Is a Drain Needed After Robotic Radical Prostatectomy With or Without Pelvic Lymph Node Dissection? Results of a Single-Center Randomized Clinical Trial

Introduction

Robot-assisted radical prostatectomy (RARP) is a well-established procedure for treating nonmetastatic prostate cancer (PCA) and can be performed with extended pelvic lymph node dissection (ePLND), which is suitable in all high-risk cases and intermediate-risk patients when the probability of lymph node invasion is higher than 5% by nomograms. ^1,2 Systematic review of literature has shown that RARP with or without ePLND can be performed safely. ³

At the end of an RARP procedure, it has been suggested that a drain should be positioned in the pelvis and removed early, while the drain could be optional in cases with extremely low risk of postoperative hemorrhage or urinary leakage. ⁴ Actually, many institutions do not routinely use a drain after RARP. Although a recent randomized controlled trial (RCT) showed no significant benefits in placing a drain in the pelvic cavity after RARP, no definitive confirmed evidences have been reported. ⁵ To address this issue, we planned an RCT, which aimed at investigating the safety of not placing a drain in the pelvic cavity after RARP with or without ePLND.

Materials and Methods

The study had institutional review board approval, but did not undergo local ethics committee approval because the policy of not placing a drain in the pelvic cavity after RARP is an option that has been already recommended by an expert panel of urologists during an international consensus conference ⁴ and because it is already a routine practice in many institutions all over the world. Each patient provided informed signed consent.

Patients were classified according to the D'Amico cancer class risk categories. ⁶ Clinical staging was computed according to the 2002 American Joint Committee on Cancer staging system for PCA. In each case, pelvic lymph node staging (cN) was performed by CT and/or by multiparametric MRI modalities. When pelvic nodes measured within 1 cm in diameter, clinical staging status was coded cN0. Enlarged pelvic nodes measuring more than 1 cm in diameter were staged as cN1 disease. The metastatic status was investigated by CT, bone scan, and occasionally by 11-choline positron emission tomography CT (11C-PET CT). The perioperative surgical risk was evaluated by the American Society of Anesthesiologists (ASA) score. ⁷ Surgical complications, which were monitored within a period of 3 months, were graded by the Clavien–Dindo score (CDS) system and reported according to the European Association of Urology panel assessment and recommendations. ^8,9 The event of hospital readmission (RAD) was also evaluated. All patients were closely monitored monthly in the 3-month follow-up visit period to detect any complication and RAD. In intermediate-risk cases, ePLND was performed when the risk of nodal metastases was greater than 5% by nomograms. ² In low-risk cases, the decision to perform ePLND was based on the risk of tumor upgrading toward high patterns in the surgical specimen. ¹⁰ In high-risk patients, ePLND was always performed.

Biopsies performed elsewhere were not revised and were accepted according to the following criteria: (1) at least 12–14 biopsy cores; (2) the reported number of positive cores; and (3) measurement of prostate volume (PV; mL). In our institution, the 14-core transperineal-guided prostate biopsy technique was used and PV was measured according to standard methods. Surgical specimens were assessed by a dedicated pathologist. All specimens were processed according to the Stanford protocol. ¹¹ Tumors were assessed by the new PCA grading group system. ^12,13 Nodal packets were tagged and submitted in separate packages. The removed lymph nodes were assessed for histopathologic analysis after hematoxylin and eosin staining. Immunohistochemical staining was performed when appropriate. In each case, the number of removed and metastatic nodes was assessed. Prostate and nodal specimens were then staged according to the 2002 American Joint Committee on Cancer staging system for PCA. ¹⁴

