Defining Early Morbidity of Radical Cystectomy for Patients with Bladder Cancer Using a Standardized Reporting Methodology

Transcription

1 european urology 55 (2009) available at journal homepage: Bladder Cancer Defining Early Morbidity of Radical Cystectomy for Patients with Bladder Cancer Using a Standardized Reporting Methodology Ahmad Shabsigh a, Ruslan Korets a, Kinjal C. Vora a, Christine M. Brooks a, Angel M. Cronin b, Caroline Savage a, Ganesh Raj a, Bernard H. Bochner a, Guido Dalbagni a, Harry W. Herr a, S. Machele Donat a, * a Department of Urology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, United States b Departments of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, United States Article info Article history: Accepted July 2, 2008 Published online ahead of print on July 18, 2008 Keywords: Radical cystectomy Complications Reporting guidelines Abstract Background: Reporting methodology is highly variable and nonstandardized, yet surgical outcomes are utilized in clinical trial design and evaluation of healthcare provider performance. Objective: We sought to define the type, incidence, and severity of early postoperative morbidities following radical cystectomy (RC) using a standardized reporting methodology. Design, setting, and participants: Between 1995 and 2005, 1142 consecutive RCs were entered into a prospective complication database and retrospectively reviewed for accuracy. All patients underwent RC/urinary diversion by high-volume fellowshiptrained urologic oncologists. Measurements: All complications within 90 d of surgery were analyzed and graded according to the Memorial Sloan-Kettering Cancer Center complication grading system. Complications were defined and stratified into 11 specific categories. Univariate and multivariate regression models were used to define predictors of complications. Results and limitations: Sixty-four percent (735/1142) of patients experienced a complication within 90 d of surgery. Among patients experiencing a complication, 67% experienced a complication during the operative hospital admission and 58% following discharge. Overall, the highest grade of complication was grade 0 in 36% (n = 407), grade 1 2 in 51% (n = 582), and grade 3 5 in 13% (n = 153). Gastrointestinal complications were most common (29%), followed by infectious complications (25%) and wound-related complications (15%). The 30-d mortality rate was 1.5%. Conclusions: Surgical morbidity following RC is significant and, when strict reporting guidelines are incorporated, higher than previously published. Accurate reporting of postoperative complications after RC is essential for counseling patients, combined modality treatment planning, clinical trial design, and assessment of surgical success. # 2008 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Memorial Sloan-Kettering Cancer Center, Department of Surgery, Division of Urology, 1275 York Ave, New York, NY 10065, United States. Tel ; Fax: address: donats@mskcc.org (S.M. Donat) /$ see back matter # 2008 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 european urology 55 (2009) Introduction Muscle-invasive bladder cancer occurs predominantly in an elderly comorbid population with a reported mean age of 68 yr in contemporary radical cystectomy (RC) series [1]. The incidence of early complications, defined as occurring either during the hospitalization or within 30 d of surgery, has been reported in the range of 20 57% [2 10]. Therefore, accounting for the impact of surgical morbidity on patient outcome is essential for treatment planning, for clinical trial design, for assessing new surgical techniques, and for perioperative patient education. A recent evaluation of the urologic oncology literature revealed that the majority of series that reported on RC morbidity had not employed a formal complication reporting system, had not utilized grading systems other than to categorizing them into major versus minor, had not accounted for comorbidities, or had not defined complications. This makes it difficult to compare data and most certainly leads to an underestimation of morbidity for the procedure [2]. In addition, the incidence of perioperative complications have often been utilized as surrogate measures of surgical competency, institutional quality of care, success of new surgical techniques, and have even been suggested as benchmarks for financial reimbursement [3]. The National Cancer Institute developed the Common Toxicity Criteria for Adverse Events (CTCAE) to be utilized as a framework for reporting adverse events associated with clinical trials, and they recently adapted it to include surgical criteria [11]; however, it is not commonly utilized in the surgical literature [12 15]. More recently the National Surgical Quality Improvement Program (NSQIP), a comprehensive, validated, outcomebased, risk-adjusted, peer-controlled mechanism, is being utilized by some to evaluate urologic procedures including RC [6]. This system, while more effective than the traditional hospital morbidity reporting systems, has limitations, including the lack of procedure-specific variables, the lack of optional listing of additional procedures secondary to complications, and the lack of a complication severity grading system to assess the impact of complications on quality of life or outcomes. In 2002, Martin et al established a list of 10 critical elements that should be included when reporting surgical complications in order to provide a more accurate and comprehensive picture of surgical morbidity [15]; many of these are incorporated by the NSQIP program. We sought to better define the type, incidence, and severity of early postoperative morbidities following RC using these 10 critical reporting elements. 