- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
RCC is a commonly diagnosed urological malignancy, with an estimated 54 390 new cases and 14 100 deaths in the USA during 2008 . Due to the increased use of abdominal imaging, there has been a stage migration, such that renal masses are being identified at much earlier stages . Radical nephrectomy (RN) remained the standard for managing cortical renal masses until advances in radiographic imaging and an increased understanding of tumour biology caused nephron-sparing techniques to resurface in the 1980s . Since then, nephron-sparing surgery (NSS) has become the standard for treating small renal masses at centres of excellence, with equivalent oncological efficacy and superior renal functional outcomes [3,4]. More specifically, compared with NSS, RN has been found to be an independent risk factor for stage III chronic renal insufficiency (CRI), which carries a substantially increased risk of hospitalization (10%), cardiovascular events (40%) and death (20%) [4–6].
RN potentially predisposes to metabolic derangements more than NSS, as additional normally functioning parenchyma is excised in RN, and the development CRI also causes biochemical imbalances . Diminished renal capacity might lead to retention of the endogenous and dietary renal acid load, contributing to metabolic acidosis (MA). Further, there is a reduction in new renal bicarbonate generation coupled with possible decreases in bicarbonate absorption, worsening the MA .
We examined the incidence of and risk factors for the development of CRI, MA, and proteinuria in patients who were treated with RN or NSS.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
In an institutional review board-approved retrospective study, we reviewed 749 patients (485 men, 264 women; 421 Caucasian, 328 African-American or other; mean age 57.9 years; mean follow-up 6.3 years) who had RN or NSS for parenchymal renal tumours at our centre from July 1987 to June 2006. Patients with urothelial tumours, a solitary kidney or incomplete records and who had RN or NSS were excluded.
The occurrence of CRI, defined as an estimated GFR (eGFR) of <60 mL/min/1.73 m2 (‘low’ eGFR), was defined as the primary endpoint, with secondary endpoints being the development of a creatinine level of ≥2.0 mg/dL (‘high’), MA (serum HCO3- of <22 mmol/L), and proteinuria (dipstick urine analysis of ‘≥1+’ protein). The eGFR was calculated using the Modification of Diet in Renal Disease (MDRD) Study Group equation: eGFR = 186 × serum creatinine−1.154 × age−0.203 × (1.21 if Black) × (0.742 if female) . Demographics, i.e. age, sex, race, body mass index (BMI); preoperative history of diabetes mellitus (DM), hypertension and smoking; disease characteristics (stage, pathology); outcomes, i.e. a low eGFR, high serum creatinine level, MA, and their time to development, and proteinuria, were recorded. We used the 2002 TNM stage classification system . We also analysed the subgroup of patients who did not have pre-existing high creatinine level, to elucidate the rate of development of CRI and MA.
The results were analysed statistically to describe factors that were significantly associated with, and/or predictive for, the development of a low eGFR, high serum creatinine or MA events after RN or NSS. Data were analysed using univariate and multivariate statistics with all potential explanatory variables included in the multivariate model. Independent variables were included in the regression models if P ≤ 0.10 in the univariate analysis. Linear regression and multivariate stepwise logistic regression statistics, using the forward procedure, were used for multivariate analysis, whereby all variables identified on univariate regression were incorporated into the model as independent variables, with the primary endpoints as the dependent variable (patients with pre-existing high serum creatinine or MA were excluded from the respective multivariate analyses). Descriptive statistics were analysed within subgroups based on whether patients had RN or NSS. Means were compared between the groups (RN vs NSS) using Student’s t-test and chi-squared tests for continuous and categorical variables, respectively. All P values were based on two-sided tests of significance, with P < 0.05 considered to indicate statistical significance.
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Table 1 shows the patient and tumour characteristics; 499 patients had RN and 250 NSS, and there were no significant differences in mean follow-up, age, BMI, race and sex distribution, history of hypertension, smoking or DM between the groups. Tumours were significantly larger in the RN group (7.1 vs 3.8 cm, P < 0.001). There was a significantly greater proportion of pT1a tumours in the NSS group, whereas there was a significantly greater proportion of pT1b, pT2, pT3a, pT3b, pT3c and pT4 tumours in the RN group. There was a significantly greater proportion of benign tumours in the NSS (14.4%) than in the RN group (5.4%, P < 0.001). There were similar proportions between the groups for clear cell, papillary and oncocytoma subtypes, but significantly greater proportions of angiomyolipoma and chromophobe tumours in the NSS group.