Patients who had had surgery for benign prostatic hyperplasia, radiation of the pelvis, and focal or androgen deprivation therapy for PCA were excluded from the study. From May to December 2016, 112 consecutive patients who underwent RARP with or without ePLND were prospectively randomized in a control and study group. Randomization respected time sequence criteria, computed by casual number table. ¹⁵ All procedures were in accordance with the ethical standards established in our country. Internal review board approval was not required for such a prospective study. A postoperative drain was placed in the pelvis of the control group; it was removed on postoperative day (POD) 1 if the output was less than 150 mL/24 hours. If the postoperative outcome was uneventful, both groups were discharged on POD 4 with the catheter removed on POD 12 without cystography. RARP was delivered by the Si da Vinci Robot System (Intuitive Surgical, Inc., Sunnyvale, CA) and was performed through the transperitoneal approach with antegrade prostatic dissection. ^4,16 A transurethral 18F Foley catheter was placed in all cases and the balloon was inflated with 7 mL of physiologic solution. Extended PLND was performed according to a standard template, which included the obturator fossa and external, internal, and common iliac lymph nodes. The operating time was calculated as the interval between incision of the first laparoscopic port and suture of the last laparoscopic port. The surgical procedures were performed by two experienced surgeons, coded as x and y, who were blinded to the results of randomization before surgery. Prophylaxis of deep venous thrombosis (DVT) with low-molecular-weight heparin was performed in all cases elected to RARP with extended PLND and with comorbidity risk factors. ⁴ In all cases, early mobilization was activated starting from the day of surgery (POD 0). DVT prophylaxis was prolonged till POD 28.

Results

The trial randomized 112 consecutive patients, of whom 2 were excluded because of higher than usual intraoperative bleeding. The two patients being excluded belonged to the study group and thus had a drain placed in the pelvic cavity, which, in one case, was not removed on POD 1. In the control group, all drains were removed by protocol on POD one. Finally, the study evaluated 110 cases, of whom 56 belonged to the control group.

Table 1 shows clinical features of the population (n = 110) stratified by the control group (n = 56) and study group (n = 54). The median age was 65 years with an ASA score of two in 75.5% of cases. Median prostate-specific antigen (PSA) was 6.5 ng/mL. and the median proportion of biopsy-positive cores was 0.3 In the patient population, PCA was classified low risk in 33 cases (30%), intermediate risk in 62 (56.4), and high risk in 15 patients (13.6%). ePLND was performed in 71 cases (64.5%). The two randomized groups were homogenous for all assessed preoperative clinical factors except for PV, which was significantly lower in the control group.

Table 2 shows the pathologic and perioperative features of the population stratified by the two randomized groups. The pathology grade group was 1 in 14 cases (12.7%), 2–3 in 77 (70.0%), and greater than 3 in 19 patients (17.3%). The disease was staged intraprostatic in 89 cases (80.9%) and extraprostatic in 21 patients with extracapsular extension in 8 (7.3%) and invasion of the seminal vesicles in 13 (11.8%). The surgical margins were negative in 82 cases (74.5%) and positive in 28 patients with focality in 17 (15.5%) and multifocality in 11 (10.0%). The median number of removed lymph nodes was 27 (range 15–56) and metastases were detected in 8 of 71 patients who underwent ePLND (11.3%). The median values of operating time and blood lost were 207 minutes (range 90–335) and 200 mL (range 50–1700), respectively. The transfusion rate was 1.8%. The two groups were homogenous for all evaluated variables, which also included the weight of the removed prostate.

In Table 3, postoperative features of the patient population are stratified by randomized groups. The median LOHS was 4 days (range 3–8). Overall, complications were detected in 27 cases (24.5%). Complications were scored as follows: grade 0 in 83 patients (75.5%), grade 0 in 20 (18.2%), grade 2 in 4 (3.6%), and grade 3a in 3 patients (2.7%). The RAD rate was 3.6%. The two randomized groups did not show any difference for both the primary endpoint, which was the CDS system (p = 0.254), and the secondary endpoints, which included LOHS (p = 0.689) and RAD rate (p = 0.970).