2. Methods Independent Review Board (IRB) approval was obtained to review 1320 consecutive RCs between January 1995 and October 2005 at Memorial Sloan-Kettering Cancer Center (MSKCC) and to enter data from them prospectively into an electronic hospital-based surgical morbidity database. A retrospective review of charts, outpatient notes, billing records, and correspondence with local physicians was performed and differences were reconciled. Of the 1320 patients, 14% were excluded due to lack of sufficient follow-up data, leaving 1142 patients for analysis. All complications within 90 d of surgery were recorded, defined, and graded according to an established five-grade modification of the original Clavien system [16] as shown in Table 1, and then they were further grouped into 11 categories as outlined in Table 2. The clinical and pathologic characteristics that were collected included the following: age, gender, pathologic stage, prior cancer treatments, prior surgeries, American Society of Anesthesiologists (ASA) score [17], Charlson- Romano comorbidity score [18], body mass index (BMI) [19], renal function (serum creatinine levels and creatinine clearance measured by the modification of diet in renal disease [MDRD] formula) [20], operative time, estimated blood loss (EBL), transfusion data, Intensive Care Unit (ICU) admission, reoperations, interventional radiology procedures, readmissions, emergency room visits, length of hospital stay (LOS), and vital status, including date of death Statistical methods Univariate and multivariable logistic regression analyses were used to evaluate variables associated with experiencing a complication, with separate analyses conducted for the outcome of high-grade (grade 3 5) complications. Variables analyzed were as follows: gender, age at surgery, BMI (continuous and categorized as >30 kg/m 2 and 30 kg/m 2 ), prior abdominal surgery, preoperative chemotherapy, preoperative radiotherapy, operative time, EBL, total units (>4 vs 4) of packed red blood cells (PRBCs) received, total units of fresh frozen plasma (FFPs) received (>4 vs4), type of urinary diversion (continent vs conduit), Charlson-Romano score ( 2 vs <2), ASA score ( 2 vs >2), preoperative creatinine levels Table 1 Postoperative complication grading system Grade Grade 0 Grade 1 Grade2 Grade 3 Grade 4 Grade 5 Definition No event observed Use of oral medications or bedside intervention Use of intravenous medications, total parenteral nutrition (TPN), enteral nutrition, or blood transfusion Interventional radiology, therapeutic endoscopy, intubation, angiography, or operation Residual and lasting disability requiring major rehabilitation or organ resection Death of patient

3 166 european urology 55 (2009) Table 2 Summary of complication types and categories Category (% of Total * ) Complication Frequency y Gastrointestinal (29%; n = 335) Ileus ** 183 SBO 82 Constipation *** 30 Clostridium difficile colitis 28 Gastrointestinal bleeding 15 Emesis 5 Anastomotic bowel leak 10 Diarrhea 10 Infectious (25%; n = 282) FUO 55 Abscess 49 UTI 113 Sepsis 51 Urosepsis 25 Pyelonephritis 29 Diverticulitis 2 Gastroenteritis 4 Cholecystitis 3 Wound (15%; n = 168) Wound seroma 5 Wound infection 106 Wound dehiscence 53 Facial dehiscence/evisceration 4 Genitourinary (11%; n = 120) Renal failure 32 Ureteral obstruction/rut 45 Urinary leak 30 Urinary fistula 3 Urinary retention 9 Parastomal hernia 4 Stomal ischemia 1 Hematuria 4 Cardiac (11%; n = 131) Arrhythmia 82 Myocardial infarction 15 Hypertension 12 Congestive heart failure 6 Angina 2 Hypotension 10 Pulmonary (9%; n = 98) Atelectasis 7 Pneumonia 45 Respiratory distress 38 Pneumothorax 1 Pleural effusion 15 Bleeding (9%; n = 104) Anemia requiring transfusion 92 Postoperative bleed other than GI 2 Wound hematoma 3 Thromboembolic (8%; n = 93) Deep venous thrombosis 60 Pulmonary embolism 36 Superficial phlebitis 2 Neurological (5%; n = 56) Peripheral neuropathy 18 CVA/TIA 6 Delirium/Agitation 21 Vertigo 1 Loss of consciousness 3 Seizure 1

4 european urology 55 (2009) Table 2 (Continued ) Category (% of Total * ) Complication Frequency y Miscellaneous (3%; n = 38) Psychological illness 6 Tendonitis 1 Dermatitis 3 Acidosis 4 Thrombocytopenia 3 Decubitus ulcer 6 Lymphocele 15 Peripheral arterial ischemia 3 Dehydration 14 Other rare complications 105 Surgical (1%; n = 8) Vascular injury 2 Bowel injury 2 Incisional hernia 3 Retained foreign body 1 n, the total number of patients within that category; SBO, small bowel obstruction (defined as clinical and radiographic findings of small bowel obstruction requiring intervention); FUO, fever of unknown origin; UTI, urinary tract infection; RUT, refluxing upper tract; GI, gastrointestinal; CVA, cerebral vascular accident; TIA, transient ischemic attack. * Reflects the percentage of patients with one or more complications within the category. ** Ileus is defined as post operative nausea or vomiting associated with abdominal distension requiring cessation of oral intake and intravenous fluid support and/or nasogastric tube (NGT) placement, or the intolerance of oral intake by postoperative day 5 resulting in patient fasting with or without NGT placement or antiemetic medication administration. *** Constipation is defined as inability to have a bowel movement by postoperative day 5 with no signs of ileus or small bowel obstruction. y Patients experiencing multiple complications of the same type are counted more than once. (<1.4 mg/dl vs 1.4 mg/dl), and organ-confined disease. All of these variables were included in the multivariable analysis with the exception of the total units of PRBCs and FFPs received, because they highly correlate with EBL. Although Charlson-Romano and ASA scores are both indicators of comorbid conditions; in our cohort they were not significantly collinear (for example, only 57% of patients with Charlson- Romano score 2 had an ASA score >2); both variables were therefore included in multivariable analyses. BMI was entered as continuous in the multivariable model. Statistical analyses were conducted using Stata 9.0 (StataCorp, College Station, TX, USA). 3. Results 3.1. Patient characteristics The study population was relatively comorbid (Table 3) with a median age of 68 yr (range: Table 3 Patients clinical and pathological characteristics Characteristic Median (IQR) or Frequency (%) Age 68 (60, 75) Gender (male) 862 (75%) Race (Caucasian) (n = 1141) 1035 (91%) Median BMI (n = 1141) 27.1 (24.3, 30.1) Very obese (BMI 30; n = 1141) 297 (26%) Overweight (BMI 25; n = 1141) 781 (68%) Prior abdominal surgery (n = 1141) 539 (47%) Prior nephroureterectomy 37 (3%) Prior hysterectomy/oophorectomy 81 (29%) Charlson-Romano score 2 (n = 1082) 329 (30%) ASA 3 4 (n = 1139) 487 (43%) Previous systemic chemotherapy 132 (12%) Previous pelvic radiotherapy 40 (4%) Serum creatinine level <1.4 mg/dl (n = 1134) 839 (74%) Hydronephrosis 206 (18%) Diversion (n = 1140) Continent 418 (37%) Ileal conduit 724 (63%) Lymph node dissection 1075 (94%) Organ-confined disease 657 (58%) IQR, interquartile range; BMI, body mass index; ASA, American Society of Anesthesiologists score.

