A contemporary standard for morbidity and outcome after radical cystectomy

Authors


Vijay A.C. Ramani, Department of Urology, The Christie NHS Foundation Trust, Wilmslow Road, Manchester, M204BX, UK.
e-mail: vijay.ramani@gmail.com

Abstract

OBJECTIVE

To report the temporal changes in peri-operative outcome over an extended period in patients undergoing radical cystectomy (RC) for all causes, irrespective of the previous treatment or pathology; and to establish a current standard of peri-operative outcome for RC by analysis of contemporary operative mortality rates (2000–5) factored for risk factors that might predict outcome.

PATIENTS AND METHODS

All patients undergoing RC between 1970 and 2005 were analysed; this was an unselected single-centre series and included patients previously treated by definitive radiotherapy, chemotherapy, and cases of RC where the primary tumour involved the bladder but was not of bladder origin.

RESULTS

In all, 846 patients had a RC, of whom 647 had a bladder primary tumour and 199 a primary tumour elsewhere (gynaecological, colorectal and others). There was a progressive reduction in 30- and 60-day mortality rates, such that the current peri-operative mortality (1999–2005) was 0.4% and 2.6%, respectively. There was a significant reduction in the re-operation rate over the decades (P = 0.01), which is currently 4.7%. Patient age was a significant factor in 30- and 60-day mortality rates (P < 0.001 for both) but there was no significant association between either American Society of Anesthesiologists grade or T stage with complication rates (P = 0.61 and 0.12, respectively).

CONCLUSION

There has been a progressive reduction in mortality related to RC, associated with both cases of RC and pelvic exenteration. The contemporary standard for 30-and 60-day mortality rates for these operations is 0.4% and 2.6%, respectively.

Abbreviations
RC

radical cystectomy

RT

radiotherapy

ASA

American Society of Anesthesiologists.

INTRODUCTION

Radical cystectomy (RC) is still considered the gold standard procedure for invasive bladder cancer [1]. Surgery of this magnitude was previously associated with high morbidity and mortality rates [2]. More recently there have been reports of large series of patients treated with RC that show a decline in both complications and mortality [3–6]. Most of these series have not included patients treated with radiotherapy (RT), or patients undergoing total exenteration, factors that might potentially increase the morbidity and mortality of such surgery [3,4,6].

Chahal et al.[5] reported a series of patients undergoing RC that included those treated with previous RT, but that series did not include patients whose primary histology was not of bladder origin. RC is often done in conjunction with the removal of other pelvic organs, yet most reported series seem to exclude non-bladder primary tumours from their series. Patient selection in these reports might lead to a bias in the outcomes reported. Morbidity and mortality rates should be an integral part of counselling and consenting patients before surgery, and so it is important to have a contemporary standard for all cases of RC.

Others have attempted to identify factors that contribute to the outcome of RC. Clinical tumour stage and grade remain the best predictors of cancer-specific survival in patients undergoing RC [7]. Other clinical variables, e.g. age, preoperative creatinine and haemoglobin levels, and pelvic irradiation, have also been shown to be associated with survival [7]. More recently, one study assessing factors affecting the peri-operative mortality of RC found that an American Society of Anesthesiologists (ASA) grade of ≥3, and significant intraoperative blood loss, were associated with higher mortality [8].

We set out to report changes in peri-operative outcomes, reported by decade, in patients undergoing radical pelvic surgery (1970–2005), including cases of salvage RC after RT, and pelvic exenterative surgery involving en-bloc bladder removal, where the primary tumour was not of bladder origin. From this we aimed to a report the temporal changes in outcome and to establish a contemporary standard for peri-operative outcome for surgery of this type. We also aimed to identify prognostic factors for peri-operative morbidity and mortality.

PATIENTS AND METHODS

Data were collected retrospectively from 1970 to 1998 and then prospectively from 1999. All patients, without exclusion, undergoing RC between 1970 and 2005 were analysed. Procedures were performed by the same surgical team at a single cancer centre encompassing three sites in the North-west of England; The Christie, University Hospital of South Manchester, and Salford Royal NHS Foundation Trusts. All operations were performed by subspecialized urological oncological surgeons based at these sites and in cases of total pelvic clearance, as part of a ‘pelvic team’ comprising urological, gynaecological and colorectal oncological surgeons.

