We used a contemporary cohort and found that roughly 50% of patients undergoing RAPN or RARP receive an intraoperative drain. The omission of drain placement was associated with a shorter length of stay in both RAPN and RARP, no difference in risk of complications after RAPN and was associated with a lower risk of complications after RARP. Additionally, patients who underwent either RAPN or RARP without drain placement had 1% risk of ultimately receiving a postoperative drain or reoperation for a urine leak/fluid collection.

Patients who underwent RAPN without intraoperative drain placement experienced shorter hospital length of stay (median [IQR] LOS 2 [1–3] days vs. 1 [1–2] days, p < 0.001) without difference in complication rates. While there were few differences in baseline characteristics amongst patients who underwent RAPN with or without drain placement, the majority of patients underwent surgery for stage 1 tumors, which is consistent with guideline recommended practice patterns [22]. Moreover, the drainless cohort had less use of postoperative antibiotics, which may have benefits not measured in this study. The dataset is limited in its ability to provide granular data on specific tumor characteristics, such as nephrometry score, that may influence a surgeon’s decision to place a drain [23], but the large number of patients in this national sample reflects real world practice patterns that suggest routine drain placement should be reconsidered.

Patients who underwent RARP without intraoperative drain placement were significantly less likely to have any complication (11.7% vs. 8.39%, p < 0.001), even after adjustment (OR 0.73; 99.58% CI [0.62–0.87]). While we did control for significant differences in the drain and drainless cohorts, including radiation history, longer operative times and extent of lymph node dissection, other unmeasured confounding variables may exist. Still, drainless RARP was associated with decreased hospital LOS on multivariable analysis (β -0.30; 99.58% CI [-0.37, -0.24]) and again, the rate of re-intervention or eventual placement of a drain in the drainless cohort was very low at 1.2%. These results again suggest that routine drain placement should be reconsidered, especially in patients where there is low concern for anastomotic leak.

Our study uses a large patient population to expand on the current data on drainless robotic urologic oncology surgeries. Beksac et al. recently used a multi-institutional database consisting of 904 patients who underwent robotic partial nephrectomy. In their study, 40% of patients underwent drainless robotic partial nephrectomy. Similar to our analysis, they concluded that drainless surgery was associated with shorter length of stay and no significant difference in overall complications or readmission rates [19]. A recent meta-analysis performed by Kowalewski et al. identified 1086 patients undergoing partial nephrectomy and 2520 patients undergoing radical prostatectomy. Their study revealed that drainless partial nephrectomy had similar rates of overall complications (OR: 0.99[0.65;1.51], p = 0.960) and re-intervention (OR: 1.16[0.31;4.38], p = 0.820). However, in their prostatectomy cohort, they found that drainless procedures were associated with a reduction in postoperative complications (OR: 0.62[0.44;0.87], p = 0.006) [21]. Another meta-analysis included 8447 patients undergoing partial nephrectomy and 1890 patients undergoing robotic radical prostatectomy. They found that patients undergoing robotic partial nephrectomy without drain had shorter length of hospital stay (mean difference: -0.84 days, 95% CI: -1.06 to -0.63; P < 0.001) and similar low-grade (P = 0.94) and high-grade (P = 0.31) complications, urinary leakage (P = 0.49), hemorrhage (P = 0.39), reintervention (P = 0.69), and readmission (P = 0.20) compared with routinely drained patients. They also found that patients undergoing drainless robotic prostatectomy had lower rate of postoperative ileus (OR 0.53, 95% CI: 0.38 to 0.74; P < 0.001), but similar low-grade (Clavien 1–2, P = 0.41) and high-grade (Clavien ≥ 3; P = 0.85) complications [24].

Given the unclear benefit of surgical drain placement in our analysis, there is a need to replace traditional dogmas with evidence-based practices, particularly as previous studies suggest that drain placement may negatively impact patient recovery. A single institutional analysis reported heightened postoperative pain with up to 24% of radical prostatectomy patients attributing postoperative pain to drain sites [25].

Our study has several limitations beyond its retrospective design and the limitations inherent to NSQIP. In particular, our data does not contain certain key intraoperative factors (such as a violation of the collecting system in a partial nephrectomy or transperitoneal vs. retroperitoneal approach) or other patient level variables (immunosuppressive disorders or peri-operative anti-coagulation requirements) that likely inform a surgeon’s decision. Second, the NSQIP database only reports 30-day complication and mortality rates. Thus, any complications related to a drainless procedure that occur after 30 days are unreportable. However, complications that may be identified by drain placement such as bleed, urine leak or lymphocele, are typically identified within 30 days. Furthermore, given the relatively limited number of nodes removed during RARP (median number of nodes: 6 vs. 5) in this dataset these findings may not be accurately extrapolated to patients undergoing extended PLND. Lastly, the placement of drains may be related to organizational or situational factors, such as the complexity of the case relative to the training stage of the surgeon, or hospital policy, and it is possible that this increased the number of patients who received drains. Due to the limitations within the dataset, we were unable to adjust for clustering. Randomized controlled trial data, including patient reported outcomes, is needed to directly compare outcomes and complications with respect to drain placement after urologic oncology surgeries in order to identify which patients will benefit most from drainage, while deferring the morbidity in those unlikely to benefit from drain placement. Furthermore, these analyses should include more detailed intraoperative data, particularly regarding the decision-making process behind the placement of a drain, such as difficult ureterovesical anastomosis, prostate size, nephrometry score or tumor complexity.