Table 1. The patient and tumour characteristics, and the renal functional outcome in all patients and in the subgroup with a high serum creatinine level before surgery
|Variable, mean (sd) or %||RN||NSS||P|
|Age, years|| 57.4 (15.5)|| 58.2 (16.0)|| 0.476|
|Gender (M/F)|| 66.5/33.5|| 61.0/39.0|| 0.624|
|Race (Caucasian vs African-American/other)|| 56.1/43.9|| 56.4/43.6|| 0.940|
|BMI, kg/m2|| 27.7 (5.0)|| 27.4 (5.4)|| 0.567|
|DM|| 21.4|| 26.4|| 0.129|
|History of smoking|| 61.7|| 66.8|| 0.174|
|Hypertension|| 60.9|| 63.6|| 0.477|
|Tumour size, cm|| 7.1 (3.0)|| 3.8 (2.0)||<0.001|
|Pathological pT Stage|| || || |
| pT1a|| 18.0|| 55.6||<0.001|
| pT1b|| 32.7|| 22.4||<0.001|
| pT2|| 25.5|| 5.6||<0.001|
| pT3a|| 6.4|| 1.6|| 0.016|
| pT3b|| 7.0|| 0||<0.001|
| pT3c|| 3.0|| 0|| 0.006|
| pT4|| 5.0|| 0.4|| 0.001|
|Benign tumours|| 5.4|| 14.4||<0.001|
|Pathology|| || || |
| Clear cell RCC|| 82.1|| 70.4|| 0.457|
| Papillary RCC|| 11.6|| 12.0|| 0.327|
| Chromophobe RCC|| 0.2|| 2.4|| 0.003|
| Oncocytoma|| 5.0|| 6.4|| 0.431|
| Angiomyolipoma|| 0.4|| 8.4||<0.001|
|Renal function before/after surgery|
|Low eGFR|| 11.2|| 14.4|| 0.212|
| 44.7|| 16.0||<0.001|
|High creatinine|| 3.2|| 5.6|| 0.115|
| 14.2|| 8.4|| 0.022|
|Proteinuria|| 7.2|| 10.0|| 0.189|
| 22.2|| 13.2|| 0.003|
|MA|| 1.6|| 4.8|| 0.010|
| 12.8|| 7.2|| 0.020|
|Subgroup; creatinine >2 mg/dL before surgery|| || || |
|No. of patients||484||235|| |
|Low eGFR|| 8.7|| 9.4|| 0.762|
| 43.0|| 11.6||<0.001|
|MA|| 0.8|| 3.4|| 0.011|
| 10.8|| 5.0|| 0.007|
|Proteinuria|| 6.2|| 5.1|| 0.559|
| 20.5|| 8.1||<0.001|
|High serum creatinine|| 11.4|| 3.0||<0.001|
|Mean time, months, to development of:|
| Low eGFR|| 20.2|| 39.4||<0.001|
| MA|| 38.6|| 51.7|| 0.012|
| High creatinine|| 37.1|| 56.6||<0.001|
Table 1 also shows the renal functional outcomes for all patients; there were no differences between the groups in preoperative low eGFR, high serum creatinine and proteinuria. A significantly greater proportion of patients in the NSS group had pre-existing MA (4.8% vs 1.6%, P = 0.01) than in the RN group. After surgery a significantly greater proportion in the RN than NSS group had a low eGFR, MA, proteinuria and high serum creatinine.
