The role of American Society of Anesthesiologists scores in predicting urothelial carcinoma of the upper urinary tract outcome after radical nephroureterectomy: results from a national multi-institutional collaborative study

Authors


Morgan Rouprêt, Hôpital Pitié-Salpêtrière, 47-83 Boulevard de l'hôpital, 75013 Paris, France. e-mail: morgan.roupret@psl.aphp.fr

Abstract

Study Type – Prognosis (cohort)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Upper urinary tract urothelial carcinoma (UUT-UC) is a rare disease, usually treated by nephroureterectomy, occurring in a population with a median age of 70 years and with frequent tobacco use and other comorbidities. We know that the American Society of Anesthesiologists (ASA) score has prognostic value in urological oncology but this has not been assessed in UUT-UC.

Using a multi-institutional French database, we have shown that the 5-year cancer-specific survival differed significantly between ASA 1, ASA 2 and ASA 3 patients (83.8%, 76.9% and 70.6%, respectively; P= 0.01). ASA status had a significant impact on cancer-specific survival in univariate and multivariate analyses, with a threefold higher risk of mortality at 5 years for ASA 3 compared with ASA 1 patients (P= 0.04).

OBJECTIVE

  • • To evaluate the impact of American Society of Anesthesiologists (ASA) scores on the survival of patients treated with radical nephroureterectomy (RNU) for upper urinary tract urothelial carcinoma (UUT-UC).

PATIENTS AND METHODS

  • • A retrospective multi-institutional cohort study of the French collaborative national database of UUT-UC treated by RNU in 20 centres from 1995 to 2010.
  • • The influence of age, gender and ASA score on survival was assessed using a univariable and multivariable Cox regression analysis with pathological features used as covariables.

RESULTS

  • • Overall, 554 patients were included. The median follow-up was 26 months (10–48 months), and the median age was 69.5 years (61–76 years). In total, 114 (20.6%) patients were classified as ASA 1, 326 (58.8%) as ASA 2 and 114 (20.6%) as ASA 3.
  • • The 5-year recurrence-free survival (P= 0.21) and metastasis-free survival (P= 0.22) were not significantly different between ASA 1 (52.8% and 76%), ASA 2 (51.9% and 75.3%) and ASA 3 patients (44.1% and 68.2%, respectively).
  • • The 5-year cancer-specific survival differed significantly between ASA 1, ASA 2 and ASA 3 patients (83.8%, 76.9% and 70.6%, respectively; P= 0.01).
  • • ASA status had a significant impact on cancer-specific survival in univariate and multivariate analyses, with a threefold higher risk of mortality at 5 years for ASA 3 compared with ASA 1 patients (P= 0.04).

CONCLUSIONS

  • • ASA classification correlates significantly with cancer-specific survival after RNU for UUT-UC.
  • • It is a further pre-operative clinical variable that can be incorporated into future risk prediction tools for UUT-UC to improve their accuracy.
Abbreviations
UUT-UC

Urinary upper tract urothelial carcinoma

RFS

recurrence-free survival

CSS

cancer-specific survival

RNU

radical nephroureterectomy

ASA

American Society of Anesthesiologists

MFS

metastasis-free survival

CSM

cancer-specific mortality

IQR

interquartile range

OR

odds ratio

INTRODUCTION

Urinary upper tract urothelial carcinoma (UUT-UC) is a rare disease, but its incidence is rising and has recently been estimated to be up to two cases per 100 000 person-years in the USA [1]. Outcome is variable, and many prognostic factors have been described including pathological and clinical parameters. Pathological factors, such as tumour stage, grade and lymph node involvement, remain the most accurate factors according to a recent literature review [2]. More recent data have suggested that lymphovascular invasion could be an independent predictor of recurrence-free survival (RFS) and cancer-specific survival (CSS) [3]. However, determining these pathologic factors requires extirpative surgery because they are currently not accurately defined by other means, even with endoscopic biopsies [4]. Recent guidelines [5] have confirmed that radical nephroureterectomy (RNU) with bladder cuff excision is the standard treatment, whereas conservative management, primarily endoscopic, is indicated in imperative or selected elective cases [5].

The clinical prognostic factors studied so far have failed to help physicians make treatment decisions, with contradictory results in the few studies available [6,7]. Comorbidities are highly prevalent in this population, with frequent exposure to tobacco and with a median age of diagnosis of approximately 70 years [8]. The American Society of Anesthesiologists (ASA) score was developed to predict perioperative outcome according to these comorbidities (see Table 1) and it has been shown to have a prognostic impact in oncological urology [9,10]. No study has assessed the prognostic significance of ASA scores in patients with UUT-UC. Our objective was to evaluate the impact of ASA scores on the survival of patients treated for UUT-UC with RNU from our multi-institutional collaborative database.

