Abstract: PD73-02
Sources of Funding: None

Introduction

Existing pre-operative measures of kidney function have limitations. We utilize clinical and histopathologic changes in the non-neoplastic kidney (NNK) to predict post-operative renal function one and two years post radical nephrectomy (RN) for renal cell carcinoma (RCC). Patients were followed for up to 5 years.

Methods

This retrospective review included all patients undergoing RN for RCC measuring <10 cm with a normal contralateral kidney between January 1, 2011 - May 1, 2015. All slides were independently re-reviewed for histopathologic changes using the Banff 97 criteria by two blinded nephropathologists for any chronic glomerular (G), tubulointerstitial (IFTA), arterial (Art), and arteriolar changes (Arl). Univariate analyses were conducted with Spearman correlation coefficients and ANOVA. Factors significantly associated with post-operative eGFR were entered into multivariable regression (MVR) models. Separate regression models were created for the clinical and histological factors. Predictive performance was evaluated with the r-square statistic. Estimated glomerular rate (eGFR) was calculated using chronic kidney disease epidemiology collaboration formula.

Results

The 167 patients had a mean age, Charlson comorbidity score, and tumor size of 61, 2.6, and 6.2cm, respectively. The group consisted of diabetics (26%), hypertensives (60%), and smokers (35%). Mean pre-op eGFR and 24-month eGFR were 78 and 51mL/min/1.73m², respectively. Severe histopathologic changes were discovered in 11% of patients despite normal pre-operative eGFRs. New kidney disease was common (54%); 8 patients progressed to dialysis and 20 died during follow-up. NNK changes were significantly associated post-op eGFR at 1 year (R2=0.52) and 2 years (R2=0.5). Severe G and IFTA were associated with an independent eGFR change of decline of -28 (p<0.001) and -16 mL/min/1.73m² (p=0.0005). Art (p=0.08) and Arl (p=0.387) changes were not associated with eGFR changes. Age (p<0.0001), body mass index (p=0.0039), and age-adjusted Charlson co-morbidity (p=0.0434) were significantly associated with post-op eGFR. Using MVR modeling a calculator was developed and accurately predicted post-op eGFR.

Conclusions

Using age, glomerular, and tubulointerstitial histopathologic variables alone, the post-operative eGFR can be accurately calculated. This calculator may guide studies on subsequent therapeutic intervention following radical or partial nephrectomy.

Funding

None