European Urology

1. Introduction

Radical cystectomy with extended pelvic lymphadenectomy represents the most effective surgical treatment of muscle-invasive transitional cell carcinoma of the bladder and is a viable alternative for high-risk or refractory superficial disease [1]. Surgery has been traditionally performed with an open approach, but recently the use of laparoscopic and robot-assisted radical cystectomy (RARC) has increased [2], [3], [4], [5], and [6]. Novel surgical techniques should ideally offer technical, functional, and oncologic enhancements. Minimally invasive surgeons have promoted decreased blood loss and shorter convalescence as benefits of laparoscopic radical cystectomy or RARC. Long-term oncologic and functional outcomes have not been adequately studied.

Despite improvements in surgical technique, technology, and perioperative care, cystectomy with urinary diversion remains a morbid operation with a substantial complication rate. Unfortunately, longitudinal comparative analyses have been difficult to interpret because of differences in reporting and nonstandardization. To objectively assess the full range of adverse events for an operation, a standardized reporting system for complications is required. Martin et al. established 10 criteria to improve the quality of reporting complications [7]. Recent studies examining cystectomy complications using a Clavien grading system have reported complication rates ranging between 48% and 74% [2], [3], [8], [9], and [10].

As only limited randomized comparisons of robotic and open cystectomy currently exist [11], objective reporting of complications based on an established system becomes increasingly important. We sought to define the type, incidence, and severity of early postoperative morbidity in a large RARC series using a standardized reporting methodology according to the Martin criteria.

2. Patients and methods

Between 2003 and 2011, 241 patients underwent RARC at City of Hope National Cancer Center. Forty-five patients were excluded because they declined participation in our bladder cancer (BCa) database or had concomitant surgery. The remaining 196 patients were consented to our institutional review board–approved cystectomy database. Indications for RARC included muscle-invasive BCa or high-grade refractory or recurrent multifocal bladder tumors not amenable to local therapy. Patients were offered neoadjuvant chemotherapy when indicated and underwent preoperative staging with computed tomography. Patients were offered orthotopic or cutaneous continent diversion or ileal conduit. Contraindications to orthotopic diversion included serum creatinine >2 ng/ml, disease at the urethral margin, or advanced disease. Contraindications to continent cutaneous diversion included poor manual dexterity, creatinine >2 ng/ml, significant comorbidity, symptoms of dementia, or advanced disease. Continent diversions were performed in 68% of cases.

Clinicopathologic and complications data were collected prospectively. In addition, independent review of all inpatient charts, outpatient notes, outside hospital notes, and correspondence with local physicians for complications was performed by academic urologists. All complications within 90 d of surgery were recorded and graded according to an established five-grade Clavien classification [12]. Clavien grade 1 or 2 complications were categorized as minor complications, and grade 3–5 complications were considered major. Preoperative comorbidities were computed using an age-adjusted Charlson comorbidity index (CCI) [13].

Presurgical patient preparation included clear liquids and GoLytely or magnesium citrate the day before surgery. Beginning in 2009, oral alvimopan was administered if there were no contraindications. Sequential compression devices were routinely used, and a second-generation cephalosporin was administered intravenously for 1 d perioperatively. In 2006, initiation of subcutaneous low-molecular-weight heparin began postoperatively with confirmation of stable hematocrit, and patients were also discharged home with a 4-wk supply.

3. Results

Patient clinical and pathologic characteristics are summarized in Table 1. Of 196 patients, 156 patients (80%) experienced one complication or more within 90 d of surgery; 475 total complications were captured, for an average of three adverse events per patient; 365 complications (77%) occurred in ≤30 d, with the remaining 110 complications (23%) occurring between 31 and 90 d postoperatively.

Complications were grouped and analyzed in a systems-based manner (Fig. 1 and Table 2). Most common were hematologic complications (20.4%), followed by infectious complications (16.2%), gastrointestinal complications (14.1%), and procedural complications (10.3%). Anemia was the most common hematologic disorder, though 98% of events were classified as minor. Minor urinary tract infections represented 42% of the infectious complications. Gastrointestinal complications mostly consisted of ileus or small bowel obstruction (67%), with the majority (84%) being minor. Ureteral or urethral anastomotic leaks made up 71% of the procedural complications captured.

Fig. 1 Breakdown of all grades and major complications by system.

Sixty-eight patients (35%) experienced a major complication during the first 90 d after surgery. Ureteral stricture occurred in 14 patients (7%), requiring percutaneous nephrostomy drainage; 38% of all infectious complications were major. The incidence of complications for patients who underwent RARC from 2003 to 2007 (n = 67) was higher than for patients who underwent RARC in 2008 and later (n = 129) (88.1% compared with 75.2%, p = 0.03). However, no significant change was seen in the incidence of major complications (35.8% compared with 34.1%).