Table 1. American Society of Anesthesiologists (ASA) physical status classification
ASA ScoreDescription
ASA 1Normal healthy patient
ASA 2Patient with mild systemic disease
ASA 3Patient with severe systemic disease
ASA 4Patient with severe systemic disease that is a constant threat to life
ASA 5Moribund patient who is not expected to survive without the operation
ASA 6Declared brain-dead patient whose organs are being removed for donor purposes

MATERIAL AND METHODS

A national database on UUT-UC was investigated to obtain the medical reports of patients managed in 20 French institutions from 1995 to 2010. After a preoperative evaluation (cystoscopy, urine cytology, intravenous urography or abdominopelvic CT scan and chest radiography or thoracic CT scan), all patients treated for UUT-UC with RNU with curative intent were included. The recorded included gender, age at diagnosis, ASA score (see Table 1), tumour characteristics (2009 UICC TNM classification, 1973 WHO grade classification, lymphovascular invasion, lymph node involvement and tumour location), management, follow-up and outcome events. Specimens were analysed by dedicated genitourinary pathologists according to standardized procedures. Patients with an incomplete data set (ASA score missing) or with metastatic disease at diagnosis were excluded from the current study. Any patients with a synchronous or previous history of urothelial bladder carcinoma were excluded.

Open RNU was performed by a standard double-access procedure: a loin, subcostal or midline incision for the ‘nephrectomy’ portion followed by an iliac incision for management of the distal ureter and extravesical excision of a bladder cuff. The kidney, ureter and a bladder cuff were excised en bloc. A regional lymphadenectomy was performed when nodal involvement was suspected from the preoperative evaluation or was discovered during the procedure. Approximately 20% of patients received adjuvant chemotherapy for high-risk pathological features, e.g T3 stage or N+ve. The database does not include accurate data on salvage chemotherapy.

For postoperative assessment the patients were followed up at 3 months, 6 months, then every 6 months for 3 years after RNU, and annually thereafter. Follow-up examinations included a history, physical examination, blood laboratory tests, urinary cytology, cystoscopic evaluation of the urinary bladder and a thoraco-abdomino-pelvic CT scan. Disease recurrence was defined as any documented failure in the operative field, bladder or the contralateral UUT. Metastatic progression was defined as any recurrence in the regional lymph nodes or distant metastases. Chart reviews corroborated the cause of death. The 5-year RFS, metastasis-free survival (MFS) and CSS data were assessed according to these definitions.

The demographic and clinicopathological features in the three ASA groups were compared using chi-squared tests for categorical variables and Kruskal–Wallis tests for continuous variables. Postoperative survival was estimated using the Kaplan–Meier method and compared between groups with the log-rank test. The oncological outcome analyses focused on RFS, MFS and CSS. Patients were censored at the last follow-up or death. Univariate and multivariate Cox proportional hazards regression analyses were conducted to ascertain the independent role of ASA status in cancer-specific mortality (CSM) variations. A P-value < 0.05 was considered significant. All statistical analyses were performed using SPSS version 17.0 statistical software (IBM Corp., Somers, NY, USA).

RESULTS

Overall, 554 patients were included in the current study. The median patient age was 69.5 years (interquartile range (IQR) 61–76), and the male-to-female ratio was 2.2 : 1. The median follow-up was 26 months (IQR 10–48). Patient comorbidities were thoroughly assessed by anaesthesiologists preoperatively to evaluate the patients' ASA physical status; 114 (20.6%), 326 (58.8%) and 114 (20.6%) patients were finally classified ASA 1, ASA 2 and ASA 3, respectively. No patient classified as ASA 4 was managed with RNU with a curative intent. Demographic and clinicopathological features of the whole study population and the three ASA groups are summarized in Table 2; no significant difference was found between the ASA groups.

Table 2. Study population characteristics
ASA ScoreAll (n= 554)ASA 1 (n= 114)ASA 2 (n= 326)ASA 3 (n= 114) P
  1. ASA: American Society of Anesthesiologists; WHO: World Health Organization; is: in situ.