Outcomes for all patients undergoing RC were analysed, including data from those who had received previous treatment with RT or chemotherapy. Those whose primary histology was not TCC of the bladder were also included, as was information from those undergoing pelvic exenteration, either anterior or total, where the primary histology was not of bladder origin. The rationale for including these patients was that the underlying surgical principles should be similar in all cases, and complication and mortality rates should be standardized for all pelvic surgery.

Variables recorded included patient demographics, ASA grade, tumour characteristics, adjuvant therapy, complication rates, and 30- and 60-day mortality rates. Complication rates were also divided into early (<30 days), and late (>30 days); these were recorded comprehensively for a wide range of variables.

In men, RC included the bladder, distal ureters, seminal vesicles and prostate. Male urethrectomy was only done when clinically indicated (prostatic and/or urethral involvement by tumour). In women, the urethra was excised routinely unless a neobladder was constructed. Ureteric stents were inserted intraoperatively and removed at 8–10 days after surgery. From 1994, all patients were routinely admitted to the hospital’s high-dependency unit for at least one night for more intensive monitoring and nursing care.

Patient groups were compared using a chi-square test, and prognostic factors for survival were determined using univariate and multivariate Cox proportional hazards regression.

RESULTS

In all, 846 unselected patients had RC from 1975 to 2005; the subcategories for tumour type are shown in Table 1. There were 536 men (63%) and 310 women (37%), with a median (range) age of 63 (23–87) years. In all, 744 patients had a RC or anterior exenteration, and 102 a total pelvic exenteration. There were 426 primary cases and 420 salvage cases after RT. In 647 of the 846 (76%) patients the primary malignancy was of bladder origin; of the remaining 199 patients, 121 (14%) had a gynaecological primary, 64 (8%) a colorectal primary and 14 (2%) had other primary tumours. Categorized by decade, there were 215 RCs in 1970–79, 183 in 1980–89, 214 in 1990–99 and 234 from 2000 to the end of 2005 (Table 2).

Table 1.  The site of primary tumour and median age of patients per site
Primaryn (%)Median age, years
Bladder TCC587 (69)64
Bladder (other) 60 (7)62
Cervix/uterus/vagina/vulva121 (14)56
Colon/rectum/anus 64 (8)61
Others 14 (2)59
Total846 (100)63
Table 2.  The number of procedures in each period divided into primary diagnoses
Primary1970–791980–891990–992000–5
Bladder (TCC)179125131152
Bladder (other) 13  5 25 17
Non-bladder 23 53 58 65
Total215183214234

The overall early complication rate, defined as at least one complication at <30 days after surgery, was 35.5% for all pelvic surgery; this compared with 39.5% for those with bladder TCC as the primary. There were 242 early medical complications and 157 early surgical complications. Over the entire series, 0.7% of patients had a bowel leak and 3.9% a urinary leak (2.1% requiring intervention). The most common medical complication was an infection of either the urinary tract or chest, occurring in 8.4% of patients, and prolonged ileus (defined as an ileus of >5 days) in 8%. Other medical and surgical complications are listed in Table 3.

Table 3.  The percentage of peri-operative complications (<30 days) divided ‘surgical’ and ‘medical’
Complications%
Surgical
Dehiscence4.3
Wound infection4.5
Early stoma problems4.3
Bowel leak0.7
Haemorrhage1.1
Urinary leak (intervention)3.9 (2.1)
Medical
Infections (UTI/chest)8.4
Ileus8
Deep vein thrombosis (pulmonary embolism)4.1 (2.8)
Cardiac (arrhymia/myocardial infarct/congestive cardiac failure)2.5
Respiratory failure2
Metabolic1

Most of the early complications were treated conservatively except for wound dehiscence and bowel leakage. The re-operation rate within the first 30 days decreased significantly from 1970–79 to the present, reducing progressively from 13.5%, 9.3%, and 7.5% to 4.7% each decade (P = 0.01; Fig. 1A). Specifically, the laparotomy rate required for a urinary or bowel leak reduced from 7% for 1970–79, to 6% for 1980–89, 1.9% for 1990–99 and 0.4% for 2000–2005 (Table 4).

