The Impact of Ischemia on Long-Term Renal Function After Partial Nephrectomy in the Two Kidney Model

Introduction

There has been a strong impetus in recent years to avoid clamping the renal hilum during partial nephrectomy (PN). While the concept is supported by historic canine studies on warm ischemia, rigorous scientific data has yet to demonstrate meaningful ischemic dose-response curves in humans. Despite the absence of high-quality data, many authors have zealously advocated for off-clamp PN (OFF-PN) in the two kidney model. Their arguments are largely based on the findings from small, retrospective institutional cohort studies that show improved perioperative renal function parameters in the short term. ^{1 –6} These reports, however, are fraught with major methodologic limitations and are unsuitable for broadly applicable, definitive conclusions.

To complicate matters, the results from a recent phase III randomized clinical trial have suggested worse overall survival and a nonsignificant increase in the risk of cancer-related death for patients undergoing PN compared with radical nephrectomy. ⁷ From this study, one could surmise that if the benefits from a renal preservation standpoint were, in fact, that clinically meaningful, the results of the trial should have been dramatically different; in other words, one would have expected the difference to favor PN. Because removing the entire kidney does not appear to adversely affect long-term clinical outcomes, one could postulate that merely clamping the kidney for a short period would have negligible effects as well.

In this context, we aim to explore the long-term consequences of on-clamp PN (ON-PN) compared with OFF-PN in terms of renal function during robot-assisted, laparoscopic, and open PN. We compare differences in estimated glomerular filtration rate (eGFR) as well as change in differential function on renal scan between ON-PN and OFF-PN groups. We also performed a post-hoc sensitivity analysis using the same comparisons after stratifying by age.

Patients and Methods

A Mayo Clinic Institutional Review Board approved retrospective analysis was performed of open, laparoscopic, and robot-assisted PN performed at our institution from November 1999 to July 2013. Of the 607 patients reviewed, 331 were included in the analysis after excluding those who underwent conversion to radical nephrectomy, patients with solitary kidneys, those who had their most recent creatinine level obtained less than 90 days after their surgery, patients who had multiple masses resected, and cases in which clamping was performed but no ischemia time was recorded.

Patient demographics, perioperative values, and oncologic outcomes were analyzed, including age, sex, body mass index (BMI), operative time, tumor size, renal scans, and preoperative and most recently recorded eGFR values using serum creatinine levels. Because of its increased accuracy for estimating eGFR above 60 mL/min/1.73 m², these values were calculated using the Chronic Kidney Disease-Epidemiology Collaboration formula. ^8,9 Mercaptoacetyltriglycine-3 renal scans were examined to determine the percent change in the ipsilateral kidney's function after PN, with 15 patients in the nonclamping group who had complete renal scan data and 115 in the clamping group. Although in some cases there were variable types of clamping, including total and selective regional vessels, clamp time was defined as the amount of time that any vessels were clamped.

Data are presented as means and standard deviations for continuous variables, and as frequencies and percentages for categoric variables. Wilcoxon rank sum tests were used to compute the P values presented in the tables, and statistical significance was defined as a two-sided P value of 0.05. Statistics were calculated using Stata v12MP^® for Windows 7.

Results

Of the 331 patients included in the study, 236 underwent ON-PN while 95 underwent OFF-PN. Half (18) of the open PNs that experienced clamping were performed using cold ischemia, while the others were clamped without. Nine of the patients experienced only selective hilar clamping, and 18 patients (14 in the robot-assisted group and 4 in the laparoscopic group) had a combination of total and selective ischemia. Some variation was noted in terms of patient demographics, because those patients who underwent ON-PN tended to be younger at the time of surgery, have larger tumors and higher R.E.N.A.L. (radius; exophytic/endophytic; nearness; anterior/posterior; location) nephrometry scores, and higher preoperative eGFRs (Table 1). This cohort also had a longer mean operative time, lower American Society of Anesthesiologists score, and shorter length of follow-up after surgery (Table 2).

