Limited non-linear impact of warm ischemia time on renal functional decline after partial nephrectomy: a propensity...

To quantificationally illustrate the impact of ischemia time (IT) on renal function decline after partial nephrectomy (PN), especially for patients with compromised baseline renal function (estimated glomerular filtration rate [eGFR] < 90 mL/min/1.73 m²).

Patients undergoing PN during 2014-2021 from a prospectively maintained database were reviewed. Propensity score matching (PSM) was employed to balance the possible covariates between patients with or without baseline compromised renal function. Specifically, the relationship of IT with postoperative renal function was illustrated. Two machine learning methods (logistic least absolute shrinkage and selection operator [LASSO] logistic regression and random forest) were applied to quantify the relative impact of each covariables.

The average drop percent of eGFR was -10.9% (- 12.2%, - 9.0%). Multivariable Cox proportional regression and linear regression analyses identified five risk factors for renal function decline, namely RENAL Nephrometry Score (RNS), age, baseline eGFR, diabetes and IT (all p < 0.05). Specifically, the relationship of IT with postoperative functional decline emerged as non-linear, with an increase from 10-30 min and a plateau afterwards among patients with normal function (eGFR ≥ 90 mL/min/1.73 m²), whereas with an increase from 10 to 20 min and a plateau afterwards among patients with compromised function (eGFR < 90 mL/min/1.73 m²). Furthermore, the coefficient's path and random forest analysis revealed that the top two most important features were RNS and age.

IT exhibits the secondarily non-linear relationship with postoperative renal function decline. Patients with compromised baseline renal function are less tolerant to ischemia damage. The use of a single cut-off interval of IT in the setting of PN is flawed.