Urinary fistula after robot-assisted partial nephrectomy: A multicentre analysis of 1 791 patients

Objective To evaluate the incidence of and risk factors for a urine leak in a large multicentre, prospective database of robot-assisted partial nephrectomy (RPN). Patients and Methods A database of 1 791 RPN from five USA centres was reviewed for urine leak as a complication of RPN. Patient and tumour characteristics were compared between patients with and those without postoperative urine leaks. Fisher's exact test was used for qualitative variables and Wilcoxon sum-rank tests were used for quantitative variables. A review of the literature on PN and urine leak was conducted. Results Urine leak was noted in 14/1 791 (0.78%) patients who underwent RPN. The mean (sd) nephrometry score of the entire cohort was 7.2 (1.9), and 8.0 (1.9) in patients who developed urine leak. The median (range) postoperative day of presentation was 13 (3-32) days. Patients with urine leak presented in delayed fashion with fever (two of the 14 patients, 14%), gastrointestinal complaints (four patients, 29%), and pain (five patients, 36%). Eight of the 14 patients (57%) required admission, while eight (57%) and nine (64%) had a drain or stent placed, respectively. Drains and stents were removed after a median (range) of 8 (4-13) days and 21 (8-83) days, respectively. Variables associated with urine leak included tumour size (P = 0.021), hilar location (P = 0.025), operative time (P = 0.006), warm ischaemia time (P = 0.005), and pelvicalyceal repair (P = 0.018). Upon literature review, the historical incidence of urine leak ranged from 1.0% to 17.4% for open PN and 1.6-16.5% for laparoscopic PN. Conclusion The incidence of urine leak after RPN is very low and may be predicted by some preoperative factors, affording better patient counselling of risks. The low urinary leak rate may be attributed to the enhanced visualisation and suturing technique that accompanies the robotic approach.