Age at diagnosis (years)     0.59
 Median69.569.169.169.9 
 Range29–8929–8834–8935–88 
Gender, n (%)     0.22
 Female175 (31.6)34 (29.8)112 (34.4)29 (25.4) 
 Male379 (68.4)80 (70.2)214 (65.6)83 (72.8) 
Pathological stage pT, n (%)     0.19
 pTa/pTis152 (27.4)38 (33.3)81 (24.8)33 (28.9) 
 pT1132 (23.8)24 (21.1)84 (25.8)24 (21.1) 
 pT257 (10.3)10 (8.8)30 (9.2)17 (14.9) 
 pT3186 (33.6)35 (30.7)119 (36.5)32 (28.1) 
 pT427 (4.9)7 (6.1)12 (3.7)8 (7.0) 
Grade (OMS 1973), n (%)     0.21
 G145 (8.1)12 (10.5)23 (7.1)10 (8.8) 
 G2195 (35.2)46 (40.4)117 (35.9)32 (28.1) 
 G3314 (56.7)56 (49.1)186 (57.1)72 (63.2) 
Lymph node status, n (%)     0.35
 pNx322 (58.1)63 (55.3)200 (61.3)59 (51.8) 
 pN0186 (33.6)43 (25.4)99 (30.4)44 (38.6) 
 pN1/246 (8.3)8 (7.0)27 (8.3)11 (9.6) 
Lymphovascular invasion, n (%)     0.97
 No453 (81.8)94 (82.5)266 (81.6)93 (81.6) 
 Yes101 (18.2)20 (17.5)60 (18.4)21 (18.4) 
Tumour location, n (%)     0.30
 Pyelo-caliceal297 (53.6)62 (54.4)178 (54.6)57 (50.0) 
 Ureter161 (29.1)37 (32.5)85 (26.1)39 (34.2) 
 Multifocal96 (17.3)15 (13.2)63 (19.3)18 (15.8) 
Follow-up (months)     0.77
 Median26.5272527 
 Range0.2–2221–1460.2–1931–222 

Overall, 208 (37.5%) patients experienced recurrence during follow-up: 38 (18.3% of the recurrent patients) patients classified as ASA 1, 122 (58.6%) as ASA 2 and 48 (23.1%) as ASA 3 had a 5-year RFS estimated at 52.8%, 51.9% and 44.1%, respectively. The statistical analysis did not show any significant difference between ASA groups (P= 0.21).

A total of 109 patients (19.7%) developed metastases, of which 20 (18.3%), 62 (56.9%), 27 (24.8%) occurred in patients classified as ASA 1, ASA 2 and ASA 3, respectively. The 5-year estimated MFS were not significantly different for any of the ASA score (76% for ASA 1, 75.3% for ASA 2 and 68.2% for ASA 3; P= 0.22).

In terms of CSS, 88 of 554 patients (15.9%) died of UUT-UC during follow-up; nine (10.2%) ASA 1, 55 (62.5%) ASA 2 and 24 (27.3%) ASA 3 patients, respectively. The Kaplan–Meier method estimated that 5-year CSS was 83.3%, 76.9% and 70.6% for each group, respectively (P= 0.01; Fig. 1). The univariate analysis confirmed this statistical significance with hazard ratios of CSM of 2.2 (95% CI 1.09–4.48) for ASA 2 patients and 3.1 (95% CI 1.46–6.78) for ASA 3 patients compared with the ASA 1 group. This was confirmed after controlling for pathological stage, tumour grade, lymph node status and tumour location in the multivariate analysis shown in Table 3.

Figure 1.

Cancer-specific Kaplan–Meier survival curves for upper urinary tract urothelial carcinoma after radical nephroureterectomy overall (left) and stratified by American Society of Anesthesiologists (ASA) score (right). Kaplan–Meier estimated five-year cancer-specific survival was 83.3%, 76.9% and 70.6% for the ASA 1, ASA 2 and ASA 3 group, respectively (P= 0.01).

Table 3. Univariate and multivariate analyses of the impact of American Society of Anesthesiologists scores and covariables on overall and cancer-specific mortality for upper urinary tract urothelial carcinoma treated with radical nephrectomy
CovariablesOverall mortalityCancer-specific mortality
Univariate analysisMultivariate analysisUnivariate analysisMultivariate analysis
HR (95% CI) P HR (95%CI) p HR (95% CI) P HR (95% CIl) P
  1. HR, hazard ratio; 95% CI, 95% confidence interval; ASA: American Society of Anesthesiologists; WHO: World Health Organization; is: in situ.