5 168 european urology 55 (2009) yr), an age-adjusted Charlson-Romano morbidity index of 2 in 64% of patients (694/1082) and 4 in 24% of patients (263/1082), an ASA score of 3 in 43% of patients, and a BMI >25 in 68% of patients and >30 in 26% of patients Operative characteristics Ileal conduits were performed in 63% of patients; orthotopic neobladder diversions were performed in 28% of patients, and continent cutaneous diversions were performed in 8% of patients. Four patients on hemodialysis did not undergo urinary diversions. The median operative time was 6.4 h (interquartile range [IQR]: 5.3, 7.7), 5.8 h for ileal conduits and 7.2 h for continent diversions. Median EBL was 1000 ml (IQR: 700, 1500). Perioperative blood product transfusions were given in 66% of patients (750/1142), including PRBCs, FFPs, and/or platelets. Of these, 55% (401/750) received 1 2 PRBCs, with only 9% (107/ 750) requiring >4 PRBCs. Intraoperative surgical complications were rare (0.7%; Table 2) Postoperative care Prior to implementation of care pathways in 2000, all patients were admitted h preoperatively. Overall, 69% of the cohort were admitted a range of 1 13 d (median, 1 d) preoperatively. ICU admission was required in 5% of patients (58/1142). During the initial admission, 2% of patients (25/1142) required reoperation related to a surgical complication, and another 1% of patients (13/1142) required reoperation related to a surgical complication following discharge. Interventional radiology procedures were required in 11% of the cohort (127/1142), 47% of these (60/127) were postdischarge. The overall median LOS was 9 d (IQR: 8, 12), increasing to 11 d (IQR: 9, 15) in those experiencing any grade complication and increasing to 11 d (IQR: 11, 27) in those experiencing high-grade (grade 3 5) complications. Patients with higher comorbidity indices had a significantly longer LOS (Spearman correlation 0.15, p < ), with the median LOS of 10 d (IQR: 8, 14) in those with a Charlson-Romano index 2, compared to 9 d (IQR 8, 12) in those with a Charlson-Romano index of <2. Emergency room visits were required in 34% of patients (382/1142), 78% (298/382) of whom required readmission to the hospital for an overall readmission rate of 26% (298/ 1142). The majority of the readmissions were for gastrointestinal complications (nausea, emesis, dehydration, diarrhea, small bowel obstruction/ ileus), infectious complications (urosepsis, deep abscess, peritonitis), wound-related complications (significant cellulitis, dehiscence), or genitourinary complications (acute renal failure, urinary leak, fistula, hydronephrosis). The inpatient mortality rate was 0.9% (10/1142). The 90-d postoperative mortality rate was 2.7% (31/ 1142), with the majority (71%) of these deaths occurring after discharge from the hospital. Fiftyfive percent of these patients (17/31) died within 1 mo of surgery, 23% (7/31) died between 30 d and 60 d after surgery, and 23% (7/31) died between 60 d and 90 d after surgery. Only one patient died of progression of disease within 90 d of surgery. Twenty-one patients (68%) died from a cardiopulmonary event, 2 from septic shock, 1 from a stroke, and 7 from unknown causes Analysis of complications A total of 1637 complications occurred in 735 patients. Overall, 64% of patients (735/1142) experienced one or more complications within 90 d of surgery. Of those experiencing a complication, 67% (493/735) experienced complication(s) during the initial hospitalization, and 58% (428/735) experienced complication(s) following discharge. Sixtyeight different types of postoperative complications were recorded and grouped into 11 categories (Table 2). As outlined in Table 4, of the 735 patients experiencing complications, 79% (582) had only minor (grade 1 2) complications. Major (grade 3 5) complications occurred in only 13% of the cohort (153/1142). Most patients (58%) experienced one or more complications within 30 d of surgery, 12% of patients experienced one or more complications between 30 d and 60 d, and 6% of patients Table 4 Summary of complications by the highest grade experienced by each patient Highest grade of complication Overall 0 30 d d d (36%) 478 (42%) 1005 (88%) 1073 (94%) (11%) 117 (10%) 40 (4%) 15 (1%) (40%) 430 (38%) 69 (6%) 39 (3%) (12%) 98 (9%) 26 (2%) 15 (1%) 4 2 (0.2%) 2 (0.2%) 0 (0%) 0 (0%) 5 19 (2%) 17 (2%) 2 (0.2%) 0 (0%)

6 european urology 55 (2009) Fig. 1 Number of major and minor radical cystectomy (RC) complications recorded relative to the number of RCs performed with the percentage of total complications by year. experienced one or more complications between 60 d and 90 d of surgery. Although complications declined in both number and severity over time, 2.5% of patients (26/1142) still suffered a grade 3 5 complication at d, and 1.3% of patients (15/ 1142) still suffered a grade 3 5 complication at d(table 4). Fig. 1 demonstrates the variation in the number of perioperative complications by year of surgery relative to the number of RCs. The most common complication categories were gastrointestinal (29%), infectious (25%), woundrelated (15%), cardiac (11%), and genitourinary (11%), as shown in Table 2. The pattern of complications changed over time, with gastrointestinal, Table 5 Summary of the most common complication experienced: (a) all grade complications; (b) Grade 3 5 complications * (a) Rank by frequency of complication Overall (n = 199) 0 30 d (n = 149) d (n = 33) d (n = 17) Category % Patients Category % Patients Category % Patients Category % Patients 1 GU (28%) GU (24%) GU (42%) GU yyy (36%) 2 Infectious (23%) Infectious (20%) Infectious (30%) Infectious yyy (36%) 3 GI y (9%) Pulmonary (11%) GI (12%) GI (18%) 4 Cardiac y (9%) Cardiac (10%) Cardiac (6%) Surgical yyyy (6%) 5 Pulmonary y (9%) DVT/PE (9%) 3% yy 3% yy DVT/PE yyyy (6%) (b) Rank by frequency of complication Overall (n = 1637) 0 30 d (n = 1316) d (n = 221) d (n = 100) Category % Patients Category % Patients Category % Patients Category % Patients 1 GI (24%) GI (26%) Infectious (33%) Infectious (37%) 2 Infectious (21%) Infectious (17%) GI (17%) GI (20%) 3 Wound (12%) Wound (12%) GU (14%) GU (16%) 4 Cardiac (9%) Cardiac (10%) Wound (9%) Wound (8%) 5 GU (9%) Pulmonary (7%) DVT/PE (7%) Bleeding (5%) n, number of complications; GU, gentourinary; GI, gastrointestinal; DVT, deep venous thrombosis; PE, pulmonary embolism. * Patients who experienced multiple complications of the same category are counted more than once. y GI, cardiac, and pulmonary each had 17 complications. yy Wound, neurologic, and miscellaneous each had one complication. yyy GU and infection each had six complications. yyyy Surgical and thromboembolic each had one complication.