Table 4 shows specific overall complication rates related to the CDS system. Urinary leakage was detected in three cases (2.7%), of whom two had a grade 3a complication, which needed the placement of a Foley-type catheter that was removed on POD 44 (one case) and 54 (one case). In one patient, urinary leakage was graded 1 and the catheter was removed on POD 21 after cystography. In three patients (2.7%), the complication was labeled as grade 1 because it needed a control cystography, which was performed on POD 4, for suspected urinary leakage; however, the catheter was removed on POD 12, according to protocol schedule. In one case (0.9%), a symptomatic lymphocele required puncture and drainage (CDS 3a). Pelvic hematoma was detected in four patients (3.6%) and did not require further procedures (grade 2 complication). Fever complicated (grade 1) postoperative outcome in five patients (4.5%). Four patients were readmitted because of fever (grade 1), pelvic hematoma (grade 2), urinary leakage (grade 3a), and symptomatic lymphocele (grade 3a).

Discussion

In high-volume centers, RARP with or without ePLND is a well-established and safe surgical procedure. ^1,4,17,18 The Pasadena Consensus Conference recommended that, at the end of RARP procedure, a drain should be placed in the pelvis and removed early, but the drain could be optional in cases with extremely low risk of postoperative hemorrhage or urinary leakage. ⁴ The decision of not placing a drain in the pelvic cavity at the end of the RARP procedure is already routine in many institutions. In a prospective study, including 143 consecutive patients who were operated by the same surgeon, it has been reported that after RARP with ePLND, a drain was left in the pelvis and removed after 24 hours if the output was less than 100 mL/24 hours. ¹⁹ Moreover, in the last 50 cases, it was planned not to place a drain if bladder neck reconstruction was not performed and anastomosis was watertight; as a result, 41 patients had no drain postoperatively. ¹⁹ Recently, a prospective randomized trial has shown that, in similar cohorts of patients, postoperative adverse events were not inferior in the no-drain group when compared with control cases. ⁵ Chenam and colleagues reported data on patients who underwent RARP from September 2012 to April 2016, including 189 patients (43 study period; 4.3 procedures per month). ¹⁹ In our study we reported data on patients who underwent RARP from May 2016 to December 2016 (6-month period; 18.6 procedures per month). Considering the number of patients performed per month, our population sample is even more representative. Moreover, in a previous RCT, the population was a parallel group trial, ⁵ while in our study, the population consecutively randomized. According to the best of our knowledge, this is the second randomized study trying to address this issue. The randomization showed that preoperatively the two groups were homogenous for all evaluated variables except for PV, which was significantly larger in the study group; however, looking at the pathologic results, the weight of the removed prostates between the two groups was not significant (Table 2), and thus, the bias depended on biopsies, which were not all performed in our institution. Perioperative variables, including operating time, blood lost, and transfusions, were not significantly different between the groups. Finally, clinically features associated with primary and secondary endpoints showed no significant differences of Clavien–Dindo grade complications, LOHS, and RAD rates between the groups (Table 3). The results of our study showed that the no-drain policy after RARP with or without ePLND was safe and equivalent to the use of a drain for 24 hours or more. The results of our study confirm those of a similar trial, performed in North America for a longer period (about 4 years), including 189 randomized patients. ⁵ In a period of 90 days, overall Clavien–Dindo complication grades were 26.8% in the drain group and 17.4% in the no-drain group, which were similar to those detected in our cohorts (28.7% and 20.4% in the control and study group, respectively).

Complications following RARP surgery with or without ePLND should be accurately reported; moreover, complications should be coded according to validated grading systems. ^3,4,9 In literature reviews, overall mean outcomes have been reported as follows: operating time 52 minutes (RARP without PLND), blood lost 166 mL, transfusion rate 2%, LOHS 1.9 days, and indwelling catheter 6.3 days. ³ In high-risk patients, median operative perioperative outcomes were as follows: 168 minutes of operating time, 189 mL of blood lost, and mean LOHS 7.2 days. ¹⁸ Yuh and colleagues have shown that median perioperative outcomes after RARP with ePLND in intermediate- and high-risk patients were as follows: operating time 180 minutes, estimated blood lost 200 mL, no transfusions, lymph node yield of 20 nodes, and LOHS 1.8 days. ¹⁹ Recently, a controlled Australian study comparing RARP vs retropubic radical prostatectomy has shown that, in the RARP arm, mean perioperative outcomes were as follows: operating time 202.03 minutes, blood lost 443.74 mL, no transfusions, 6.5 nodes removed, LOHS 1.55 days, and indwelling catheter 8.5 days. ²⁰ In our study, overall perioperative outcomes are reported in Table 2, showing that operating time was longer than that reported by Yuh and colleagues and by the Australian study, but blood lost was lower than the Australian study and similar to that reported by Yu and colleagues; however, the median number of removed nodes was larger than the other two studies. ^19,20 As a standardized internal guideline, the study protocol considered patient discharge on POD 4 and catheter removal on POD 12. As a result, there was a longer LOHS and time of indwelling catheter that was removed on POD 12 without cystography in 107 patients.

