1 INTRODUCTION

Radical prostatectomy is currently the gold-standard treatment for clinically localised prostate cancer (PCa).¹ From the time of its introduction, robot-assisted radical prostatectomy (RARP) has changed perspectives regarding surgical options and gained popularity over open and purely laparoscopic prostatectomy.² RARP affords numerous benefits, such as reduced blood loss and fewer days of hospitalisation^{3, 4} while maintaining equivalent oncological and functional outcomes.^{4, 5} Despite the rapid diffusion of RARP as the standard of care internationally, the need for routine insertion of a pelvic drain remains equivocal.⁶ The rationale for the insertion of a drain following RARP is multifaceted, with many indications continuing to be anchored to past paradigms not necessarily relevant to the ‘new era’ of robotic surgery.

A urinary leak from the urethrovesical anastomosis is one of the most significant short-term complications of RARP. The aim of our study was to assess the yield of pelvic drain outputs in predicting early-onset leak from a vesicourethral anastomosis and determine the necessity of routinely placing a pelvic drain following RARP.

2 METHODS

3 RESULTS

The cohort included 204 patients with a median age of 66 years (IQR 62–70 years) and a median PSA level of 7.3 ng/dL (IQR 5.5–11 ng/dL). Their baseline clinical, surgical, and pathological characteristics stratified by urinary leak status are presented in Table 1. The preoperative, operative, and pathological characteristics were not different between the urinary leak group and the no-leak group (Table 1). The early-onset anastomotic urinary leak rate during hospitalisation was 3.4% (n = 7 patients).

The median D-POD1 was significantly higher (P < .001) among patients with a urinary leak (median 250 mL, IQR 130–625 mL) compared with others (median 80 mL, IQR 50–140 mL). The patients who sustained early-onset urinary leaks consistently presented with a higher D/U ratio on POD1 (median 0.22 vs 0.04, respectively; P < .001; Table 2). These differences were no longer significant on POD2. The receiver operating characteristics curve (Figure 1) was used to evaluate the model's predictive performance for early-onset vesicourethral anastomotic leakage. The receiver operating characteristics curve determined that D-POD1 = 227 mL was the threshold for predicting an anastomotic leak (area under the curve 0.88, 95% confidence interval 0.78–0.99; P < .001). A univariate logistic regression analysis demonstrated a significant association between D-POD1 >227 mL and the presence of a urinary leak (odds ratio 35, 95% confidence interval 6–200; P < .001). The minimum number needed to predict a single patient with a urinary leak was 29.

4 DISCUSSION

Robotic technology bestows the possibility of preserving key anatomic structures and minimising the perioperative complications during radical prostatectomy, thereby vastly changing the perception of prostate surgery. RARP has now become the new ‘gold standard’ approach for the surgical treatment of localised PCa. The vesicourethral anastomosis made possible by the robotic approach is more accurate and tighter compared with other approaches, resulting in fewer occurrences of anastomotic urine leak and subsequent urinoma.¹⁰ Previous studies on this new approach had focused mainly on the major issues of oncological and functional outcomes. We now consider that enhanced imaging capabilities, better handling of tissues, and the finer movements afforded by the robotic approach may have decreased the level of concerns from a leak. This possibility was supported by more recent studies that estimated the incidence of anastomotic leak during RARP to range between 0.1% and 6%.^{11, 12}

Danuser et al¹³ questioned the need for pelvic drain placement in RARP and suggested that it may be significantly less than in open prostatectomy techniques. Even earlier, Sharma et al¹⁴ presented a concept of omitting the drain altogether. There are, however, numerous considerations underlying the rationale of placing a pelvic drain during this type of surgery, especially its role as an early indicator of anastomotic leak, bleeding, and lymphatic leak. A recent randomised controlled trial showed no significant benefit in placing a drain in the pelvic cavity after RARP.⁶ Moreover, drain placement has been associated with surgical site infections,¹⁵ pain at the drain site,¹⁶ and bleeding.¹⁷ Niesel et al¹⁶ found that one-third of patients had post-operative pain associated with either the incision or drain site after radical prostatectomy and retroperitoneal drainage after gynaecologic abdominal surgery was found to increase post-operative hospital stay with higher complication rates (P < .001). These may interfere with the ‘fast-track’ pathway following surgery by delaying hospital discharge and consequently increasing hospital costs.

Our study focused on the potential yield of a pelvic drain in predicting post-operative early-onset urinary leak from the vesicourethral anastomosis, and the results confirmed that the drain output represents a valid prediction tool for urinary leak. Interestingly, the significance between the drain outputs of patients who sustained urinary leaks and those who did not vanished on POD2. These findings suggest that drain output is a valid predictor of urinary leak, but that its effectiveness is less evident after the first post-operative day. This could be related to the increase in lymph secretion caused by the irritative effect of drainage in the pelvic cavity. In addition, with the early-onset anastomotic urinary leak rate identified by cystography to be as low as 3.6%, the ‘number needed to drain’ to be able to predict a single patient with urinary leak on cystography is 29 patients. Considering these findings alongside the relatively low rate of urinary leak complications associated with a robotic approach, we conclude that, unless specifically indicated otherwise, it may be safe to forgo routine placement of a pelvic drain.

There are several limitations to this study. First, it is a single-institution retrospective cohort study with a sample size that may limit the significance of the findings. Second, the low number of urinary leaks precluded the performance of a multivariate analysis. Despite these limitations, we believe that this study contributes clinically relevant findings to the ongoing uncertainty of the value of drain placement following RARP.

In conclusion, the placement of a pelvic drain can serve as a useful tool to predict early-onset anastomotic urinary leaks in the post-operative management of patients after RARP. The findings of our analysis, however, lead us to consider that its routine use in RARP may be avoided and saved to selected patients, based on the relatively low rate of this complication. Further prospective randomised trials are needed to establish our results.

AUTHOR CONTRIBUTIONS

Conception and design—Eugenio Bologna, Hilly Perlman and Ziv Savin. Acquisition of data—Eugenio Bologna, Hilly Perlman, Idan Zeeman, Tomer Bashi, and Karin Lifshitz. Analysis and interpretation—Eugenio Bologna, Hilly Perlman Snir Dekalo, Roy Mano, and Ziv Savin. Drafting of the manuscript—Eugenio Bologna, Hilly Perlman and Ziv Savin. Critical revision of the manuscript—Avi Beri, Ofer Yossepowitch, Snir Dekalo, and Roy Mano, Yuval Bar-Yosef. Statistical analysis—Ziv Savin and Roy Mano. Supervision—Ziv Savin.

CONFLICT OF INTEREST STATEMENT

All authors have nothing to disclose.