To determine short-term differences in renal function evolution between patients with renal cell carcinoma (RCC) submitted to radical nephrectomy (RN) and living kidney donors matched for age and gender.
To assess the role of co-morbidity as a risk factor for developing an estimated glomerular filtration rate (eGFR) of <60 mL/min/1.73 m2.
Patients and Methods
In this retrospective study patients undergoing Radical Nefrectomy (RN) between January 2000 and February 2011 for suspicion of localised RCC were matched by age and gender to living kidney donors.
Renal function was compared between the groups using the Modification in Diet and Renal Disease (MDRD) equation at 1 year after RN.
Charlson co-morbidity score, incidence of hypertension, diabetes and cardiovascular disease were compared and assessed as predictors for developing an eGFR of <60 mL/min/1.73 m2.
In all, 196 patients were included, 98 in each group.
The mean age was respectively 60.6 (RCC group) and 59.1 years (donors).
The 1-year postoperative mean eGFR (available in 89 patients with RCC and 87 donors) was similar, at a mean (sd) of 56.7 (16.4) mL/min/1.73 m2 in patients with RCC and 56.2 (9.8) mL/min/1.73 m2 in donors (P = 0.83).
In patients with RCC the incidence and severity of co-morbidities was significantly higher.
A preoperative eGFR of 60–89 mL/min/1.73 m2 was the only independent risk factor for developing a postoperative eGFR of <60 mL/min/1.73 m2 (odds ratio 4.4, confidence interval 2.1–9.5, P < 0.001, 95% confidence interval).
In our cohorts with advanced age the 1-year follow-up eGFR was similar in both groups.
Despite increased co-morbidity in the RCC group there was no increased decline in renal function.
Only reduced preoperative eGFR could be identified as risk factor for developing a postoperative eGFR of <60 mL/min/1.73 m2.
Various studies have shown that RN for RCC may lead to long-term adverse renal outcomes, e.g. deterioration of the estimated GFR (eGFR) or hospitalisation [1, 2]. An eGFR of <60 mL/min/1.73 m2 has been described in 45–70% of the patients after RN [3, 4]. Although the number of patients that eventually requires renal replacement therapy is low (2.0– 6.2%) [1, 2], chronic kidney disease (CKD) deteriorates the quality of life and is a risk factor for cardiovascular disease (CVD) and overall mortality .
In contrast to patients undergoing RN for RCC, living kidney donors rarely develop end-stage CKD (0.3–0.5%) after RN [6, 7].The fact that donor selection is primarily based on good health and limited co-morbidity, has been the advocated reason .
By contrast, Timsit et al.  reported that donors and patients with RCC with a low co-morbidity load presented with the same percentage of decrease in renal function at 1–4 years after RN, despite the patients RCC being older and having lower baseline renal function.
The primary objective of the present study was to determine short-term differences in renal function evolution between patients with RCC submitted to RN and living kidney donors at an age when RCC is predominant. The secondary objective was to assess the role of co-morbidity as a risk factor for developing an eGFR of <60 mL/min/1.73 m2.
Patients and Methods
For this single centre retrospective longitudinal cohort study, all patients aged 18–75 years who underwent (open or laparoscopic) RN for suspicion of RCC between January 2000 and February 2011 in our centre were identified. Patients with an abnormal contralateral radiological kidney, solitary kidney or synchronous RCC, patients with metastatic or locally advanced cancer at RN (pT4/N1–2), previous surgery for RCC and pre-existing severe renal insufficiency (defined as an eGFR of <30 mL/min/1.73 m2) were excluded.
In addition, we retrieved all living kidney donors aged 18–75 years operated upon in our hospital during the same period from a prospective database.
Selected patients undergoing RN for solitary renal masses (RCC group) were matched for age and gender to the most suitable group of kidney donors (donor group).
Due to the retrospective nature of this study approval of the medical ethics committee was not required.
Renal function was evaluated by calculating the eGFR using the four variable MDRD equation . To obtain maximal follow-up measurements we contacted patient's GPs for creatinine values and co-morbidity development. The eGFR was assessed preoperatively, and at 6 months and 1 year postoperatively. To mirror the stages of kidney disease defined by the National Kidney Foundation , the eGFR was divided in the following stages ≥90; 60–89; 30–59; 15–29; <15 mL/min/1.73 m2.