Table 1 also shows the subgroup analysis of renal functional outcomes in patients with a high preoperative serum creatinine level (484 RN and 235 NSS). There were no significant preoperative differences between RN and NSS groups in the proportion with proteinuria or a low eGFR, but a significantly greater proportion of patients in the NSS group had pre-existing MA. After surgery, a significantly greater proportion in the RN than the NSS groups had a low eGFR, proteinuria and MA, and developed de novo high serum creatinine. In patients with de novo development of a low eGFR, high serum Creatinine and MA, the time to development of these conditions was significantly longer in the NSS group than in the RN group (Table 1).
Table 2 shows the multivariate analyses for factors associated with the development of a low eGFR, high serum creatinine and MA; those associated with the development of low eGFR included DM, RN, hypertension, BMI ≥30 kg/m2, age ≥60 years and smoking. Factors associated with the development of high serum creatinine were age ≥60 years, DM, hypertension, smoking and RN. Factors associated with the development of MA after surgery were male gender, age ≥60 years, BMI ≥30, RN, preoperative low eGFR and high serum creatinine.
Table 2. Multivariate analysis of factors associated with the development of a low eGFR, high serum creatinine and MA after surgery
|Variable||Odds ratio (95% CI)||P|
|History of DM (yes vs no)||8.96 (2.21–36.39)|| 0.002|
|Procedure (RN vs NSS)||5.32 (2.42–11.67)||<0.001|
|History of hypertension (yes vs no)||4.55 (1.70–12.17)|| 0.003|
|BMI ≥30 kg/m2 (yes vs no)||3.51 (1.25–9.89)|| 0.017|
|Age (≥60 vs <60 years)||2.91 (1.51–5.62)|| 0.002|
|Smoking history (yes vs no)||2.44 (1.2–4.96)|| 0.014|
|High serum creatinine|
|Preop low eGFR||18.99 (7.81–46.19)||<0.001|
|History of DM (yes vs no)||10.00 (5.43–18.51)||<0.001|
|History of hypertension (yes vs no)||7.41 (2.13–25.64)|| 0.002|
|Smoking history (yes vs no)||5.29 (2.58–10.87)||<0.001|
|Procedure (RN vs NSS)||3.08 (1.64–5.78)||<0.001|
|Age (≥60 vs <60 years)||2.00 (1.12–3.55)|| 0.019|
|Gender (M vs F)||2.50 (1.17–5.33)|| 0.019|
|Age ≥60 years (yes vs no)||3.13 (1.54–6.67)|| 0.002|
|BMI ≥30 (yes vs no)||3.52 (1.84–6.74)||<0.001|
|Procedure (RN vs NSS)||9.82 (3.50–27.58)||<0.001|
|Preop low eGFR (yes vs no)||9.71 (4.40–21.44)||<0.001|
|Preop high creatinine (yes vs no)||5.92 (1.60–21.85)|| 0.008|
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
NSS has emerged as the preferred option for managing small renal tumours [3,4,10], comparing favourably with RN from the standpoint of long-term oncological control [3,4,11]. For renal function, NSS is associated with improved outcomes in both imperative cases and in the case of a normal contralateral kidney . Meanwhile, the adverse consequences of RN have been increasingly realised. In patients with a normal preoperative serum creatinine level and a normal contralateral kidney, RN is a significant risk factor for developing new onset CRI [4,6,12]. Although risk of progression to end-stage renal disease and dialysis-dependence might be low for such patients , additional significant complications of CRI include hypertension, anaemia, malnutrition, neuropathy, reduced quality of life, and increased risk of cardiovascular disease and mortality [5,13,14].
The association between CRI and MA is also well established [15–17], although the exact prevalence of MA in CRI is unknown. Severe CRI can lead to MA as a consequence of insufficient renal tubular bicarbonate reabsorption or abnormal bicarbonate production. The acidosis resulting from CRI is generally mild to moderate (plasma bicarbonate 12–22 mmol/L) [16,17], however, even mild degrees of acidosis can have significant adverse consequences , including muscle-wasting, hypoalbuminaemia, bone disease, impaired insulin sensitivity, and abnormalities in thyroid function and basic metabolic rate. Early studies of MA in CRI suggested that MA occurs when severe CRI develops (GFR ≈20 mL/min or less) [7,19]. However, recent evidence suggests that advanced age (>60 years) and lesser degrees of CRI are also associated with the development of MA [20,21]. Given these associations, we hypothesized that significant rates of MA might be detectable in patients after RN or NSS.