In terms of surgical approach, a higher percentage of laparoscopic and robotic cases were clamped (68% and 87.5%, respectively) while only 53% of open cases were clamped. Factors such as BMI and length of hospital stay were not statistically different between the patients who experienced ischemia and those who did not.

On analyzing mean change in eGFRs from preoperative and most recently recorded serum creatinine levels, it was found that the mean change in eGFR was the same for patients in both groups (−6.07 mL/min/1.73 m² for ON-PN and −6 mL/min/1.73 m² for OFF-PN) (Table 2). On examination of postoperative renal scans, the differential renal function of the affected kidney was not found to be statistically different between the two groups (−0.77±8.35% change for ON-PN and −1.1±3.95% change for OFF-PN).

To control for differences in type of ischemia, we repeated the analysis of change in eGFR with the 45 patients who received either cold ischemia or selective hilar clamping removed (Table 3). Again, there was found no significant difference between the ON-PN and OFF-PN cohorts (P=0.87). Also, because of the large range in clamp times, we stratified the clamping cohort into two groups: Those clamped for under 30 minutes (n=179) and those clamped for 30 minutes or longer (n=57) (Table 4). No significant differences in change in eGFR were found between the two groups (P=0.95) nor when comparing zero ischemia to ischemia times under 30 minutes (P=0.66) or 30 minutes or longer (P=0.86).

In an effort to address the potential effect of age on the long-term renal function outcomes, we performed a post-hoc sensitivity analysis of eGFR stratified by age at time of surgery. We stratified the patients into three similar-sized cohorts based on tertiles of age in years (Table 5). Because the OFF-PN group had a longer mean follow-up time than the clamping group, new creatinine values were obtained for OFF-PN patients who were similar in follow-up time to ON-PN. Once the new eGFRs were calculated, comparisons were obtained within the three age groups, and none of the mean changes in eGFR was found to be statistically significant.

Discussion

To summarize the findings of this study, we demonstrate that there is no apparent long-term renal dysfunction in patients undergoing ON-PN, at least compared with patients undergoing OFF-PN. This is in direct contradistinction to the previously published work from some authors who insist that a zero-ischemia approach is superior. ^1,2 Our study, however, which is one of the largest of its kind with the longest follow-up, corroborates the findings in other studies that have shown that ON-PN is not only an acceptable approach in the two kidney model, but has relatively little effect on the long-term performance of the kidney. ^{10 –13} In this analysis, no significant difference was found in the long-term mean change in eGFR between the clamping and nonclamping cohorts. Based on this, one could argue that on-clamp approaches to PN in the two kidney model have equivalent long-term renal function outcomes.

Several other studies have likewise demonstrated that on-clamp approaches have minimal impact on renal function, at least in the short-term. In a study by Koo and colleagues, ¹⁴ 11 nonclamped patients were compared with 10 clamped patients during PN. They found no difference in the estimated blood loss, perioperative complications, or mean postoperative creatinine levels. Bhayani and coworkers ¹³ reported that there were no significant differences in the postoperative mean change in eGFR between the three approaches to PN: Off-clamp (42 patients), on-clamp up to 30 minutes (48 patients), and on-clamp for more than 30 minutes (28 patients). ¹³ In an analysis by Lee and associates, ¹⁵ 369 patients who experienced warm ischemia, cold ischemia, or off-clamp approaches were compared, and no difference in 1-year mean change in eGFR was observed.

In the present study, which has the longest mean follow-up of any published report on the matter, we found that the mean change in eGFR was not significantly different between the on- and off-clamp cohorts, with the on-clamp group experiencing a mean change of −6.07 mL/min/1.73 m² and the off-clamp group with a change of −6 mL/min/1.73 m². The mean follow-up time for the most recently recorded serum creatinine level was 757 days for the clamping group and 1373 days for the nonclamping. Taken together, it seems as though clamping for a short period (usually less than 30 minutes) is inconsequential to the long-term renal function of a patient with two kidneys.