Four patients (2%) died in ≤30 d, and eight patients (4.1%) died ≤90 d after surgery; two patients each died of sepsis, respiratory insufficiency, and BCa progression; one patient died of cardiac arrest; and one patient died of unknown causes. Patients with early mortality were older and had higher CCIs (p≤ 0.03).

In univariable analysis, American Society of Anesthesiologists (ASA) score, preoperative hematocrit ≤35%, estimated blood loss (EBL), operative time, and being a candidate for continent diversion were predictive of complications of any grade within 90 d of surgery (Table 3). In multivariable analysis looking at only presurgical indicators, ASA score, preoperative hematocrit, and being a candidate for continent diversion were significant predictors of any complications. In multivariable analysis of all presurgical and postsurgical indicators, age, ASA score, preoperative hematocrit, EBL, and operative time remained predictors of complications of any grade.

For prediction of major complications (Table 4), preoperative hematocrit ≤35% and EBL were significant on univariable analysis. In multivariable analysis of presurgical and all indicators, CCI, preoperative hematocrit, and orthotopic neobladder diversion were significant predictors of major complications.

Stratifying complications by diversion type, 45 of 62 patients with an ileal conduit (73%), 44 of 48 patients with a continent cutaneous pouch (92%), and 67 of 86 patients with an orthotopic neobladder (78%) experienced a complication. The respective rates for major complications were 17 of 62 patients (27%), 15 of 48 patients (31%), and 36 of 86 patients (42%). Patients with ileal conduits were older and had higher CCI, higher ASA score, and lower body mass index (BMI) (all p≤ 0.03) (Table 5 and Table 6). Patients receiving an ileal conduit had a worse pathologic stage than other patients; 50% of these patients were found to be ≥pT3 (p < 0.01).

Forty patients (20%) had a preoperative hematocrit <35%. Of 196 patients, 86 (43.9%) required a blood transfusion perioperatively. Patients requiring blood transfusion had lower preoperative hematocrits (36.9 compared with 42.1), lower BMI, higher preoperative serum creatinine, higher CCI or ASA score, and female gender (all p≤ 0.03). Patients with higher pathologic stage required transfusion more frequently (p = 0.02). Transfusions were significantly associated with an increased LOS (9 d compared with 11.5 d, p < 0.01). Patients with increased age, higher CCI, or higher pathologic stages also had a longer LOS (p≤ 0.03). There was a trend toward longer LOS in the continent cutaneous and ileal conduit groups (p = 0.06).

4. Discussion

Radical cystectomy remains the gold standard surgical therapy for locally aggressive urothelial carcinoma. Awareness of the significant associated morbidity is important for patients and clinicians to establish realistic expectations. Complications are related to the technical complexity of the operation, as well as the increased age of patients, preexisting comorbidities, and perioperative metabolic and fluid shifts. We report a prospective comprehensive complication analysis of a consecutive series of RARCs. All complications within 90 d were recorded and graded according to the Clavien system by urologic oncologists. Our series demonstrates that RARC can be performed with oncologic prudence in regard to surgical margin status and lymph node yields, two important factors related to recurrence and long-term survival [14] and [15]. The positive margin rate (4.1%) was low despite 36% of patients with extravesical disease, and the median lymph node yield (28) was consistent with other academic centers [16] and representative of an oncologic template dissection.

Most studies evaluating morbidity of radical cystectomy do not use a formal grading system, account for comorbidities, or clearly define various complications [17]. These factors likely lead to an underestimation of the true morbidity for the procedure. To standardize reporting, complication grading systems [12] and criteria for reporting surgical outcomes have been adopted [7].

We found that 80% of patients experienced an adverse event in the first 90 d, with the majority of these complication (76%) being minor. Most complications (77%) occurred ≤30 d after surgery. Mortality at 90 d was 4.1%. These results are comparable to recently published series that used a similar reporting methodology [2], [3], and [8]. Smaller series reviewing initial perioperative outcomes of RARC report complication rates ranging from 34% to 63% [3], [4], and [5]. Shabsigh et al. used a standardized reporting methodology in a series of 1142 open radical cystectomy patients [9]. At 90 d postoperatively, 64% of patients experienced a complication; 13% of the complications were major, the mortality rate was 2.7%, and the blood transfusion rate was 66%. In a series of 156 RARC patients, Hayn et al. identified 52% of patients having experienced a complication within 90 d of surgery [5]. In contrast to these studies, our study defined any blood transfusion as a complication, which could explain our increased complication rate. We also recorded a higher major complication rate, though direct comparison is challenging, as ileal conduit was performed in 63% [9] and 93% [5] of patients in these respective studies. The majority of our major complications (62%) were secondary to urinary diversion and often warranted higher Clavien classification because of procedural intervention (percutaneous drain or nephrostomy).

