Cystoscopy at the Time of Benign Hysterectomy: A Decision Analysis

摘要

Urinary tract injury (bladder or ureter) occurs in up to 4.3% of all hysterectomies. The risk of morbidity is significantly decreased when a urinary tract injury is detected intraoperatively. Failure to detect a bladder or ureteral injury intraopera-tively may result in peritonitis, urinoma, or fistula formation, with significant morbidity and need for subsequent treatment. Cystoscopy may be used to detect intraoperative urinary tract injury, but how often it should be used is controversial. In 2001, a cost-effectiveness analysis of routine cystoscopy performed to identify ureteral injury at the time of hysterectomy {Obstet Gynecol 2001;97:685-692) reported that routine cystoscopy was cost saving when the rate of ureteral injury exceeded 1.5% at the time of abdominal hysterectomy or 2% in the case of vaginal or laparoscopically assisted vaginal hysterectomy. A number of changes in clinical practice since that time make this question worth revisiting. Because urinary tract injury occurs rarely and not easily studied in a randomized trial, decision analysis is a useful method for evaluating the cost associated with varying strategies for use of cystoscopy. The aim of this decision analysis was to quantify costs of routine cystoscopy, selective cystoscopy, or no cystoscopy at the time of benign hysterectomy. Three separate decision analysis models using TreeAgePro were created for each hysterectomy modality. Models evaluated included cystoscopy following abdominal, laparoscopic/robotic, and vaginal hysterectomy from the perspective of a third-party payer. Bladder and ureteral injuries detected intraoperatively and postoperatively were modeled. Detection of ureteral injury included both false-positive and false-negative results. The model included potential costs of diagnostics (imaging and repeat cystoscopy) as well as treatment (office/emergency room visits, readmission, ureteral stent placement, cystotomy closure, ureteral reimplantation). Costs of peritonitis, urinoma, and vesicovaginal/ureterovaginal fistula were also included in the model. Complication rates were obtained from published literature. When procedure codes could not accurately capture the cost for additional length of stay or workup related to complications, costs were estimated from Medicare reimbursement and published literature. The decision tree analyzed possible outcomes for patients undergoing benign hysterectomy with plan for no cystoscopy, selective cystoscopy, or routine cystoscopy. Selective cystoscopy was used as an intermediate option. With the selective strategy, it was assumed that surgeons would make an intraoperative decision regarding their assessment of urinary tract injury risk for that specific patient. Based on prior literature, incidence of bladder injury was 1.75%, 0.93%, and 2.91% for abdominal, laparoscopic/robotic, and vaginal hysterectomy, respectively, and ureteral injury incidence was 1.61%, 0.46%, and 0.46%, respectively. Cost of hysterectomy with no cystoscopy varied from $885 to $1122. Performing selective cystoscopy added $13 to $26 of the cost of surgery. Routine cystoscopy added $51 to $58 above the cost of selective cystoscopy. Selective cystoscopy becomes cost saving with the increasing risk of urinary tract injury. Selective cystoscopy costs less than no cystoscopy (based on surgical route), when bladder injury exceeds 4.48% to 11.44% or ureteral injury exceeds 3.96% to 8.95%. However, before routine cystoscopy is cost saving, the risk of bladder injury must exceed 20.59% to 47.24%, and ureteral injury, 27.22% to 37.72%. Model robustness was assessed using multiple 1 -way sensitivity analyses, and no reasonable thresholds for model variables were identified other than the rate of ureteral and bladder injury. It is clear that use of selective cystoscopy in patients suffering a complication is more likely to be beneficial with intraoperative detection of injury and prompt treatment.