Standardized analysis of complications after robot-assisted radical cystectomy: Korea University Hospital experience

Open radical cystectomy with urinary diversion is considered the gold standard treatment for patients with invasive bladder cancer [1]. Despite improvements in surgical technique, technology, and perioperative care, cystectomy with urinary diversion remains a morbid operation with a substantial complication rate [2]. Recently, interest in laparoscopy and robot-assisted radical cystectomy (RARC) has increased with the goal of reducing procedure-related morbidity in the field of urology [3]. RARC has already superseded the use of pure laparoscopic radical cystectomy and is increasingly becoming an available option at major tertiary-care centers [4]. Therefore, RARC is representative of minimally invasive surgery for muscle-invasive transitional cell carcinoma of the bladder.

However, because adverse event reporting is highly variable and nonstandardized in the urologic literature, data on complications have mostly been reported by multiple surgeons. Also, most reports used only extracorporeal urinary diversion (ECUD) with a small skin incision to bring out the intestine following completion of robotic cystectomy [5]. Given that intracorporeal urinary diversion (ICUD) has only recently been attempted, studies regarding it have been rarely reported. In addition, it is difficult to analyze and interpret the data because the reporting system is not standardized. Also, most of the research is based on surgical outcomes that do not take into account the surgeons' technique. Therefore, reporting of complications on the basis of an established system is required.

We aimed to systemically analyze the complications after RARC performed by a single surgeon at a tertiary referral hospital including ICUD by use of a standardized reporting methodology.

The clinical and pathological characteristics of the patients are summarized in Table 1. The demographic data and perioperative outcomes in patients who underwent ECUD or ICUD were compared as well. There were no significant differences in demographic data between the ECUD and ICUD groups. The mean age of the patients was 63.3 years (range, 39-79 years). Mean overall operative time was 496 minutes (range, 310-700 minutes) and mean EBL was 247 mL (range, 50-600 mL). Mean length of hospital stay was 17.3 days (range, 8-47 days). Thirteen patients (25%) underwent laparoscopic adhesiolysis owing to prior abdominal surgery before completion of the docking process. For urinary diversion, 27 patients (51.9%) had an ileal conduit, and 21 (40.4%) had an orthotopic neobladder. Four patients (7.7%) had an ureterocutaneostomy; two of them had a single kidney and the others underwent concomitant nephroureterectomy because ureteric frozen pathology was consistently positive during the operation. The mean number of LNs harvested was 24.6 (range, 4-71), with 30.5 (range, 8-71) in extended PLND.

Table 1
Patient characteristics and perioperative outcomes

• Click for larger image
• Click for full table
• Download as Excel file

Urinary diversion was performed both extracorporeally and intracorporeally. Up to the 37th case, it was performed extracorporeally. Urinary diversion was then performed intracorporeally in the more recent cases. In comparison with the ECUD group, mean overall operative time for the ICUD group was prolonged (464 minutes vs. 615 minutes, p<0.001), but mean EBL was lower (269 mL vs. 165 mL, p=0.003). In addition, cases of extended LND and the mean number of LNs harvested were greater in the ICUD group (p=0.01 and p=0.02, respectively).

Fifty percent of the patients (26 of 52) experienced at least one complication of any grade within 90 days of surgery, and 11 patients (21.2%) experienced a major complication (Table 2). The rate of complications was compared according to the diversion method as well. In the ECUD group, 56% of the patients (23 of 41) experienced at least one complication of any grade within 90 days of surgery. Twenty patients (48.8%) and nine patients (22%) each experienced a minor and a major complication. On the other hand, in the ICUD group, 27% of the patients (3 of 11) experienced at least one complication of any grade within 90 days of surgery; one patient (9.1%) and two patients (18.2%) each experienced a minor and a major complication. The incidence of minor complications for patients in the ECUD group was higher than for patients in the ICUD group (48.8% vs. 9.1%). There was a significant difference in the Fisher exact test (p=0.03) (Table 2).