Transfusion rates for contemporary open cystectomy series range between 38% and 66% [9] and [18]. Our perioperative transfusion rate was 44%. Although EBL at the time of RARC is reduced, many patients ultimately require a transfusion. Possible reasons include patient comorbidity, frequent baseline anemia, and suppression of intraoperative EBL through pneumoperitoneum. In our study, transfused patients were older and had more comorbidities, higher pathologic stage, and lower preoperative hematocrit compared with nontransfused patients.

Median operative time of 432 min is longer than many open series and could affect postoperative morbidity. The time difference may be explained by the learning curve and the fact that the majority of patients underwent a continent diversion that took 80–120 min longer, on average (p < 0.01). Our median LOS of 9 d is similar to most published open series, which range from 6 to 11 d [9], [18], [19], and [20]. Our 23% incidence of ileus or small bowel obstruction is also similar to the incidence in open series [9] and [18].

Most surgical complications and major complications were urinary diversion related. This finding may be partly explained by the high percentage of continent diversions performed. Other reporters have given similar complication rates when comparing continent and conduit diversions [21] and [22]; however, no randomized controlled studies exist, and historical results may be influenced by a selection bias that favored continent diversion in healthier patients. In our study, patients with continent diversions were younger and healthier, and they had higher baseline hematocrit and lower pathologic stage. We believe that complex urinary diversions lend themselves to a slightly higher rate of diversion-related complications. Another reason may be that the urinary diversion is created through a minilaparotomy and then placed back inside the abdomen. Inadvertent kinking or twisting of the ureters and/or diversion may occur, leading to the aforementioned complications.

Multivariable analysis showed that age, ASA score, preoperative anemia, operative time, and EBL were statistically significant predictors of complications of any grade. However, when predicting major complications, only preoperative anemia, increased comorbidity, and orthotopic neobladder diversion were significant. Age [9] and [23], ASA score [8] and [10], and EBL [3] have previously been identified as significant predictors of complications. Other reports have also demonstrated that blood transfusion increases postoperative morbidity, specifically infectious complications [24] and [25]. The underlying reason is debated. Some evidence suggests that transfusion leads to direct immunosuppression [26]. Other evidence suggests that transfusion merely identifies individuals with greater underlying comorbidities, who are therefore at increased risk for complications. The predisposition for major complications in orthotopic diversion is possibly related to procedural intervention for ureteral and urethral neobladder anastomotic leaks.

The main limitations of our study include lack of randomization and smaller sample size compared with large open series [27]. Although to our knowledge this study represents the largest single-institution analysis of RARC, the statistical results may still be underpowered. Additionally, while the Clavien system represents an objective and graded method of measurement, the system remains intervention-based, and reproducibility biases may exist. While our data were prospectively collected, analysis was performed retrospectively. Finally, our patient population represents the experience of a tertiary cancer center and may not coincide with the experience of community hospitals. Although complication rates are high, we did demonstrate a decrease in incidence of total complications over time. As robotic technology improves, so will the ability to safely and efficiently perform intracorporeal diversions that may help lower complication rates. Also, with randomized comparative studies, we may better understand which patients are best suited for RARC.

5. Conclusions

When a standardized reporting methodology is used, RARC is associated with a high complication rate, though this rate is comparable to that of contemporary open series that used similar reporting guidelines. Our study reinforces the need for accurate and standardized reporting when scrutinizing surgical techniques and comparing published series.

Author contributions: Bertram E. Yuh had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Yuh, Wilson, Jankowski, Lau, Chan.

Acquisition of data: Yuh, Wilson, Nazmy, Jankowski, Menchaca, Torrey, Linehan, Lau, Chan.

Analysis and interpretation of data: Yuh, Wilson, Ruel.

Drafting of the manuscript: Yuh, Wilson, Nazmy, Jankowski.

Critical revision of the manuscript for important intellectual content: Yuh, Wilson, Nazmy, Jankowski, Torrey, Linehan, Lau, Chan.

Statistical analysis: Yuh, Wilson, Ruel.

Obtaining funding: None.

Administrative, technical, or material support: Menchaca, Ruel.

Supervision: Yuh, Wilson.

Other (specify): None.

Financial disclosures: Bertram E. Yuh certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.

Funding/Support and role of the sponsor: None.

Acknowledgment statement: The authors thank Nicola Solomon, PhD, for editorial assistance and critical review.