Baseline co-morbidity load was assessed by the Charlson co-morbidity score  and by a history of hypertension, diabetes and CVD.
RCC tumours were staged according to the 2009 TNM staging system .
Propensity score matching using a multivariate logistic regression model to create single scores was used for matching the patients with RCC with kidney donors in a 1:1 ratio based on age and gender.
To assess the differences between the RCC and donor groups the independent samples t-test was used for continuous (normal) variables and the chi-square test for categorical variables.
Independent risk factors for developing an eGFR of <60 mL/min/1.73 m2 at 1 year postoperatively were evaluated by univariate analyses using the independent samples t-test for continuous (parametric) variables (age and body mass index [BMI]) and the chi-square test and odds ratio (OR) for categorical variables (preoperative eGFR stage, presence of diabetes, hypertension, CVD, Charlson score of ≥2 and gender). Multivariate analyses (preoperative eGFR stage, age, BMI, Charlson score of ≥2 and gender) were done using logistic regression (all variables entered). Variables included in the multivariate analyses were statistically significant on univariate analyses or known with relation to renal function deterioration.
A multiple imputation and subsequent analyses with the pooled data were done to assess the effect of missing data on eGFR outcomes.
All tests were two-tailed and a P < 0.05 was considered to indicate statistical significance. Matching was done using R statistical software version 2.13.2. Data was analysed using Predictive Analytics Software Statistics 18.0.2 (Armonk, New York, United States of America).
Both the RCC and living donor groups contained 98 patients. Baseline characteristics of both groups are shown in Table 1. None of the living donors had proteinuria at baseline and donor RN was performed laparoscopically.
Table 1. Differences in baseline characteristics of RCC and donor RN patients.
†BMI was available in 96 RCC patients and in 94 donors; ‡American Society of Anesthesiologists (ASA) score was available in 92 RCC patients and 95 donors; *P < 0.05.
The RCC group consisted of 92 RCCs and six benign tumours. The median (range) pathological tumour size was 61 (11–180) mm. Pathological T-stage was divided as follows: T1a, 14 (15.2%); T1b, 28 (30.4%); T2a, 11(12.0%); T2b, nine (9.8%); T3a 28 (30.4%); T3b two (2.2%).
Postoperative differences in renal function are shown in Table 2. Creatinine measurements at 1 year after RN were available in 89 patients with RCC and 87 donors. Four patients with RCC died before the 1-year follow-up was reached (two from other causes and two from progressive RCC). Renal function measurements were not available (in the hospital or with the GP) for five patients with RCC and six donors; five of the donors had no GP or came from abroad.
Table 2. Postoperative renal function and changes vs baseline renal function at 6 months and 1 year after RN.
*Renal function was estimated using the MDRD equation.
The 1-year postoperative eGFR in patients with RCC was distributed as follows: ≥90 mL/min/1.73 m2 in four (4.5%); 60–89 mL/min/1.73 m2 in 29 (32.6%); 30–59 mL/min/1.73 m2 in 52 (58.4%) and 15–29 mL/min/1.73 m2 in four (4.5%) patients. In the donor group, at 1 year 33 donors (37.9%) had an eGFR of 60–89 mL/min/1.73 m2 and 54 (62.1%) had an eGFR of 30–59 mL/min/1.73 m2 (P = 0.04). No patient in the donor group developed an eGFR of <30 mL/min/1.73 m2.
Figure 1 shows the changes from preoperative to postoperative renal function per preoperative stage. The percentage decrease for each patient between baseline eGFR and the eGFR at 1 year postoperatively is shown in Figure 2.
The preoperative eGFR of the four patients with RCC that developed a postoperative eGFR of 15–29 mL/min/1.73 m2 were 42, 44, 46 and 61 mL/min/1.73 m2. The postoperative evaluation of renal function showed that hypertension in all of them was associated with nephrosclerosis (1 patient), diabetes type II and micro-albuminuria (1), metabolic syndrome (1) and diabetes type II and chronic NSAID use (1).
