Does Glomerular Filtration Rate at Discharge After Partial Nephrectomy Predict Long-Term Glomerular Filtration Rate Stability?

Introduction

Reliable estimation of long-term renal function after partial nephrectomy (Pnx) is an important piece of information to provide patients undergoing surgery for small renal masses. The clinical utility of glomerular filtration rate (GFR) as a measure of renal function is valuable given its association with hospitalizations, cardiovascular events, cardiovascular mortality, and all-cause mortality ^{1 –3} in patients with estimated GFR (eGFR) <60 mL/(minute ·1.73 m²). Although Pnx is known to preserve postoperative (post-op) renal function more so than radical nephrectomy, loss of renal function after Pnx is still a major concern for patients and treating physicians alike. ^{4 –6} Approximately one-third of patients undergoing Pnx with a preoperative (pre-op) GFR >60 mL/(minute ·1.73 m²) will experience a decline in renal function to chronic kidney disease (CKD) Stage III after Pnx. ^4,7

Prior studies have sought to investigate factors such as preserved parenchyma, pre-op GFR, and post-op GFR to see if they could predict long-term renal function after Pnx. ^{8 –13} A 2008 study by Lane and colleagues analyzing GFR for a 10-year period after nephrectomy (partial or radical) suggested that long-term renal function is associated with GFR measured in the immediate post-op period. ⁸

Factors affecting baseline renal function, among many others, include hypertension (HTN) and diabetes mellitus (DM), both of which interestingly have been shown to influence GFR recovery after radical nephrectomy. ¹⁴ A study published in 2016 found DM status to be an independent predictor of renal function at 1 year after radical nephrectomy. ¹⁵ In addition, HTN and DM have been shown to negatively impact GFR postradical nephrectomy, and the effect of such comorbidities on GFR is sustained at 12 months. ¹⁴ A clear understanding of renal function trajectory after Pnx is paramount to an effective informed consent discussion in the process of deciding intervention. ¹⁶

As such, in an effort to understand renal function trajectory in our highly comorbid inner city population, we compared the relationship between baseline and day of discharge (DC) GFR to that of long-term GFR. We were interested in understanding long-term GFR stability for patients undergoing Pnx in hope of being able to provide future patients with a clear projection of renal function trajectory in the post-op period. We also evaluated other clinical and pathological factors that might demonstrate important influences in this relationship. Our ethnically diverse cohort is representative of an inner-city patient population, one that is often underrepresented in similar studies. In addition, the high prevalence of comorbidities in our cohort allowed us to further explore the effects of HTN and DM on renal functional recovery.

Methods

The patient cohort included in this study was generated through Clinical Looking Glass (CLG) software, a database query program developed at our institution. Patients who underwent minimally invasive Pnx at Montefiore Medical Center between January 2010 and December 2016 were gathered into a cohort, retrospectively reviewed, and added to our IRB-approved database. All Pnx were transperitoneal with consistent use of standard bulldog clamp technique. Inclusion criteria consisted of having all of the following measurements available: GFR at pre-op, at DC, and 12–18 months after DC, referred to as “long-term.” eGFR values were calculated using the Modification of Diet in Renal Disease equation. ¹⁷

Multiple paired t-tests were performed to evaluate differences in means between variables of interest. Multivariate linear regression was performed to assess factors predictive of GFR after DC between 12 and 18 months. These factors included estimated blood loss (EBL), warm ischemic time (WIT), tumor size, length of stay, pre-op GFR, DC GFR, race, CKD, DM, and HTN. In our cohort, CKD was defined as an eGFR of <60 mL/(minute ·1.73 m²) on at least two occasions 90 days apart, with or without markers of kidney damage. All statistical tests were performed using IBM's SPSS version 20.

Results

A total of 317 patients who underwent Pnx were identified. In that 155 of these patients were excluded for missing data. In total 162 patients ultimately met inclusion criteria. Demographic and pathological data for our ethnically diverse patient population are described in Table 1. Notably, 37% of patients were African American, 30.9% Hispanic, 21% white, and 11.1% representatives of other ethnicities. The patient population was highly comorbid recording 71% with HTN, 31% with diabetes and 51.3% with obesity. Mean EBL and WIT across all patients were 183.4 ± 269.4 mL and 25.4 ± 10.0 minutes, respectively.