In the present contemporary review of 749 of well-matched patients who had RN or NSS (Table 1) there were significant differences in the incidence of a high creatinine level and low eGFR after RN and NSS (Table 1). Both of these criteria have been used as thresholds for moderate CRI, although the National Kidney Foundation reported that actual or eGFR values are more accurate reflections of kidney function than serum creatinine level . For the development of a high creatinine level (14.2% for RN and 8.4% for NSS, P = 0.022, Table 1) and low eGFR (44.7% for RN vs 16.0% for NSS, P < 0.001, Table 1), our results are similar to previously published reports. Using the MDRD equation and an eGFR criterion of <60, Lucas et al. reported that 21 of 36 patients (60%) undergoing RN developed de novo CRI, vs 12 of 98 (12%) undergoing nephron-sparing approaches. Similarly, the Memorial-Sloan Kettering group  reported a 69% (142/204) incidence of de novo CRI (using an eGFR of <60 as the criterion) in the RN group, vs 17% (50/287) in the NSS group. The incidence of new-onset CRI using an eGFR of <60 in the present analysis corroborates the increasing evidence showing greater preserved renal function after NSS than RN. We also assessed the incidence of new-onset CRI using a serum creatinine threshold of ≥2.0 mg/dL, similar to Lau et al.. With this criterion, the Mayo group found that the 10-year incidence of CRI was significantly higher in the RN than the NSS group (22.4% vs 11.6% at 10 years, risk ratio 3.7, 95% CI 1.2–11.2, P = 0.01). In the present study there was a similar trend for the incidence of CRI (14.2% for RN and 8.4% for NSS, P = 0.022). Both of the present analyses of renal function, using serum creatinine or eGFR, showed decreased rates of de novo moderate CRI in patients who were treated with NSS vs RN. Substantially more patients received a designation of having moderate CRI when using eGFR thresholds rather than serum creatinine criterion; this discrepancy was also noted by Herts et al., who found that nearly 2.5 times as many patients were classified as having moderate CRI when using eGFR rather than serum creatinine.
A novel finding in the present series is the identification of significant differences in rates of de novo MA between the RN and NSS groups (Table 1). To our knowledge, there are no previous reports of this kind. It is striking that patients with renal tumours treated with RN or NSS had higher rates of MA than would be expected based on degrees of CRI. Most patients who developed de novo CRI also developed MA, while in nephrology reports based on patients with medical renal disease, MA is suggested to be associated only with severe CRI. More recently, a mechanism implicating MA in exacerbating declining GFR through the endothelin A receptor has been outlined, suggesting that MA actually has a substantial role in exacerbating CRI . Theoretical reasons for the increased prevalence of MA in the present series include the influence of increased dietary acid load in these patients. It is also possible that the advanced age of the present patients (mean age 57.7 years) made them more susceptible to the development of MA [20,21]. Although the exact mechanism is yet to be elucidated, NSS was associated with a significantly lower incidence of de novo MA than was RN.
The present study showed significantly higher rates of proteinuria in the RN than the NSS group (22.2% vs 13.2%, P = 0.003, Table 1). Qualitative dipstick proteinuria is an important clinical finding with long-term implications. Ishani et al. found that patients with a qualitative dipstick proteinuria reading of 1+ (hazard ratio 3.1, 95% CI 1.78–5.42) or 2+ (15.7, 10.33–23.87) have a higher risk of developing end-stage renal disease than those with a reading of negative/trace proteinuria. Further, this risk is magnified if the eGFR is low . Similar results were reported by others, who also showed worse long-term clinical outcomes when patients were stratified by qualitative urine protein content . Admittedly, qualitative urine dipstick analysis is not the most sensitive measure of assessing proteinuria. Lau et al. used a quantitative protein/osmolality ratio (proteinuria = ratio >0.12), which is more accurate than qualitative urine analysis and provides an estimate of 24-h protein levels, to compare proteinuria in patients receiving RN and NSS. They found a significantly higher risk of proteinuria in patients having RN than NSS (55.2% vs 34.5%, P = 0.01). The present study indicates that RN predisposes to proteinuria more so than NSS, which is associated with adverse renal and cardiovascular consequences. A dipstick analysis during preoperative planning might be an important marker when considering operative strategy, and even such a simple test should prompt the urologist beforehand to explore the option of NSS.