Age (continuous variable) 1.02 (1.007–1.04)0.0061.02 (1.002–1.04)0.0281.02 (1.006–1.04)0.011.02 (1.003–1.04)0.025
ASA score  0.001 <0.001 0.014 0.012
 ASA 2 vs ASA 12.12 (1.14–3.91)0.0161.99 (1.06–3.75)0.0322.21 (1.09–4.48)0.0272.00 (0.96–4.14)0.062
 ASA 3 vs ASA 13.49 (1.81–6.75)<0.0013.57 (1.84–6.94)<0.0013.14 (1.46–6.78)0.0033.14 (1.44–6.84)0.004
Tumour location  0.006 0.092 <0.001 0.009
 Ureter vs pelvi-caliceal1.19 (0.77–1.83)0.4271.25 (0.80–1.98)0.3161.52 (0.92–2.51)0.1001.66 (0.99–2.81)0.054
 Multifocal vs pelvi-caliceal2.07 (1.31–3.26)0.0021.68 (1.05–2.69)0.0302.90 (1.73–4.84)<0.0012.23 (1.31–3.79)0.003
Pathological stage pT  <0.001 0.001 <0.001 0.002
 pT1 vs pTa/pTis1.32 (0.67–2.59)0.4201.17 (0.58–2.35)0.6473.33 (1.2–9.28)0.0212.64 (0.93–7.50)0,067
 pT2 vs pTa/pTis2.65 (1.27–5.52)0.0091.98 (0.89–4.39)0.0936.50 (2.22–19.05)0.0013.54 (1.13–11.04)0,029
 pT3 vs pTa/pTis4.52 (2.58–7.90)<0.0013.00 (1.51–5.95)0.00212.19 (4.84–30.69)<0.0016.03 (2.16–16.82)0.001
 pT4 vs pTa/pTis8.87 (4.32–18.23)<0.0015.08 (2.11–12.25)<0.00121.97 (7.61–63.41)<0.0019.45 (2.85–31.27)<0.001
Grade (WHO 1973)  0.003 0.274 <0.001 0.380
 G2 vs G12.17 (0.65–7.21)0.2021.78 (0.53–5.95)0.3482.88 (0.37–22.06)0.3082.19 (0.28–17.1)0.454
 G3 vs G16.12 (1.92–19.51)0.0022.40 (0.70–8.23)0.16314.86 (2.06–107.14)0.0074.78 (0.62–36.91)0.133
Lymph node status  <0.001 0.081 <0.001 0.195
 pNx vs pN01.17 (0.77–1.77)0.4380.96 (0.62–1.49)0.8651.21 (0.75–1.96)0.430.95 (0.57–1.58)0.850
 pN1/2 vs pN04.31 (2.62–7.09)<0.0011.78 (1.02–3.10)0.0404.97 (2.87–8.61)<0.0011.64 (0.90–2.99)0.106
Lymphovascular invasion         
 Yes vs no2.06 (1.33–3.19)0.0011.04 (0.64–1.68)0.862.25 (1.38–3.67)0.0011.04 (0.61–1.78)0.861

DISCUSSION

Current contemporary management decisions in urological oncology are underpinned by guidelines from key international regulatory bodies, levels of evidence, grades of recommendation and the development of accurate risk prediction models. These latter tools are complex statistical modelling techniques that necessitate the interpretation and correlation of specific variables with defined endpoints to generate a patient-specific risk probability [11]. In general, these variables can be clinical, pathological, radiological or even molecular. They can be used in a pretreatment or post-treatment setting to predict the risk of a variety of outcomes including disease stage, disease recurrence and survival. Survival is a heterogeneous endpoint because it encompasses overall, cancer-specific, recurrence-free and metastasis-free survival.

For UUT-UC, the after clinical and pathological variables have been studied to examine their relationship with oncological outcomes; patient age [7], gender [6], tumour location and multifocality [12], tumour grade [13], tumour stage [14], lymphovascular invasion [3], tumour architecture [15] and associated carcinoma in situ[16]. In various collective forms, they have been incorporated into nomogram models to predict either CSS [17] or disease stage at time of NU [18].

The ASA classification has been clearly proven to be associated with perioperative and postoperative outcomes in surgery in general. Wolters et al. [19] assessed 6301 surgical patients for the association of ASA score with perioperative and postoperative risk factors. On univariate analysis, significant association (P < 0.05) between ASA class and perioperative complications (blood loss, duration of intensive therapy unit stay, duration of postoperative ventilation) was found. The risk of postoperative complications was significantly higher for ASA 3 (odds ratio (OR) 2.2) and ASA IV (OR 4.2) patients. It is also reported that the ASA classification is superior to other notable scoring systems including the Charlson score [10] and the New York Heart Association classification of cardiac insufficiency [20] in predicting operative outcomes.

