- To determine renal function eligibility for cisplatin-based chemotherapy using our experience with radical cystectomy (RC) patients.
chronic kidney disease
The Chronic Kidney Disease Epidemiology Collaboration (equation)
Modification of Diet in Renal Disease
methotrexate, vinblastine, doxorubicin, and cisplatin
Standard of care options for muscle-invasive urothelial carcinoma of the bladder include radical cystectomy (RC) preceded by neoadjuvant cisplatin-based chemotherapy . This notion is supported by level I evidence showing improved overall survival with neoadjuvant cisplatin-based chemotherapy and meta-analyses showing an ≈5% improvement in overall survival at 5-year after RC [2, 3]. Consistent with this, the National Comprehensive Cancer Network (NCCN) guidelines state that cisplatin-based chemotherapy should be ‘considered’ for localised cT2 urothelial carcinoma of the bladder and should be ‘strongly considered’ for cT3N0M0 urothelial carcinoma . For cT4 urothelial carcinoma of the bladder, chemotherapy is further suggested as the preferred initial management strategy .
While cisplatin is the mainstay of chemotherapy regimens for numerous malignancies, eligibility for cisplatin can be rather restrictive, especially for tumours that primarily afflict older patients and those at higher risk for concomitant comorbidities. A recent working group was assembled to define guidelines regarding eligibility for proper use of cisplatin . With a review of the literature and a survey of 120 international medical oncologists, a consensus definition of patients unfit for cisplatin-based chemotherapy was delineated as follows: (i) WHO or Eastern Cooperative Oncology Group (ECOG) performance status of 2 or Karnofsky performance status of 60–70%, (ii) creatinine clearance of <60 mL/min, (iii) grade ≥2 hearing loss or (iv) peripheral neuropathy according to the Common Terminology Criteria for Adverse Events, or (v) Class III heart failure according to the New York Heart Association . Note that chronological age and presence of a solitary kidney do not by themselves deem a patient unfit for cisplatin . An example of the restrictive eligibility for cisplatin use in patients with urothelial carcinoma of the bladder is the fact that overall median age at time of death from disease is 78 years, while the median age for patients enrolled in phase III trials assessing cisplatin efficacy is closer to 64 years [5, 6]. This discrepancy, combined with the fact that many RC candidates either have baseline chronic kidney disease (CKD) or disabling symptoms at presentation including haematuria, pelvic pain, or renal obstruction, are each examples of why only a minority of RC patients in the USA receive neoadjuvant cisplatin-based chemotherapy .
We investigated our experience with RC specifically to determine renal function eligibility for cisplatin-based chemotherapy, both in the neoadjuvant and adjuvant setting, and to analyse features associated with a perioperative change in renal function eligibility status.
After obtaining Institutional Review Board approval, we reviewed the Mayo Clinic Cystectomy Registry and identified 768 patients treated with RC for urothelial carcinoma between 1980 and 2005. Patients were included in this study if they had both a preoperative (≤30 days) and postoperative (60–120 days) measurement of serum creatinine/calculated GFR. Patients who received systemic chemotherapy or radiation before surgery (86 patients) were excluded.
The GFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, which includes serum creatinine, age, race, and gender and is expressed in mL/min/1.73m2 . The preoperative GFR was calculated from the serum creatinine closest to surgery, while the postoperative GFR was calculated from the serum creatinine at 3 months after RC (range 60–120 days). To limit the immediate postoperative impact of rapid fluid shifts, indwelling ureteric stents, acute kidney injury, and other transient factors effecting measurement of renal function, we chose 60–120 days as the period for the postoperative assessment of GFR. As normal renal function was required for phase III trials studying neoadjuvant cisplatin-based chemotherapy  and a consensus panel concluded that patients with a GFR of <60 mL/min are unfit for cisplatin , we used 60 mL/min as the threshold for a patient to be considered eligible for cisplatin both preoperatively and postoperatively.
The CKD-EPI equation includes the same variables (with different coefficients) as the Modification of Diet in Renal Disease Study Equation (MDRD), although it has recently been reported to classify fewer patients as harbouring stage III CKD and is thought to be a more accurate predictor of mortality than the MDRD equation in an international study involving >1 million participants . Nevertheless, for completeness, we also calculated GFR using the MDRD equation and repeated all analyses.
Summary data are presented as the median with interquartile range (IQR) or number and percentage as appropriate. Perioperative change in GFR was calculated as the postoperative GFR minus the preoperative GFR. Features associated with a change in GFR were evaluated using linear regression models. We also evaluated features (age, gender, race, urinary diversion, and preoperative hydronephrosis) associated with new onset CKD (GFR <60 mL/min) after RC among the subset of patients with normal renal function (GFR ≥60 mL/min) before RC using a modified logistic regression model (Firth penalized regression), which accounts for the relative infrequency of non-Caucasian patients. All P values are two-sided and statistical significance was set at 0.05.