The issue of not placing a drain in the pelvic cavity has also been considered in dated series, which included patients who underwent open radical prostatectomy. Indeed, it was shown that the decision of not placing a drain in the pelvic cavity did not increase the risk of complications if the bladder neck was preserved or sharply reconstructed. ^21,22 Later on, other investigators realized that this policy could also apply to RARP cases. ²³ However, these trials were noncontrolled, and thus, the validity of the results was biased by patient selection criteria. ^21,22 The issue was again considered by a panel of experts who recommended that a drain in the pelvis after RARP could be omitted in selected cases. ⁴ Finally, the recent North American randomized trial has shown that, in similar cohorts of patients, postoperative adverse events were not inferior in the no-drain group when compared with control cases. ⁵ Our study is the first prospective controlled trial confirming that the no-drain policy is equivalent to drain after RARP with or without ePLND in terms of complications rate, LOHS, and RAD rate.

Postoperative complications following RARP include lymphorrhea, lymphocele, pelvic hematoma, urinary leakage, disrupted anastomosis, and penile shortening. ³ Studies dealing with predictors of complications after RARP are limited and confusing. ^3,4 It has been suggested that comorbidity, PV, PSA, biopsy Gleason score, and surgical experience could be considered. ⁴ Literature reviews show that mean complication rates of RARP are 9% with a range between 3% and 26%. ^3,4 In high-risk series, overall complication rates range from 3% to 30%. ¹⁸ Yuh and colleagues reported an overall complication rate of 18%, while the complication rate was 6.7% in the Australian study. ^19,20 The most important complications are scored as grade 3, according to the CDS system. Yuh and colleagues reported 4.9% of grade 3 complications, while the Australian investigators assessed a rate of 1.1% of complications graded as 3a. ^19,20

In our investigation, the overall complication rate was 24.5%, which was closer to the results of Yuh and colleagues than the Australian study, but this may be related to the extension of the nodal yield, which was larger in the study of Yuh and colleagues (median 20) and in this present investigation (median 27) than that of the Australian study (mean 5.50 nodes). The rate of grade 3a CDS complications was 2.7%, which was lower than the results of Yuh and colleagues, but higher than the Australian trial (RARP arm).

Complications of challenging cases, who were operated by RARP, are controversial and controlled trials focused on this issue are lacking. ^3,4 In our study, cases with previous transurethral resection of the prostate and salvage RARP did not enter randomization. The median body mass index was 25.4 kg/m² and there were no differences between the two groups. Median values of PV measured significantly larger in the study than the control group; however, median prostate weight did not differ between the groups.

There are several limitations to our study. First, it was a single-institution trial with only two surgeons performing RARP; so far, our results might not apply to other centers. Second, the results might be limited by the small size of the population that was, however, adequate for the aim of the study. However, beyond these limitations, our study addresses and gives evidence that not placing a drain after RARP with or without ePLND is safe and equivalent to the drain policy. Since this internal clinical trial, our routine policy became “no drain” except those cases where the urethrovesical anastomosis was particularly difficult and possibly not watertight.

Conclusions

In a modern cohort of PCA patients who underwent RARP with or without ePLND, a single-center RCT showed that the no-drain policy is equivalent to drain after RARP with or without ePLND in terms of complication rate, LOHS, and RAD rate. The option of placing a postoperative drain for the first 24 hours could be considered in cases of difficult urethrovesical anastomosis.