To evaluate retroperitoneal robot-assisted partial nephrectomy (RAPN) against transperitoneal approach in a multi-institutional prospective database, after accounting for potential selection bias that may affect this comparison.

Post-hoc analysis of the prospective arm of the Vattikuti Collective Quality Initiative database from 2014 to 2018. Six hundred and ninety consecutive patients underwent RAPN by 22 surgeons at 14 centers in 9 countries. Patients who had surgery at centers not performing retroperitoneal approach (n = 197) were excluded. Inverse probability of treatment weighting was done to account for potential selection bias by adjusting for age, gender, body mass index, comorbidities, side of surgery, location/size/complexity of tumor, renal function, American Society of Anesthesiologists score, and year of surgery. Operative and perioperative outcomes were compared between weighted transperitoneal and retroperitoneal cohorts.

Ninety-nine patients underwent retroperitoneal RAPN; 394 underwent transperitoneal RAPN. Hospital stay in days—median 3.0 (Interquartile range [IQR] 2.0-4.0) transperitoneal vs 1.0 (1.0-3.0) retroperitoneal; P < .001, and blood loss in mL—125 (50-250) transperitoneal vs 100 (50-150) retroperitoneal; P = .007—were lower in the retroperitoneal group. There were no differences in operative time (P = .6), warm ischemia time (P = .6), intraoperative complications (P = .99), conversion to radical nephrectomy (P = .6), postoperative major complications (P = .6), positive surgical margins (P = .95), or drop in estimated glomerular filtration rate (P = .7).

In a multi-institutional setting, both retroperitoneal and transperitoneal approach to RAPN have comparable operative and perioperative outcomes, except for shorter hospital stay with the retroperitoneal approach.