The baseline characteristics of the 768 patients, who underwent RC without previous therapy, are given in Table 1. The median age was 68 years and most patients were Caucasian and male. In all, 209 (27%) patients had hydronephrosis noted on preoperative imaging. An incontinent diversion (e.g. ileal conduit) was performed in 557 (73%) patients, while 209 (27%) patients underwent a continent urinary diversion with the vast majority of these representing an ileal neobladder.
|Age, years||68 (61–74)|
|Serum creatinine, mg/dL||1.2 (1.0–1.4)|
|GFR (CKD-EPI), mL/min||60 (48–71)|
|Gender (n = 767):|
|Pathological T stage:|
|Urinary diversion (n = 766):|
|Ileal conduit||557 (73)|
|Continent (e.g. neobladder)||209 (27)|
Among all patients, the median (IQR) preoperative GFR was 60 (48–72) mL/min and 363 (47%) patients had a preoperative GFR of ≥60 mL/min. Thus, 405 (53%) patients had a preoperative GFR of <60 mL/min and were not considered eligible for neoadjuvant cisplatin-based chemotherapy based on renal function status. As might be expected, GFR was age-dependent; among 271 patients aged <65 years, 167 (62%) were eligible for cisplatin-based chemotherapy before RC, whereas only 196 (39%) of 497 patients aged ≥65 years were eligible for cisplatin with a preoperative GFR of ≥60 mL/min (P < 0.001). Additionally, GFR was also significantly better 3 months after RC in younger patients with 188 (69%) patients aged <65 years vs 193 (39%) patients aged ≥65 years having a postoperative GFR of ≥60 mL/min (P < 0.001).
Among the 363 patients who had a preoperative GFR of ≥60 mL/min, 91 (25%) had a postoperative GFR of <60 mL/min which rendered them ineligible for cisplatin after RC. On the other hand, among the 405 (53%) patients who had a preoperative GFR of <60 mL/min and would not have been eligible for neoadjuvant cisplatin, 109 (27%) had an improvement of renal function to a GFR of ≥60 mL/min after RC. Figure 1 shows the relationship between preoperative and postoperative GFR in a scatter plot format.
Meanwhile, among the subset of 209 patients with preoperative hydronephrosis, the median (IQR) preoperative GFR was 52 (39–66) mL/min vs a median (IQR) preoperative GFR of 62 (52–73) mL/min in patients without hydronephrosis (P < 0.001). Among the 132 patients with preoperative hydronephrosis who had a preoperative GFR of <60 mL/min, 37 (28%) had improvement in their renal function after RC and had a postoperative GFR of ≥60 mL/min. Among the 77 patients with preoperative hydronephrosis who had a preoperative GFR of ≥60 mL/min, 24 (31%) had a GFR of <60 mL/min after RC.
We next evaluated features associated with a perioperative change in GFR using linear regression analyses. Variables included age, race, and gender, which are the components of the CKD-EPI equation along with preoperative (GFR and hydronephrosis) and perioperative (choice of urinary diversion) parameters associated with renal function. Table 2 shows the univariable and multivariable features associated with change in GFR. A negative estimate indicates a negative change in GFR, while a positive estimate indicates a positive change in GFR after RC. As shown in Table 2, higher baseline GFR, older age, and continent urinary diversion were significantly associated with a negative change in GFR after RC. While preoperative hydronephrosis was significantly associated with a positive change in GFR univariately, statistical significance was lost in a multivariable model. When we evaluated features associated with new onset CKD (postoperative GFR of <60 mL/min) among the 363 patients with a preoperative GFR of ≥60 mL/min, older age was significantly associated with this change in cisplatin eligibility status in a multivariable analysis (odds ratio 1.06, 95% CI 1.03–1.09, P < 0.001), while the other features (gender, race, urinary diversion, and preoperative hydronephrosis) were not significant either univariately or in a multivariable model.
|Gender (ref. – female)||–3.30||0.011||–1.03||0.4|
|Race (ref. – White)||2.06||0.8||1.23||0.9|
|Urinary diversion (ref. – ileal conduit)||–3.91||<0.001||–2.66||0.014|
In addition, recognising that some patients currently get split-dosed or reduced dose cisplatin with mildly reduced renal function, we report that 162 (21%) and 172 (22%) patients had an eGFR of 50–60 mL/min before and at 3 months after RC, respectively. Of note, replacing the CKD-EPI equation with the MDRD equation for calculation of GFR resulted in similar findings for all results (data not shown).