Figure 1.

A, Re-operation rates (%) and B, 30- and 60-day mortality rates (%) for the defined periods (1970–79, 1980–89, 1990–99, and 2000–05).

Table 4.  The procedure (%) at re-operation over the decades
Procedure1970–791980–891990–992000–05
Laparotomy (bowel/urinary leak)761.90.4
Small bowel obstruction1.31.60.90
Refashion stoma1.40.500.8
Remove/insert pack0.503.20.8
Re-suture of incision4.21.11.42.6
Total13.59.37.54.7

Similarly, the late complication rate, defined as at least one complication at >30 days after surgery, was 25.5% overall, and 26.7% for those with a bladder TCC primary. The most common late complications included uretero-ileal stricture (1.8%), stoma-related, including prolapse, retraction and stenosis (2.4%), incisional hernia (5.9%), sinus or fistula formation (2.6%), abscess (1.6%) and interval bowel obstruction (1.4%). Of the 22 patients who developed a sinus or fistula, 14 occurred between 1970 and 1979 (6.5% of the group), compared with only one (0.4%) between 2000 and 2005. The early and late complication rate for the 102 total exenteration cases was 25.5% and 22.5%, respectively.

There was a progressive reduction in the 30- and 60-day mortality rates over the decades, from 13.5% and 15.8% in 1970–79, 4.4% and 8.2% in 1980–89, 3.7% and 5.1% in 1990–99 to the current standard mortality of 0.4% and 2.6%, respectively, in 2000–2005 (Fig. 1B). The 30- and 60-day mortality rates for the 102 patients who had total exenteration was 2.0% and 3.9%, respectively.

Risk factors for peri-operative outcomes were divided into static and dynamic, where static are predetermined (age, ASA grade, pathological stage, etc.) and dynamic are potentially modifiable (use of adjuvant RT/chemotherapy or surgical technique). When divided into decades, the overall 30- and 60-day mortality was 2.6% and 4.2% for patients aged <60 years, 5.6% and 7.7% for 60–69 years, and 10.5% and 14.0% for those aged >70 years. The 30- and 60-day mortality rates for these ages reduced over the study period. For patients aged <60 years the respective mortality rates for 1970–79 were 6.1% and 7.3%, vs 0% and 1.4% for 2000–2005. Similarly, for patients aged 60–70 years, the respective mortality rates reduced over time and are currently 0% and 2.2% for 2000–2005, and for those aged >70 years 1.5% and 4.4%, respectively. Age was a statistically significant risk factor (both P < 0.001) for the 30- and 60-day mortality, but not for complications (P = 0.09) in these three age groups.

ASA grade was a more recent addition and therefore representative of a modern group of 149 patients; 14 were ASA grade 1, with an overall complication rate of 46.2%, 102 were grade 2 with a complication rate of 39.4%, and 31 were grade 3, with a complication rate of 51.6%. There was no significant association between ASA grade and complication rate (chi-squared P = 0.61).

Overall, 266 patients had a pathological stage of ≤ T1, 134 stage T2, 79 stage T3 and 70 stage T4 disease. The complication rates for these groups were 57.9%, 52.2%, 49.3% and 47.5%, respectively. There was no significant relationship between pathological T stage and complication rates (P = 0.12). For those with stage ≤ pT1 the 30- and 60-day mortality rates were 5.2% and 7.1%, respectively. This compared with 2.2% and 5.2% for pT2, 8.3% and 11.1% for pT3 and 6.5% and 8.5% for pT4. There was no significant association between pathological stage and either mortality rate (P = 0.12 and 0.22, respectively).

In all, 420 patients had RT before surgery; there was no significant difference between the complication rates after surgery in these patients, at 36.9% after RT vs 34.0% with no previous RT (P = 0.40).

