastomotic supply originating from the internal pudendal artery should have been investigated by directing the vascular catheter into this vessel. Finding such collaterals would have shown that supplementary surgery was indicated, thereby preventing the subsequent haemorrhage. The idea that all uterine bleeding can always be arrested by embolising the uterine arteries alone is, in my view, mistaken.