7 170 european urology 55 (2009) infectious, and wound-related complications (in declining order) being the most common in the first 30 d and infectious, gastrointestinal, and genitourinary complications accounting for the majority of complications between 30 d and 90 d postoperatively (Table 5A). When only high-grade complications were considered (Table 5B), genitourinary complications were the most prominent in all three time periods. Table 2 demonstrates the frequency of individual complications within each category. Ileus and/or a small bowel obstruction were the most common complication, occurring in 23% of patients. Anastomotic bowel leakage was rare, occurring in only 0.9% of patients, and intra-abdominal or pelvic abscesses occurred in 4% of patients. Urinary infections and/or urosepsis occurred in 14% of patients. Woundrelated complications occurred in 15% of patients, and culture-positive wound infections occurred in 9.3% of patients. Superficial dehiscence occurred in 4.6% of patients managed with local wound care, and fascial dehiscence requiring reoperation occurred in 0.4% of patients. New acute renal failure or exacerbation of chronic renal deficiency occurred in 2.8% of patients, and 4% of patients had postoperative radiographic evidence of ureteral obstruction or hydronephrosis. Twenty-eight patients (2.6%) developed symptomatic urinary leakage, three of these resulted in a urinary fistula. The most common cardiac morbidity was arrhythmia, affecting 7.2% of patients. Myocardial infarction occurred in 1.3% of patients. The most common pulmonary complications were pneumonia (in 3.9% of patients) and respiratory distress (in 3.3% of patients) related to decreased respiratory effort, atelectasis, bronchospasm, or pulmonary edema. Deep venous thrombosis developed in 5.3% of patients, and pulmonary embolism developed in 3.2% of patients; however, 1.2% of patients died of an acute cardiopulmonary event (myocardial infarction/pulmonary embolism [MI/PE]) within 30 d of surgery Predictors of postoperative morbidity (Table 6) A univariate analysis to evaluate predictors of any grade (grade 1 5) complication identified gender ( p = 0.003), prior pelvic radiotherapy ( p = 0.04), EBL ( p = 0.04), number of PRBCs ( p < ) or FFPs ( p = 0.005) transfused, and ASA score ( p = 0.03) as significant. Using multivariate analysis for any grade complication, gender ( p = 0.002), ASA score ( p = 0.023), and type of urinary diversion ( p = 0.015) were significant predictors, while prior pelvic radiotherapy ( p = 0.06) and age ( p = 0.08) trended toward significance. Regarding high-grade (grade 3 5) complications only, univariate analysis also identified EBL ( p = 0.013), number of PRBCs ( p < ) or FFPs ( p < ) transfused as well as age ( p = 0.02) as significant predictors, while ASA score was marginally predictive ( p = 0.053). Using multivariable analysis, age ( p = 0.04), prior abdominal surgery ( p = 0.03), and EBL ( p = 0.04) were significant predictors, while ASA score ( p = 0.053) and prior chemotherapy ( p = 0.055) trended toward significance. 4. Discussion Postoperative complication rates are often used as surrogates to indicate surgical competency and quality of care, and they have even been suggested as benchmarks for financial reimbursement, emphasizing the need for reliability of the reporting process for postoperative complications. Although recommendations for standardized reporting methodology exists [15], it is not routinely employed for reporting surgical complications in the urologic oncology literature [12], making it impossible to reliably compare the outcomes among different institutions, surgeons, or surgical techniques. An ideal methodology for reporting adverse events related to surgical therapy should include 10 established basic reporting criteria [12,15,21]: (1) a clear description of the method of data acquisition, (2) an indication of the duration of follow-up, (3) an indication of whether or not outpatient complication data are included, (4) definitions/inclusion criteria of at least one complication, (5) an indication of the mortality rate and cause of death, (6) an indication of the morbidity rate (number of patients and the total number of complications recorded), (7) an indication of procedure-specific complications, (8) the utilization of a grading system to clarify severity of complications, (9) an indication of the median or mean LOS, and (10) an indication of the methodology utilized to assess patient risk stratification (eg, Charlson-Romano index, ASA scoring). In this study, the overall complication rate was 64%, which is high in comparison with other studies [1 4,8,21 23]. Although this disparity in complication rates among series may be related to differences in surgeon experience, technique, and type of institution, it is more likely due to discrepancies in comorbidities among patient populations [24] and differences in the methodologies of collecting and reporting data, since these series all occur in academic institutions with experienced surgeons [7,12,25,26].