DISCUSSION

We report our experience of a large series of unselected patients undergoing RC at one UK cancer centre. Unlike many previous series [3–6], we incorporated all cases requiring RC, including anterior and total pelvic exenteration, considering cases with histology of both bladder and non-bladder origin. We also included patients undergoing salvage RC after RT. Stein et al.[3] reported selectively only 1054 of a total of 1471 consecutive patients undergoing RC, and similarly, Madersbacher et al.[6] reported on 507 patients from a total of 647. The rationale for including a nonselective approach including all consecutive cases in our series is that the surgery for these cases followed the same principles throughout, and although there might be subtle differences, the complication and 30- and 60-day mortality rates should be comparable irrespective of tumour origin, stage, or previous treatment.

Our results showed no significant relationship between outcome for the different subtypes undergoing RC in terms of complication rate, 30- and 60-day mortality, and T stage. The main contributing factor for this is likely to be the surgical and anaesthetic techniques. Patients are optimized medically before intervention, surgery is undertaken under close haemostatic control, with the pelvic dissection performed as a vascular operation, ‘skeletonising’ the iliacs, flush-ligating all pelvic vessels and avoiding large shifts in fluid volume. Patients are then invasively monitored and aggressively managed by an experienced and unified team, comprising surgeons and anaesthetists, after surgery. This procedure has become standard practice and therefore can be replicated to good effect in all patients undergoing RC. Not surprisingly, advancing age was associated significantly with a poor outcome, as has been reported previously [9]. Despite this, advanced age is not an exclusion criterion for surgery, as the complication rates remain acceptably low in carefully selected patients. There was no association between ASA grade and complication rates in the present data. It is accepted that there were few patients undergoing a full ASA-based evaluation, and this only occurred in most of the 1999–2005 cohort. ASA grade has been identified previously as a risk factor for complications [10] and definitive statements relating to this will require prospective evaluation of more patients.

There was a significant decline in the 30- and 60-day mortality over the decades, and a concomitant and significant decline in re-operation rates. The 30-day mortality reported here, at 0.4%, is lower than that in most other series, which report rates of 0.8–4.5%[5,11,12], and only Chang et al.[13] reported a lower 30-day mortality rate of 0.3%. However, that series did not include cases of salvage RC after RT, or total pelvic exenteration. We believe that the outcomes in the present report now provide a benchmark for surgery of this type, which can be used not only for comparing the outcome between centres, but also with the outcome for RC using other treatments, such as standard or robot-assisted laparoscopic RC. The low mortality rates reported here might be explained by standardization of surgical/anaesthetic techniques and management before and after surgery, but another factor is likely to have contributed to these outcome data. The volume of major pelvic surgery through our cancer centre is high and this is likely to have contributed to the better outcomes. It has been shown clearly that high volumes are associated with better outcomes for patients when complex surgery is being undertaken [14].

The contemporary re-operation rate for the present series (<30 days) was 4.7%; this is lower than that recently reported of 6.2% by Novotny et al.[4]. However, it is important that while this benchmark value is low, complications relating to RC for whatever reason are high, with similar rates reported in other series to those documented here [3,4,6]. The collection of data concerning complication rates and the definition of what actually constitutes a ‘complication’ is important. There is no existing standardization of this definition and so in some series, e.g. a wound infection or prolonged ileus, might not have been classified as a complication, whereas in the present prospective data collection, these are classed as complications. We have attempted to address this issue by defining complications as medical and surgical. Whilst this is not a novel approach, this type of subclassification might be of use in discriminating between complications which are of major importance from those which are less so. Thus, complications such as bowel leakage or re-operation for bleeding, etc., which are more commonly life-threatening, might be viewed in an appropriate context by comparison with lesser complications, such as ileus or wound infection, which are very common after RC but which are of lesser consequence to morbidity and mortality after surgery. Such a subclassification and standardization of reporting would make comparative audit and the determination of outcome related to clinical governance much more relevant in the monitoring of how an individual unit/centre is performing.

In conclusion, we report a large series of RC from a single UK cancer centre, including an unselected series of patients undergoing all types of pelvic exenteration. We showed a sequential decade-by-decade reduction in the mortality rate for patients undergoing surgery of this type. The contemporary standard for 30- and 60-day mortality rates is now 0.4% and 2.6%, respectively, but morbidity remains high. Pathological stage and previous RT were not statistically significant risk factors for complication rates or 30- and 60-day mortality, but age was significantly associated with mortality.

CONFLICT OF INTEREST

None declared.

Ancillary