Table 2
Complication summary

• Click for larger image
• Click for full table
• Download as Excel file

Complications were grouped in systems-based categories (Table 3). Fifty complications occurred in 52 patients, and hematologic complication (transfusion) was the most common. Twenty-five percent of patients (n=13) required blood transfusion, and urinary tract infection (n=7) and fever (n=4) were the most common infectious complications. Wound dehiscence, anastomotic leakage, urinary tract obstruction, mechanical obstruction, and thromboembolism occurred as major complications in 11 patients. One patient developed mechanical obstruction. Therefore, the patient underwent an additional colorectal operation for repair. Five patients needed a wound revision owing to wound dehiscence. All of them underwent RARC with the ECUD method. Among them, three patients underwent wound closure, which was performed under general anesthesia and thus was categorized as a major complication. Two patients underwent simple suturing under local anesthesia as a minor complication. Anastomotic leakage occurred in four patients; three of them required catheterization for drainage of ureteral anastomotic leakage, and one patient underwent an operation for repair of urethral anastomotic leakage. Urinary tract obstruction occurred in three patients. One patient had to undergo transuretero-ureterostomy, another patient needed to have ureterocutaneostomy stoma revision, and the other only underwent temporary catheterization. Vascular thrombo-embolitic events occurred in two patients. Atherosclerosis obliterans of the left toe occurred in one patient; therefore, he needed a procedure for stent insertion. Medication was enough for the other patient who developed a deep vein thrombosis.

Table 3
Complication summary in systems-based categories

• Click for larger image
• Click for full table
• Download as Excel file

Regarding predictors of major complications, in the univariable and multivariable analyses, EBL (over 300 mL) was significant in predicting a complication (Table 4). Concerning predictors of minor complications, EBL (over 300 mL), operative time, and method of urinary diversion (extra- vs. intracorporeal) were significant negative predictors in the multivariable analysis, whereas EBL (over 300 mL) and method of urinary diversion were significant negative predictors in the univariable analysis. EBL (over 300 mL) was a significant negative predictor of total complications in the univariable analysis, and EBL (over 300 mL) and operative time were significant negative predictors of total complications in the multivariable analysis. No patients died within 90 days.

Table 4
Univariable and multivariable logistic regression analysis of complications

• Click for larger image
• Click for full table
• Download as Excel file

Standardized reporting of postoperative complications is vital for treatment planning, clinical trial design, assessing innovative surgical techniques, and informed patient counseling [9]. However, adverse event reporting is highly variable and nonstandardized in the urologic literature, and the complication data have mostly been reported by multiple surgeons. Most studies evaluating the morbidity of radical cystectomy do not use a formal grading system, account for comorbidities, or clearly define various complications [9]. Consequently, it is difficult to interpret the data owing to the nonstandardized reports and inconsistent surgical outcomes by multiple surgeons with different surgical techniques.

We aimed to analyze the complications after RARC performed by a single surgeon at a single center by use of a standardized reporting methodology. At our institution, 50% of patients (26 of 52) experienced at least one complication of any grade within 90 days of surgery, and 11 patients (21.2%) experienced major complications (Table 2). These results show that RARC has fewer complications than in recently published open series reported by use of a similar methodology. Shabsigh et al. [10] used a standardized reporting methodology in a series of 1,142 open radical cystectomy patients. Within 90 days postoperatively, 64% of patients experienced a complication, and the major complication rate was 13%. Recently, smaller series reviewing initial complication rates of RARC have reported complication rates ranging from 34% to 63% [3, 11, 12]. Pruthi et al. [13] reported an overall complication rate of 36% among 100 consecutive RARC cases but included complications that occurred within only 30 days of cystectomy. Ng et al. [14] reported a 48% complication rate within 90 days of RARC by use of standardized reporting methods. More recently, in a series of 156 RARC patients, Hayn et al. [11] identified 52% of patients having experienced a complication within 90 days of surgery, and the major complication rate was 24%. Yuh et al. [2] reported that the 90-day complication rate was 80% by use of the modified Clavien-Dindo system, and the major complication rate was 35%. The 50% complication rate and 21.2% major complication rate in our study are comparable to those of contemporary robot series.

