We appreciate the criticisms offered in your recent letter to the editor.1 Certainly, selection by indication bias is an issue in any observational series. We selected cases based on the diagnosis of accreta, whether antenatal or intrapartum. Because of concern for bias in management according to timing of diagnosis, we separated the cases into those suspected antenatally and those diagnosed intrapartum when comparing the potential impact of interventions on outcome.
The authors express concern regarding the consistency of medical and surgical management of patients given the long study period and inclusion of patients from two centres. The same practice group of both maternal-fetal medicine specialists and gynaecologic oncologists practiced at both of the included institutions, the University of Utah Hospital and the Latter Day Saints Hospital, in Salt Lake City, Utah. The majority of attending physicians involved in these cases were (and continue to be) in practice at the same institutions. Both facilities are tertiary care hospitals with Level III Neonatal Intensive Care Units and both are also Level I Trauma Centers; available resources are similar between institutions. While differences in practice patterns over time remains a legitimate concern, the practitioners involved remained consistent.
Regarding ureteral stent placement, there may be some bias as to practitioner preference, but most of our gynaecologic oncologists endorse ureteral stent placement prior to skin incision for any case involving suspected placenta accreta, regardless of other patient characteristics (such as BMI, degree of suspected invasion). The greatest blood loss in these cases occurs during dissection of the bladder from the lower uterine segment and ureteral stents allow for more rapid, continuous localization of the ureters, which may reduce operative time and blood loss while also reducing the likelihood of ureteral injury.
The authors express concern regarding the comparison of those who did and did not undergo prophylactic hypogastric ligation and how such a procedure can be prophylactic in cases of anterior previa. For the purpose of this study, only those women with suspected accreta were included in this comparison (ligation versus no ligation) for obvious reasons of selection bias. ‘Prophylactic’ referred to those ligations performed after delivery of the baby and prior to initiation of hysterectomy as determined by the operative note. The practice in cases of suspected placenta accreta in our institutions is to perform a classical uterine incision beginning above the placental margin (a fundal uterine incision) to avoid placental disruption. The hysterotomy was closed and, if the surgeon elected to do so, prophylactic hypogastric ligation was performed at that time. Performance of the procedure was at the discretion of the attending physician and based, largely, on personal preference as indicated in the manuscript. We did not suggest that hypogastric ligation is never effective, simply that we could not show a benefit in terms of reducing morbidity or the need for large volume transfusion in our cohort of patients.
We agree that our cohort was subject to bias and we acknowledged this limitation in our manuscript. The only way to determine the best way to care for women with placenta accreta is through properly designed clinical trials. Until such data are available, case series such as ours can provide some guidance to clinicians regarding others experience with this morbid and increasingly common condition.