DC GFR measurements were recorded at a mean of 3 ± 1.9 days post-op and long-term GFR measurements were 14 ± 0.46 months postoperatively. Mean pre-op GFR was 70.75 ± 22.95 mL/(minute ·1.73 m²), mean DC GFR was 68.33 ± 27.67 mL/(minute ·1.73 m²), and mean long-term GFR was 66.53 ± 25.46 mL/(minute ·1.73 m²). Mean change in GFR from pre-op to long-term was −4.23 ± 17.17 mL/(minute ·1.73 m²), whereas mean change in GFR from DC GFR to long-term GFR was −1.80 ± 19.78 mL/(minute ·1.73 m²).

As seen in Table 2, there was no significant difference between DC GFR and long-term GFR (68.33 and 66.53 mL/(minute ·1.73 m²), respectively; p = 0.25). No significant difference was found between pre-op GFR and DC GFR (70.75 and 68.33 mL/(minute ·1.73 m²), respectively; p = 0.06). However, a significant difference was observed between pre-op GFR and long-term GFR (70.75 and 66.53 mL/(minute ·1.73 m²), respectively; p = 0.002).

On multivariate linear regression analysis (Table 3), both pre-op GFR (β = 0.53; 95% confidence interval [CI] = 0.26–0.81; p ≤ 0.001) and DC GFR (β = 0.31; 95% CI = 0.09, 0.54; p = 0.007) were found to be strong predictors of long-term GFR (model R ² = 0.61). A comparison of the beta values obtained from regression analysis shows each point of pre-op GFR predicts 0.53 U of long-term GFR, suggesting that pre-op GFR alone is a strong predictor of long-term GFR. Other variables—including sex, age, ethnicity, CKD, DM, HTN, EBL, WIT, tumor size, and length of stay—assessed on the multivariate model were not found to be significant predictors of long-term GFR.

Discussion

Our analysis of 162 patients undergoing Pnx for small renal masses demonstrates that mean GFR measured at time of discharge is not significantly different from mean long-term GFR measured at a mean of 14 months. Furthermore, the results of multivariate analysis show that both pre-op GFR and GFR at the time of discharge are predictive of long-term GFR. This relationship indicates that long-term renal function after Pnx can be approximated closely by either pre-op or DC GFR. These results are supported by a study by Lifshitz and coworkers who found post-op day 1 GFR percent change to independently predict the final GFR percent decline. ¹¹ Similarly, Dawidek and colleagues demonstrated that 6–12-week post-op nuclear renography measurements in conjunction with eGFR did not differ from renal function parameters at 1 year, contributing more specific information on the operated kidney's contribution to renal function trajectory after Pnx. ¹⁸ Two other studies analyzed the predictive value of the long-term “new baseline GFR,” measured as the highest GFR obtained in the 6 weeks after nephrectomy. ^9,19 Highlighting the importance of understanding long-term GFR trajectory, a 2015 study by Lane and coworkers determined a new baseline GFR of <45 mL/(minute ·1.73 m²) significantly predicted all-cause mortality for a period of 10 years. ⁹ In addition, a 2017 study by Zabell and colleagues found long-term new baseline GFR significantly predicted 10-year nonrenal cancer mortality. ¹⁹ Being able to predict one's long-term renal function after Pnx offers physicians an increased level of confidence in being able to discuss surgery as an option with patients undergoing Pnx.

Several studies have evaluated factors that are predictive of long-term GFR function. Many studies have demonstrated pre-op GFR's predictive capabilities, suggesting that it is a clinically useful significant predictor of long-term GFR after nephrectomy. ^{8,10,12,13,15} Our study found that both pre-op and DC GFR were significant predictors of long-term renal function in an ethnically diverse comorbid patient population, suggesting that GFR measurements immediately before and after surgery are clinically meaningful tools for projecting long-term renal function after Pnx.

We believe that a major strength of this study is its diverse patient population with respect to both age and ethnicity, reflecting an inner-city patient population that is often excluded from similar studies. None of the studies cited in this article listed Hispanic patients as a demographic category. The studies by Lane and colleagues in 2008, 2011, and 2015 analyzed patient populations that were roughly 90% white.

Our study has several limitations that should be known. Our long-term evaluation of GFR trajectory included a limited duration of follow-up, with the mean time of 14 months after DC. Additional limitations include the retrospective nature of our study and its restriction to patients from a single institution. The patient population we analyzed may not be representative of all populations, particularly those that have lower incidence of HTN and DM or those with different ethnic compositions.

Conclusion

Long-term GFR at a mean of 14 months after discharge from Pnx can be projected by assessment of either pre-op GFR or GFR measured at time of discharge from the hospital. In an ethnically diverse patient population, GFR experiences a plateau after Pnx, resulting in minimal change in long-term renal function. This is valuable data to present to patients undergoing Pnx.