8 european urology 55 (2009) Table 6 (a) Univariate and (b) multivariate logistic regression analyses used to evaluate variables associated with any complication and complications grade 3 or higher Variables analyzed Outcome Any complication grade 1 5 n = 735 events Complication grade 3 5 n = 153 events Odds ratio (95% CI) p value Odds ratio (95% CI) p value (a) Gender (male vs female) 0.64 (0.48, 0.86) (0.59, 1.28) 0.5 Age at surgery (5 yr) 1.04 (0.98, 1.10) (1.02, 1.20) 0.02 BMI (kg/m 2 ) 1.01 (0.99, 1.03) (0.98, 1.04) 0.6 Very obese (BMI >30 kg/m 2 ; yes/no) 1.15 (0.87, 1.53) (0.88, 1.86) 0.2 Prior abdominal surgery (yes/no) 1.12 (0.88, 1.43) (0.59, 1.18) 0.3 Preoperative chemotherapy (yes/no) 1.17 (0.79, 1.72) (0.37, 1.24) 0.2 Preoperative radiotherapy (yes/no) 2.27 (1.04, 4.97) (0.60, 3.20) 0.4 Operating Room Time (100 min) 1.01 (0.90, 1.12) (0.96, 1.28) 0.17 Estimated Blood Loss (500 ml) 1.09 (1.00, 1.18) (1.02, 1.23) Total PRBCs received (>4 vs4) 4.07 (2.29, 7.23) < (2.04, 5.08) < Total number of FFP received (>4 vs4) 17.9 (2.43, 131) (3.47,14.53) < Type of urinary diversion (continent vs conduit) 1.16 (0.90, 1.49) (0.66, 1.34) 0.7 ASA (3 4 vs 1 2) 1.32 (1.03, 1.69) (1.00, 1.97) Charlson-Romano score (2vs <2) 1.03 (0.78, 1.35) (0.85, 1.77) 0.3 Abnormal preoperative creatinine level (>1.4 mg/dl; yes/no) 0.88 (0.67, 1.16) (0.90, 1.90) 0.15 Organ-confined disease (yes/no) 1.19 (0.93, 1.52) (0.71, 1.41) 1 (b) Gender (male vs female) 0.60 (0.44, 0.83) (0.52, 1.24) 0.3 Age at surgery (5 yr) 1.07 (0.99, 1.15) (1.01, 1.24) 0.04 BMI (kg/m 2 ) 1.02 (0.99, 1.05) (0.97, 1.04) 0.6 Prior abdominal surgery (yes/no) 0.93 (0.71, 1.21) (0.46, 0.97) 0.03 Preoperative chemotherapy (yes/no) 1.12 (0.74, 1.72) (0.25, 1.01) Preoperative radiotherapy (yes/no) 2.27 (0.96, 5.35) (0.61, 4.02) 0.3 Operating room time (100 min) 0.96 (0.84, 1.09) (0.96, 1.35) 0.13 Estimated blood loss (500 ml) 1.08 (0.98, 1.18) (1.00, 1.23) 0.04 Type of urinary diversion (continent vs conduit) 1.50 (1.08, 2.08) (0.70, 1.73) 0.7 ASA (3 4 vs 1 2) 1.38 (1.04, 1.82) (0.99, 2.14) Charlson Score (2 vs <2) 0.91 (0.68, 1.22) (0.67, 1.48) 1 Abnormal preoperative creatinine (>1.4; yes/no) 0.87 (0.64, 1.17) (0.74, 1.71) 0.6 Organ-confined disease (yes/no) 1.18 (0.91, 1.54) (0.72, 1.51) 0.8 BMI, body mass index; PRBC, pack of red blood cells; FFP, fresh frozen plasma pack; ASA, American Society of Anesthesiologists score. Each of the 10 reporting criteria may affect complication rates. For example, a commonly omitted reporting criterion is the indication of the duration of follow-up [10]. If this study had only reported inpatient complications, the overall complication rate would have been 43% (493/1142) as opposed to 58% at 30 d postoperatively, and 64% at 90 d postoperatively. Furthermore, if outpatient complication data had not been reported, the impact of urgent care visits (34% of patients), reoperation rates (1% of patients), angiointerventional procedures (5.3% of patients), and readmissions (26% of patients) would not be apparent. This is especially important with continent urinary diversion, where some complications are not apt to occur until 2 6 wk postoperatively (when drainage tubes are removed). In this cohort 12% of patients suffered a complication between 30 d and 60 d postoperatively, and 6% of patients suffered a complication between 60 d and 90 d postoperatively (Table 4). In addition, almost half (45%) of procedure-related deaths in this study occurred between 30 d and 90 d postoperatively and would not have been reported in the traditional 30-d time period, supporting consideration of a longer follow-up period to define the morbidity and mortality of RC [24]. For example, in this study the inpatient mortality rate was only 0.9%; however, the 30-d mortality rate was 1.5%, and the 90-d mortality rate was 2.7%. Providing definitions for at least the more common procedure-specific complications would allow better comparison among series and would improve the accuracy of the data when multiple individuals are recording data. For example, there are at least six different definitions of ileus in the urologic literature [12], which may account for

9 172 european urology 55 (2009) disparity in reported rates. We defined ileus as the inability to tolerate solid food by postoperative day 5, the need to place a nasogastric tube (NGT), or the need to stop oral intake due to abdominal distension, nausea, or emesis. Using this definition, our rate of postoperative ileus was 23%, which is twice as high as those reported in some RC series and is comparable with other studies where this rate ranges from 13% to 23.5% [2 4,8,21 23]. Utilization of a grading system identifies the severity of a complication using objective reproducible measures (Table 1) and requires complications to be defined in a specific manner as opposed to subjective definition of what constitutes a major complication by each author. If this study had reported only the major complications (grade 3 5), the overall complication rate would have been only 9.8% for inpatients and 10% at 30 d postoperatively, as opposed to a 43% complication rate for inpatients and a 58% complication rate 30 d postoperatively when all grades were reported. Whether minor (grade 1 2) complications are important to report is debatable; by definition (Table 1), they include open wounds receiving local care and events requiring total parenteral nutrition, transfusion, intravenous fluids, or medications to treat. Such complications may require readmission (26% in this series) and can affect the timing of planned adjuvant therapies. In addition, when determining the predictors for postsurgical complications (Table 6), we found that the severity or grade of the complication included in the analysis can affect the results. For example, a multivariable analysis of predictors for experiencing any grade (1 5) of complication found that