Fif ty complications occurred in 52 patients, and transfusion was the most common complication (n=13). Transfusion rates f or contemporary open radical cystectomy series have ranged from 38% to 66% [10, 15]. For robot series, Yuh et al. [2] and Hayn et al. [11] reported rates of 43.9% and 16%, respectively. In our study, the transfusion rate was 25%. The results suggest that the robotic surgery helps to reduce blood loss and the need for blood transfusion.

The mean operative time of 496 minutes is longer than in many open series and other RARC series. Hayn et al. [11] reported a median operative time of 378 minutes and Yuh et al. [2] reported 432 minutes. The time difference may be explained by the urinary diversion type and learning curve. Hayn et al. [11] reported that 7% of patients (11 of 156) underwent orthotopic neobladder formation, and Yuh et al. [2] reported that 44% of patients (86 of 196) underwent orthotopic neobladder formation [2, 11]. In our study, 21 patients (40.4%) underwent orthotopic neobladder, and 11 patients (21%) underwent ICUD in the RARC group. We also need to consider the associated learning curve.

In open RC, Shabsigh et al. [10] reported that age, prior abdominal surgery, and EBL were significantly associated with experiencing a grade III or higher complication. Similarly, Yuh eh al. [2] reported that comorbidity, preoperative hematocrit, and orthotopic diversion were significantly associated with experiencing major complications. In our study, as predictors of minor complications, EBL (over 300 mL), operative time, and method of urinary diversion were significant in the multivariable analysis. EBL (over 300 mL) and operative time were signif icant negative predictors of total complications in the multivariable analysis. EBL (over 300 mL) was a significant predictor of major complications in the univariable and multivariable analyses.

Urinary diversion performed by using a total ICUD method can have additional benefits, including less incisional pain, decreased bowel exposure and desiccation, and the potential for decreased fluid imbalances [16]. Because ICUD has only recently been attempted, studies related to it are scarce. Hence, determining whether to create a complete ICUD versus performing the reconstruction extracorporeally is still a controversial and critical issue [8]. Pruthi et al. [16] compared 32 cases of RARC with ICUD (n=12) or ECUD (n=20). Although the operation time was longer in ICUD patients (5.3 hours) than in ECUD patients (4.2 hours), the complication rates within 30 days were similar between the two groups (42% vs. 40%). Guru et al. [17] reported a comparison between 13 ICUD and 13 ECUD procedures, with four complications in the ICUD group and five in the ECUD group in a 3-month follow-up. In our study, we also compared the rates of complications between the urinary diversion methods. We found a difference in operative time between the ECUD group (464 minutes) and the ICUD group (615 minutes). As a result, the incidence of minor complications was higher for patients in the ECUD group than for patients in the ICUD group (48.8.0% vs. 9.1%). There were significant differences in the Fisher exact test (p=0.03) (Table 2). Yet, there was no significant difference in major complications. As a predictor of minor complications, ECUD was a significant factor in the univariable and multivariable analyses (p<0.05). The reason that ECUD is regarded as a risk factor is due to the fact that ICUD, which was performed later in the series, had fewer complications because the learning curve had already been achieved. Also, the number of patients who underwent ICUD may have been too few to achieve statistical significance. Recently, a study of an international robotic cystectomy consortium with 935 patients reported that the 90-day complication rate was not significantly different between the groups, but a trend favoring ICUD over ECUD was noted (41% vs. 49%, p=0.05) [18]. As mentioned before, ICUD is considered a leader in the minimally invasive trend, although it has vulnerability. For the practical application of ICUD, more global data with respect to long-term outcomes and a larger series of cases are required.

This study had limitations. It was a retrospective and nonrandomized study and may have contained selection bias. The sample size was relatively small, which limits the statistical power of the analysis as well. Because the learning curve varies according to the surgeon's technical characteristics, supporting data from other surgeons with similar experience levels will be needed to identify the advantages of RARC in the early learning period.

The present study showed that the complication rate reported by use of a standardized methodology after robotic radical cystectomy is still considerable, although comparable to those of contemporary robot series. EBL (over 300 mL), operative time, and diversion methods were predictors of complications after RARC. ICUD has a lower risk of postoperative complications. Larger, randomized studies are required to compare complications between open radical cystectomy and robotic radical cystectomy in order to define the role of robotics in radical cystectomy.