When assessing patients with a preoperative of eGFR ≥60 mL/min/1.73 m2, new onset of an eGFR of <60 mL/min/1.73 m2 developed in 41 of 74 patients with RCC (55.4%) and 54 of 86 donor patients (62.8%) (P = 0.34).
Univariate analyses showed significant associations between developing an eGFR of <60 mL/min/1.73 m2 (P < 0.001), gender (P = 0.01) and a preoperative eGFR of 60–89 mL/min/1.73 m2 (P < 0.001). In multivariate analyses (including 155 patients who had all data available at baseline and at the 1-year follow-up) including preoperative eGFR stage, age, BMI, Charlson score ≥2 and gender the only independent factor associated with developing a postoperative eGFR of <60 mL/min/1.73 m2 was a preoperative eGFR of 60–89 mL/min/1.73 m2 (OR 4.5, CI 2.1–9.5, P < 0.001).
Calculations for postoperative renal function outcomes using pooled multiple imputation results were similar to results with the original data. With the imputed data the 1-year postoperative eGFR was 56.6 and 56.8 mL/min/1.73 m2 respectively in patients with RCC and donors (P = 0.93); creatinine was 118.6 and 111.1 μmol/L (P = 0.09); and the percentage decrease in eGFR compared with baseline was 32.1% and 38.0% (P = 0.003).
In the present study we assessed the evolution of renal function after RN at an age when RCC is predominant. Our results suggest that RN at advanced age results in a significant 1-year decline in renal function in both patients with RCC and living kidney donors independently of the presence of co-morbidity. A preoperative eGFR of 60–89 mL/min/1.73 m2 predicted the development of a post-RN eGFR of <60 mL/min/1.73 m2.
After RN, decrease in renal function has been reported in both patients with RCC and living kidney donors [6, 12-14]. However, the mean age and co-morbidity of living donors were traditionally considerably lower than the corresponding figures in patients with RCC. In older living donors, with a mean age in the range of our RCC cohort, a similar eGFR decrease has been reported [15, 16]. Renal function decline in previous studies seemed to be dependent on age at donation [3, 12]. It was more prominent in donors aged ≥60 years than in younger ones. Up to 80% of the elderly might develop a post-RN eGFR of <60 mL/min/1.73 m2, while this figure accounted for 31% in the younger patients .
Conversely comparison between living kidney donors and ‘healthy’ patients with RCC (meeting criteria for living donation) showed no difference in post-RN renal function decrease (≈30% in both groups), despite of significant differences in mean age at RN (57.9 years in patients with RCC vs 48.5 years in donors) and baseline renal function . This later study suggests that age and baseline renal function might not be the most important determinants in renal function outcomes after RN and that other factors e.g. co-morbidity, may play an important role.
We hypothesised that matching groups for age should minimise its effect as a risk factor on post-RN renal function outcomes and underscore the effect of co-morbidity. As renal function is subjected to gender differences a further match on gender would overcame this confounder [17, 18]. However, the present two cohorts presented similar 1-year eGFR despite the higher co-morbidity load in the RCC group. Furthermore, the decline in renal function of the donor group was higher than that of the RCC group.
In the present study the only explanatory reason for this finding was the better preoperative renal function of the donor group (higher eGFRs) and the trend to a lesser eGFR decrease among patients with lower baseline renal function as displayed in Figure 1. Taken into account the median tumour size and the high rate of T1b–T2 tumours the possibility remains that the tumour-bearing kidney was contributing lesser to renal function in more than half of the RCC group while the remaining kidney was the major contributor to renal function. Because of the retrospective nature of the study no radioisotope studies assessing separately the relative contribution to total renal function were available in the RCC group and thus this hypothesis could not be confirmed.
While we did not expect any effect of age in the present analysis, the striking point is that in these two age/gender-matched cohorts we could not demonstrate any association between co-morbidity and development of an eGFR of <60 mL/min/1.73 m2. In the age range of the present study, baseline eGFR was the only predictive factor for development of an eGFR of <60 mL/min/1.73 m2 after RN. Specifically a baseline eGFR of 60–89 mL/min/1.73 m2 was associated with a four-fold higher chance of developing an eGFR of <60 mL/min/1.73 m2 at 1 year after RN.