gender, ASA score >2, and type of urinary diversion were predictive variables ( p < 0.05). However, a multivariable analysis of predictors for high-grade (grade 3 5) or major complications identified advancing age, prior abdominal/pelvic surgery, EBL, and ASA score >2 as significant predictors, while gender and urinary diversion were not identified as significant predictors. This could partly be explained by the reduced power associated with the smaller number of major complications (n = 153) versus any complication (n = 735). Nevertheless, this finding indicates that the distribution of the grade of complications (for example, one series having only high-grade complications and another having a majority of lowgrade complications) may account for the some of the disparity in predictors of complications among series. Defining the morbidity of a study cohort, using risk stratification methodology has been shown to significantly influence frequently utilized outcome measures such as perioperative mortality and LOS [24]. These measures in particular are often utilized to compare the morbidities of new surgical procedures, where patients with fewer comorbidities are often selected for the new procedure to try to limit complications, and this must be taken into account when making conclusions regarding the morbidity of new surgical techniques. In the present series, a Charlson-Romano index of 2 significantly ( p < ) affected LOS, a commonly used outcome to suggest that minimally invasive procedures are less morbid than open. Finally, evaluating complications over the 90 d postoperative period rather than the traditional inpatient or 30-d postoperative period, allows us to see the natural evolution of complications over time and has the potential of identifying modifiable factors that can be targeted for quality improvement [6]. Interestingly, gastrointestinal complications (ileus/intermittent small bowel obstruction) tend to persist as the second or third most common complication between 30 d and 90 d postoperatively. Although they tended to be low-grade (grade 1 2) in nature, they often required readmission. Both infections and genitourinary complications were common, notably constituting 30 42% of the highgrade complications postdischarge (Table 5). The limitations of this study are that we may have underreported complications treated outside MSKCC. Comparison with other RC series will be difficult if similar reporting guidelines were not utilized. And finally, the rate of complications experienced in a high-volume tertiary cancer center may not be transferable to the community setting where surgical experience and the patient s comorbid state and may not be comparable. 5. Conclusions This study demonstrates the high rate of postoperative morbidity following RC when a standard reporting methodology is utilized and the value of considering reporting methodology when comparing series. Furthermore, this study confirms prior observations [24] that have indicated the significant impact patient comorbidity may have on surgical outcomes such as LOS, emphasizing the importance of taking this into account when assessing new surgical techniques. Accurate reporting of complications is essential for preoperative counseling, for identifying modifiable risk factors to reduce complication rates, for planning combined modality treatments, for clinical trial design, and for a more accurate assessment of surgical success. It is our hope that our experience will encourage others to

10 european urology 55 (2009) begin utilizing the 10 reporting criteria and methodology described in this report when publishing their results so future meaningful comparisons among retrospective series may be possible in lieu of randomized trials. Author contributions: S. Machele Donat 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: Donat, Shabsigh. Acquisition of data: Donat, Shabsigh, Kurtas, Vora, Brook, Raj. Analysis and interpretation of data: Donat, Shabsigh, Savage, Cronin. Drafting of the manuscript: Donat, Shabsigh. Critical revision of the manuscript for important intellectual content: Donat, Shabsigh, Bochner, Dalbagni, Herr. Statistical analysis: Savage, Cronin. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: None. Other (specify): None. Financial disclosures: I certify 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. References [1] Konety BR, Allareddy V, Herr H. Complications after radical cystectomy: analysis of population-based data. Urology 2006;68: [2] Chang SS, Cookson MS, Baumgartner RG, Wells N, Smith Jr JA. Analysis of early complications after radical cystectomy: results of a collaborative care pathway. J Urol 2002;167: [3] Cookson MS, Chang SS, Wells N, Parekh DJ, Smith Jr JA. Complications of radical cystectomy for nonmuscle invasive disease: comparison with muscle invasive disease. J Urol 2003;169: [4] Frazier HA, Robertson JE, Paulson DF. Complications of radical cystectomy and urinary diversion: a retrospective review of 675 cases in 2 decades. J Urol 1992;148: [5] Ghoneim MA, el-mekresh MM, el-baz MA, el-attar IA, Ashamallah A. Radical cystectomy for carcinoma of the bladder: critical evaluation of the results in 1026 cases. J Urol 1997;158: [6] Hollenbeck BK, Miller DC, Taub D, et al. Identifying risk factors for potentially avoidable complications following radical cystectomy. J Urol 2005;174: [7] Joniau S, Benijts J, Van Kampen M, et al. Clinical experience with the N-shaped ileal neobladder: assessment of complications, voiding patterns, and quality of life in our series of 58 patients. Eur Urol 2005;47: [8] Meller AE, Nesrallah LJ, Dall Oglio MF, Srougi M. Complications in radical cystectomy performed at a teaching hospital. Int Braz J Urol 2002;28: [9] Meyer JP, Drake B, Boorer J, Gillatt D, Persad R, Fawcett D. A three-centre experience of orthotopic neobladder reconstruction after radical cystectomy: initial