RN (vs NSS) was identified as a risk factor for developing a low eGFR (odds ratio 5.32, P < 0.001), high creatinine (odds ratio 3.08, P = 0.001) and MA (odds ratio 9.92, P < 0.001) in the multivariate analyses (Table 2). Lau et al. also described a greater risk of creatinine failure in patients who had RN than in a matched group receiving NSS. Huang et al. identified patients who had RN (vs NSS) as having a greater risk of developing CRI, defined as an eGFR of <60 (hazard ratio 3.82, 95% CI 2.750–5.32, P < 0.001). Glomerulosclerosis and subsequent declining renal function in the remnant kidney was previously reported in patients with a solitary kidney receiving NSS; furthermore, the decline in renal function was shown to depend on the amount of kidney excised [27,28]. This same mechanism of hyperfiltration, glomerulosclerosis and renal insufficiency is likely responsible for worse renal function in patients who have RN, as there is a lower total of nephrons to disperse the additional filtered load. DM, hypertension, smoking, and age ≥60 years were also identified as risk factors for a low eGFR and high creatinine (Table 2). In a prospective analysis of >23 000 individuals, these same variables were identified as risk factors for developing CRI . Huang et al. also identified hypertension, increased comorbidity (Charlson-Romano Index ≥2), and advancing age, in addition to RN, as risk factors for developing stage III CRI. In the present multivariate analysis outlining risk factors for developing MA, a low preoperative eGFR and high serum creatinine were associated with a higher risk of developing MA, presumably due to the well-described mechanism of CRI . There is also an inverse correlation between age and plasma bicarbonate , which was also the case in the present analysis. We also found obesity to be a risk factor for developing postoperative MA. This is attributable to hypertension, atherosclerosis and other kidney-damaging changes associated with obesity, and to increased amounts of serum free-fatty acids in obese individuals [30,31].
The preservation of GFR and maintenance of acid-base homeostasis are useful measures of the merits of NSS. However, the ultimate measures include the effect on dialysis rates, cardiovascular morbidity and patient survival. Although we did not evaluate these endpoints, recent studies have advanced the discussion on the ultimate effect of NSS. In a retrospective cohort study of >10 800 patients from the Surveillance, Epidemiology and End Results database, NSS was associated with significantly lower rates of dialysis and renal transplantation than was RN . However, the incidence of adverse cardiovascular outcomes did not differ by treatment . Thompson et al. compared overall survival between patients with T1a tumours treated with RN and NSS. In a subset analysis of 327 patients aged <65 years, RN was associated with a lower overall survival than NSS . These studies address critical aspects of the ultimate impact of NSS; further study of these long-term endpoints is required.
The weaknesses of our study include the retrospective design, including inherent biases and limitations. The use of the abbreviated MDRD equation to estimate GFR is an additional limitation. GFR was not measured in the follow-up algorithm for our patients, and equations for estimating GFR introduce a degree of inaccuracy. Nevertheless, the MDRD equation is widely used in other studies to evaluate renal functional outcomes after RN and NSS. Despite these limitations, our findings are supported by inclusion of many patients with an intermediate-term follow-up. To our knowledge, this study represents the first attempt to measure the effects of renal tumours and surgical management on long-term renal acid-base homeostasis. Our findings show that the surgical management of renal tumours significantly affects the development of proteinuria and acid-base homeostasis, and that NSS can mitigate this effect to a significant degree. Further study is needed to corroborate our findings and further clarify the nature of these associations.
In conclusion, patients who had RN had a significantly higher incidence of CRI, proteinuria and MA than a contemporary well-matched cohort treated with NSS. In addition to RN, age ≥60 years, DM, hypertension and smoking history are significantly associated with progression to moderate-severe CRI after surgery.