A 1-year postoperative eGFR of >90 mL/min/1.73 m2 in 4.5% of the patients in the RCC group is anecdotal and reflects the normal range variations. Conversely, the 4.5% of patients with severe CKD at 1 year is of concern. Of these four patients with RCC, three had a low eGFR before RN (between 42 and 46 mL/min/1.73 m2) and two of them presented with micro-albuminuria.
The present results, reinforced by the lack of progression to severe CKD in the donor group, suggest that criteria for living donor may expand to those elderly patients with a certain degree of co-morbidity provided their baseline eGFR is >89 mL/min/1.73 m2. Consequently an increasing pool of individuals with advanced age may still be considered for living kidney donation under strict adherence to pre-donation screening guidelines including evaluation of proteinuria/micro-albuminuria [19-21]. Conversely considerations should be directed to assessing safety of donation in the elderly in the presence of basal CKD II.
From the urological point of view, the present results are a plea for a careful evaluation before considering surgical treatment of renal tumours. Indications on nephron-sparing surgery (NSS) in patients at risk of developing renal insufficiency should be expanded in an attempt to avoid detrimental quality of life or serious events  and RN could still be considered in the elderly population with an excellent basal renal function even in tumours technically amenable to NSS.
We do acknowledge that the above conclusions deserve critical consideration. Firstly, renal function was assessed at short follow-up hindering conclusions on the influence of co-morbidity on long-term outcomes. Although several studies have indicated that changes in renal function occur mainly during the first year after RN [8, 16, 23], GFR still declines with age on average 0.8 mL/min/1.73 m2 in healthy individuals. It is considered a ‘normal’ senescence process well tolerated in healthy old but still co-morbidity might appear or aggravate complicating renal function outcomes . Secondly, proteinuria determination was not available in all the present patients with RCC precluding assessment of its weighted value on renal function outcomes. Macroalbuminuria (≥300 mg albumin/24 h urine) has a higher impact on the risk of decline in renal function than impaired baseline renal function  and recent data indicates its usefulness on the decision treatment algorithm in patients with suspected RCC . Thirdly, in the present study eGFR was used as a proxy for CKD stages. We applied the MDRD formula, which is known to increase in accuracy when eGFR is <60 mL/min/1.73 m2 with respect to the Cockcroft and Gault formula . Fourth, the present study sample is relatively small. However, age and gender matching allowed for independent assessment on the value of co-morbidity as a predictor in a given age range. Lastly, 1-year renal function was not available in a few patients for medical or social reasons. Efforts were maximised to collect follow-up data and rates of an eGFR of <60 mL/min/1.73 m2 were similar in both cohorts. Noteworthy after multiple imputations the present results did not change.
Despite these limitations, the present results suggest that living donation programmes may be expanded to include highly selected elderly patients and support the current policy on NSS in RCC when feasible allowing the option of a RN in patients with excellent renal function. Further prospective research on the long-term effects of co-morbidity and other factors, e.g. proteinuria on renal function in patients with RCC and elderly living donors should be pursued.
In conclusion, after RN patients with RCC and living kidney donors matched for age and gender had similar renal function at 6 months and 1 year postoperatively, despite a higher co-morbidity load and lower baseline renal function in the patients with RCC. Only a preoperative eGFR of 60–89 mL/min/1.73 m2 was an independent risk factor for developing an eGFR of <60 mL/min/1.73 m2. No association was found between the presence of co-morbidity and renal functional outcomes.
We acknowledge Tessa de Jong and Janneke Vervelde, transplant coordinators of the living kidney donation programme in the AMC Amsterdam. Their help in providing data on living kidney donors was invaluable. We thank Laura Sandbergen for her help in the data collection.
We also acknowledge the Cure for Cancer foundation for their help in supporting our research with an unrestricted research grant.
Conflict of Interest
Jean De La Rosette is a paid consultant for BSC (Boston Scientific Corporation). No other conflicts disclosed.