results. BJU Int 2004;94: [10] Quek ML, Stein JP, Daneshmand S, et al. A critical analysis of perioperative mortality from radical cystectomy. J Urol 2006;175: [11] Turna B, Frota R, Kamoi K, et al. Risk factor analysis of postoperative complications in laparoscopic partial nephrectomy. J Urol 2008;179: [12] Donat SM. Standards for surgical complication reporting in urologic oncology: time for a change. Urology 2007;69: [13] Callcut RA, Breslin TM. Shaping the future of surgery: the role of private regulation in determining quality standards. Ann Surg 2006;243: [14] Trotti A, Colevas AD, Setser A, et al. CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 2003;13: [15] Martin 2nd RC, Brennan MF, Jaques DP. Quality of complication reporting in the surgical literature. Ann Surg 2002;235: [16] Clavien PA, Sanabria JR, Strasberg SM. Proposed classification of complications of surgery with examples of utility in cholecystectomy. Surgery 1992;111: [17] Al-Homoud S, Purkayastha S, Aziz O, et al. Evaluating operative risk in colorectal cancer surgery: ASA and POSSUM-based predictive models. Surg Oncol 2004;13: [18] Koppie TM, Serio AM, Vickers AJ, et al. Age-adjusted Charlson comorbidity score is associated with treatment decisions and clinical outcomes for patients undergoing radical cystectomy for bladder cancer. Cancer 2008;112: [19] Hafron J, Mitra N, Dalbagni G, Bochner B, Herr H, Donat SM. Does body mass index affect survival of patients undergoing radical or partial cystectomy for bladder cancer? J Urol 2005;173: [20] Dash A, Galsky MD, Vickers AJ, et al. Impact of renal impairment on eligibility for adjuvant cisplatin-based chemotherapy in patients with urothelial carcinoma of the bladder. Cancer 2006;107: [21] Brannan W, Fuselier Jr HA, Ochsner M, Randrup ER. Critical evaluation of 1-stage cystectomy reducing morbidity and mortality. J Urol 1981;125: [22] Skinner DG, Crawford ED, Kaufman JJ. Complications of radical cystectomy for carcinoma of the bladder. J Urol 1980;123: [23] Novotny V, Hakenberg OW, Wiessner D, et al. Perioperative complications of radical cystectomy in a contemporary series. Eur Urol 2007;51:

11 174 european urology 55 (2009) [24] Hollenbeck BK, Miller DC, Taub DA, et al. The effects of adjusting for case mix on mortality and length of stay following radical cystectomy. J Urol 2006;176: [25] Jaques DP. Measuring morbidity. Ann Surg 2004;240: [26] Miller DC, Filson CP, Wallner LP, Montie JE, Campbell DA, Wei JT. Comparing performance of morbidity and mortality conference and national surgical quality improvement program for detection of complications after urologic surgery. Urology 2006;68: Editorial Comment on: Defining Early Morbidity of Radical Cystectomy for Patients with Bladder Cancer Using a Standardized Reporting Methodology Richard E. Hautmann Department of Urology, University of Ulm, Germany richard.hautmann@uniklinik-ulm.de This study [1] is an extraordinary, valuable contribution, and the authors should be congratulated. In a previous paper [2], the corresponding author of this paper [1] did a MEDLINE search and analyzed the quality of complication reporting in the urologic literature. The conclusion was that the disparity in the quality of surgical complication reporting in urologic oncology makes it impossible to compare the morbidity of surgical techniques and outcomes. Common measures for surgical outcomes include estimated blood loss, operative time, analgesic use, length of hospital stay, time to return to work, perioperative death and complication rates, and hospital costs; however, no definitions for complications or guidelines for reporting surgical outcomes have been universally accepted. This makes it impossible to compare the surgical outcomes among institutions, individual surgeons, or surgical techniques. Given comparable expertise with regard to indication, volume of cystectomy, professional staffing, and surgeon volume, surprisingly identical outcomes may be achieved [3]. The need for standardized reporting of complications is especially critical in urologic oncology because of the elderly population and their medical comorbidities. The success of multimodality therapies in the treatment of urologic cancers often relies on the timing of chemotherapy or radiotherapy with the surgical intervention and can be affected by perioperative complications. Minimally invasive techniques are gaining increased popularity in the belief that they will minimize the perioperative morbidity and thereby facilitate recovery. Because randomized trials comparing minimally invasive techniques to open surgery are difficult and lacking in urology, it seems imperative that we establish universally accepted criteria for reporting surgical morbidities and outcomes to establish the efficacy of surgical techniques and improve the quality of patient care. Radical cystectomy has been assessed the highest values in terms of difficulty of surgery for any procedure in urology. Radical cystectomy is also the most difficult robotic procedure and more so if the diversion is performed totally intracorporeally. Only the availability of a standardized reporting of complications can provide hard data on volume outcome relationships and can encourage patients facing certain urologic procedures to seek treatment at high-procedure-volume centers [4]. Radical cystectomy quality has a major impact on invasive bladder cancer survival and consequently plays a significant role in clinical trial design and trial outcome. Who performs the surgery and where and how well it is done matter. Negative surgical margins and 10 lymph nodes removed were associated with better overall survival independent of patient age, pathological stage, nodal status, and whether chemotherapy was given. This cooperative group trial shows that the quality of cystectomy and pelvic lymph node dissection directly affects the chances of survival, and it