In urological terms, the ASA score has been evaluated in prostate and renal cancer to predict both survival [9,10] and perioperative and postoperative outcomes [21]. The ASA score has also been correlated with outcomes from laparoscopic urological surgery [22]. de Cassio et al. [9] studied 145 patients after nephrectomy for RCC. The ASA score was found to impact CSS on both univariate and multivariate analyses. Five-year CSS was 95.7, 71.1 and 39.8% for ASA 1, ASA 2 and ASA 3, respectively (P= 0.007). ASA score also influenced overall survival (P < 0.001). ASA 3 patients had a significant increased risk of developing metastases (OR 10; P= 0.001). Froehner et al. [10] compared ASA score with Charlson score as predictors of survival after radical prostatectomy. Using a database of 444 patients with a mean follow-up 5.9 years, both classifications were able to predict comorbid (non-cancer) and overall survival. The ASA score was superior to the Charlson score in defining at-risk groups. Han et al. [21] retrospectively analysed 551 patients after nephrectomy for RCC. No difference in perioperative or postoperative complications was seen but ASA 3 patients had a higher risk of transfusion (42% vs 28%; P= 0.001). Matin et al. [22] evaluated 399 patients who had undergone ‘laparoscopic urological surgery’. Age > 65 years was not associated with complications on univariate and multivariate analysis. Patients > 65 years were hospitalized for longer (43 vs 24 h, P= 0.02). Patients with higher ASA scores were more likely to receive blood transfusions. On univariate analysis, no difference in rate of complications between ASA groups was seen but on multivariate analysis (ASA 1 vs ASA 3), the ASA 3 patients had a higher rate of complications (P= 0.04).

Herein, we have evaluated our large multi-institutional database on UUT-UC in respect of ASA influence on survival outcomes after RNU in 554 patients. Using univariate and multivariate regression analysis, we have shown that the 5-year CSS differed significantly between ASA 1, ASA 2 and ASA 3 patients (83.8%, 76.9% and 70.6%, respectively; P= 0.01). We found no influence on RFS (P= 0.21) or MFS (P= 0.22). ASA status had a significant impact on CSS in univariate and multivariate analyses, with a threefold higher risk of mortality at 5 years for ASA 3 compared with ASA 1 patients (P= 0.04). The statistical difference in CSS (P= 0.01) but not RFS (P= 0.21) or MFS (P= 0.22) is a point of interest. It implies that the ASA score does not directly impact biological cancer phenomena (e.g. recurrence or metastasis) but it is plausible that reduced CSS for the higher ASA score was a consequence of the patients not being eligible for salvage chemotherapy upon the development of recurrence or metastasis. However, although the statistic analysis did not show a difference there were definite trends for absolute difference when comparing ASA 1 with ASA 3. For 5-year RFS, a +8.7% difference for ASA 1 (52.8% vs 44.1%) was seen. For 5-year MFS, +7.8% for ASA 1 (76% vs 68.2%) and for 5-year CSS, +12.7% for ASA 1 (83.3% vs 70.6%). We did not include the data on adjuvant or salvage chemotherapy because in a previous study of patients in our database we showed that giving adjuvant chemotherapy did not improve survival outcomes after RNU for UUT-UC [23].

We would like to address some limitations of our study. Our results were analysed in a retrospective fashion because the relative rarity of this disease makes a prospective study difficult to implement. Additionally, various physicians treated the patients over a long period of time and there was a lack of centralized pathological review. However, all centres in the study followed the recommendations of our national society of urology and the lack of a centralized pathological review is balanced by the relative standardization of practices that are imposed by our national urological and pathological society guidelines [24–26]. Also, the interobserver variability of ASA scoring between different anaesthesiologists in different institutions cannot be ruled out, as has already been shown [27].

To our knowledge, this is the first study to report the significant association of ASA score and survival outcome in UUT-UC patients after RNU. This identifies the ASA score as a predictive variable for UUT-UC and it should now be considered for inclusion in any future predictive modelling technique for UUT-UC outcomes after RNU.

The ASA classification correlates significantly with CSS after RNU for UUT-UC. It is a further preoperative clinical variable that can be incorporated into future risk prediction tools for UUT-UC to improve their accuracy.

CONFLICT OF INTEREST

None declared.

Ancillary