In the present large cohort of RC patients, ≈50% of patients were not eligible for cisplatin-based chemotherapy, both before and after RC, based on having a GFR of <60 mL/min. Furthermore, we found that older patient age and continent urinary diversions were associated with a negative change in GFR after RC in a multivariate analysis. Importantly, we also report that about one in four patients who were eligible for neoadjuvant cisplatin have a decline in renal function after RC and become unfit  for cisplatin-based chemotherapy and again that older age was significantly associated with this outcome. While we used the CKD-EPI equation for estimation of GFR, a formula that has recently been reported to classify fewer patients as harbouring CKD and may be a more accurate predictor of overall survival , we obtained similar results when using the MDRD equation for estimating GFR. Taken together, this information should prove useful during patient counselling before surgical extirpation of invasive bladder cancer.
Neoadjuvant cisplatin-based chemotherapy is an option for localised cT2–4 urothelial carcinoma of the bladder ; it should be ‘considered’ for patients with cT2N0M0 and ‘strongly considered’ for patients with cT3N0M0 disease . While methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) for three cycles was previously used in a phase III trial supporting neoadjuvant chemotherapy , randomised trials comparing gemcitabine and cisplatin vs MVAC in the metastatic setting supported equivalent oncological efficacy with diminished toxicity using the two drug regimen . Thus, while level I evidence supporting the two drug regimen of gemcitabine and cisplatin in the neoadjuvant setting are lacking, it is not uncommon for the two drug regimen to be used in lieu of MVAC before RC. Regardless, both regimens include cisplatin and thus, eligibility for this drug is of paramount importance for patients with invasive urothelial carcinoma of the bladder. Additionally, while retrospective data supports the use of carboplatin rather than cisplatin-based chemotherapy before RC for non-organ confined bladder cancer , it is important to highlight that no level I evidence supports the use of carboplatin in a neoadjuvant setting for patients with localised urothelial carcinoma of the bladder.
Previously, Dash et al.  reported that among patients with locally advanced urothelial carcinoma at Memorial Sloan-Kettering, 24–52% (depending on patient age and method used to calculate renal function) of patients were ineligible for cisplatin-based chemotherapy after RC. Furthermore, Canter et al.  evaluated the Fox Chase experience with patient eligibility for cisplatin-based chemotherapy among patients with bladder cancer. In their analysis of 194 patients using the Cockgroft-Gault and MDRD equations, 33–41% of patients before RC and 29–40% of patients after RC were not eligible for cisplatin-based chemotherapy; they similarly observed that preoperative GFR and older patient age were independently associated with a negative change in GFR after RC . The present results, with more patients and including more patients treated with a continent urinary diversion (209 vs 14 patients) support those of Canter et al., and further suggest that a continent urinary diversion is associated with a negative change in GFR after RC. Additionally, we present novel evidence that older patient age is significantly associated with new onset CKD after RC among patients with normal preoperative renal function. Collectively, these studies highlight that many patients with bladder cancer are unfit for cisplatin-based chemotherapy and certain populations, including older patients and those planning a neobladder, are more likely to have a decline in renal function after RC and should be counselled as such.
We recognise several limitations of the present work. First, the results were obtained via a retrospective review and are subject to the many limitations inherent with such an approach. Second, we relied on a single measure of GFR estimation before and after RC, which does not account for trends in GFR over time. Furthermore, patients with hydronephrosis before RC were typically managed with immediate RC, although we recognise that such patients may have become eligible for cisplatin therapy with upper tract drainage and a resulting improvement in renal function. In addition, traditional formulas used to calculate renal function tend to underestimate measured creatinine clearance as reported by Raj et al. ; although we used the CKD-EPI equation, which is less likely to label patients with CKD than the MDRD equation . Third, we evaluated eligibility for cisplatin based on renal function alone and recognise the possibility that additional patients would have been considered ‘unfit’ for cisplatin based on Grade ≥2 hearing loss, peripheral neuropathy, or performance status. Fourth, the present observation that continent urinary diversion was independently associated with a negative change in GFR is quite possibly secondary to reabsorption properties of a neobladder  or cutaneous pouch mimicking a calculated reduction in renal function given that catheter drainage was not routinely performed when evaluating GFR. Thus, for patients who would be candidates for adjuvant cisplatin-based chemotherapy after a continent urinary diversion, accurate measurement of renal function with proper urinary drainage should be considered. Nevertheless, the present results, in conjunction with others, show that a substantial proportion of patients with invasive bladder cancer are not optimal candidates for cisplatin-based chemotherapy; thus, further investigation of non-cisplatin regimens in the neoadjuvant/adjuvant setting is needed.
In conclusion, the present results suggest that ≈50% of patients undergoing RC are not eligible to receive perioperative cisplatin-based chemotherapy based on renal function alone. Nearly one in four patients eligible to receive cisplatin before RC had a perioperative decrease in renal function and were no longer eligible for this treatment after RC. Certain patient characteristics and surgical factors are more likely to experience a negative change in GFR after RC and should be counselled accordingly.