This hypothesis is further supported, at least conceptually, by the findings of a randomized, phase III trial published by the European Organisation for Research and Treatment of Cancer (EORTC). ⁷ This trial demonstrated that PN did not significantly improve overall or cancer-specific survival for patients with tumors less than 5 cm in comparison with radical nephrectomy. These data were surprising, because one would expect that preserving renal parenchyma would mitigate, to some degree, future adverse clinical outcomes. Survival was worse in the PN group, however, which illustrates the difficulty of accounting for potential biases incumbent to nonrandomized comparisons because many of the previous retrospective studies before the publication to the EORTC trial demonstrated that PN was associated with improved overall and cancer-specific survival.

This selection bias certainly exists in the present study as well. Close examination of the data in the present study, however, reveals that, with the exception of age, the cohort demographics and clinical characteristics would, if anything, favor the OFF-PN group (smaller, less complex tumors in patients with lower BMIs). As such, we attempted to account for the potential effect of age on the interpretation of the data and performed a secondary analysis of the eGFR and differential function on renal scan after stratifying by tertiles of age. In many respects, stratification is a preferred method for addressing confounding because it avoids many of the assumptions and challenges associated with model-based statistics (which variables to include in the model, how to handle collinearity, whether risk is evenly distributed in the controlled factor, etc). In any case, even after stratification, no apparent difference in long-term renal function was noted, which further supports the notion that ON-PN (with appropriate time-constraints) may not adversely affect a patient's long-term renal function.

Because our cohorts are heterogenous, we performed additional analysis to control for potential confounders. In addition to zero ischemia techniques, superselective clamping has been touted as preserving renal function over total hilar clamping. ¹⁶ While we have published on selective vs total clamping previously ¹² demonstrating no observable benefit to selective clamping, we did additional analysis by excluding selectively clamped patients and cold ischemia patients. As demonstrated in Table 3, there was no statistically significant difference with these patients excluded. Finally, we further stratified the analysis by length of ischemia (≥30 min vs <30 min) because of the historical assumption that 30 minutes is the cutoff for permanent damage. Again, as seen in Table 4, no statistically significant difference was demonstrated between either group or when the groups were compared with the off-clamp cohort separately.

Despite these strengths, these findings need to be interpreted in the context of certain limitations. First, as mentioned, the analysis is retrospective and observational, and because of its study design has certain inalienable drawbacks and biases (such as incomplete data, confounding from other unknown variables, etc). Second, this is a single institutional case series and may not be representative of the entire population of PN patients in the United States or elsewhere. Third, different surgeons were included in the dataset, which may have influenced the selection of patients for a particular approach. The choice of approach (open, laparoscopic, or robotic) and the decision to clamp the kidney were also surgeon dependent. Fourth and finally, no consideration was afforded for the type of ischemia used (i.e., no differentiation was made between warm and cold or selective and total clamping of the kidneys).

Despite these limitations, we believe that these data uniquely inform the understanding of the late effects of clamping during PN. Because the late effects of renal function were minimal in this study, one could argue against zero ischemia or superselective approaches in patients with two kidneys on the basis of increased perioperative risk of hemorrhage alone, which has been demonstrated in reports. Certainly, avoiding ischemia to any organ is ideal, and if OFF-PN is a safe option, then it is reasonable. It is important to recognize, however, that attempts at sparing a few minutes of ischemia that may result in a complication such as hemorrhage may have more dire consequences than a few points of eGFR in the two kidney model. Surgeons should feel comfortable, especially early in their experience with minimally invasive PN, clamping the renal hilum if it is in the best interest of the patient and perioperative outcomes.

Conclusions

In the two kidney model, warm ischemia does not appear to affect long-term renal function outcomes after PN. These data provide evidence that ON-PN regardless of surgical approach is perfectly acceptable in the appropriately selected patient with two kidneys. Heroic measures to minimize ischemia, which may compromise an otherwise safe surgical technique, are unlikely to be of long-term benefit.