is surgeon-dependent [5]. This paper [1] should have a major impact on future cystectomy outcome studies. References [1] Shabsigh A, Korets R, Vora KC, et al. Defining early morbidity of radical cystectomy for patients with bladder cancer using a standardized reporting methodology. Eur Urol 2009;55: [2] Donat SM. Standards for surgical complication reporting in urologic oncology: time for a change. Urology 2007;69: [3] Hautmann RE, Volkmer BG, Schumacher MC, et al. Longterm results of standard procedures in urology: the ileal neobladder. World J Urol 2006;24: [4] Hollenbeck BK, Taub DA, Miller DC, et al. The regionalization of radical cystectomy to specific medical centers. J Urol 2005;174: [5] Herr HW, Faulkner JR, Grossmann HB, et al. Surgical factors influence bladder cancer outcomes: a cooperative group report. J Clin Oncol 2004;22: DOI: /j.eururo DOI of original article: /j.eururo

12 european urology 55 (2009) Editorial Comment on: Defining Early Morbidity of Radical Cystectomy for Patients with Bladder Cancer Using a Standardized Reporting Methodology Renzo Colombo Department of Urology, University Vita-Salute San Raffaele, Milan, Italy colombo.renzo@hsr.it Neoadjuvant and adjuvant chemotherapy settings are referred with consistently different rates of administration from different institutions in different countries [1 3]; however, there is a general agreement that perioperative chemotherapy remains infrequently used to date. Suboptimal design of trials based on the insufficient number of patients has limited the clinical application of chemotherapy as an adjuvant approach. In addition, although similar results (5 6% net benefit in overall survival) are presumed for adjuvant chemotherapy when compared with neoadjuvant chemotherapy, this finding is still definitively unproved [4]. Apart from the overall limited clinical outcomes, age, medical morbidity, severity of renal failure, and patient reluctance affect the extant confined role of adjuvant chemotherapy. Donat et al [5] investigated the potential impact of surgical complications related to radical cystectomy on the timing of adjuvant chemotherapy for stage III bladder cancer. This monoinstitutional experience, including numerous series of patients submitted to radical cystectomy over a period of 10 yr, certainly has many merits. The study specifically focused on the potential for surgical complications to affect the patient s ability to use an adjuvant chemotherapy approach. Few publications [6,7] have treated in detail the limiting effect of cystectomy-related morbidity on adjuvant therapy. To reach their target, the authors adopted a detailed and reproducible methodology based on a grade system (modified Clavien 5-grade system) that allowed for rigorous assessment of both category type and severity of surgical complications. Because adjuvant chemotherapy should be given within 90 d postoperatively, the authors adequately decided to extend the overall period of time for the postoperative evaluation of surgical complications up to 90 d. This represents an additional value when compared with the majority of previous studies in which surgical complications were generally described within d. According to the results of Donat et al s study [5], 30% of patients could have been excluded from a timing-adjuvant chemotherapy due only to grade 2 5 surgical complications; however, as clearly admitted, since the authors were not able to identify how many patients actually received chemotherapy, this study appears to be more speculative than of practical relevance. The key point is to know how many patients, mainly those with grade 2 complications (about 60%), may have recuperated in time to start chemotherapy or may have been able to complete treatment even though the start of therapy may have been delayed. We cannot, however, definitively assume that a delay of the prescribed timing can minimize or cancel the benefit of the adjuvant approach. Although not the final point of the study, an assessment of the reasons why so many patients did not receive adjuvant chemotherapy would have been of interest and provided useful information on the relative burden of surgical complications. In addition, when looking at such a high rate of high-risk patients, the results from this study at a high-volume, tertiary American center could be difficult to translate directly to the majority of urologic centers, particularly in Europe. Although the results of this study cannot definitively suggest preference for neoadjuvant therapy, they represent a practical example of what can be deduced when a standardized methodology for surgical complication reporting is adopted. This methodology, including five grades and 11 specific categories and used for the definition of type, incidence, and severity of morbidities following radical cystectomy, was extensively presented by the authors in a contemporary paper [8]. According to their experience, when using such a rigorous method of collection and evaluation, the overall rate (64%) of postoperative complications is consistently higher than is generally shown in the literature, and the majority of complications (51%) are of moderate severity (grades 1 2). Gastrointestinal complications are unquestionably the most frequent (29%) early on, and the highest grades of complications (grades 3 5) occur mainly within the first 30 d postoperatively. Multivariate analysis shows that American Society of Anesthesiologists (ASA) score and type of urinary diversion are significant predictors for any grade of postoperative complication. The need for a uniformly accepted definition of some surgical morbidity (ie, small bowel obstruction, ileus, constipation) and a standard classification of severity of complications has been universally advocated for many years. Although this risk stratification methodology comes from a tertiary referral institution managing a special