A clinical score to predict recovery in end-stage kidney disease due to acute kidney injury.

Abstract

Acute kidney injury (AKI) is a major contributor to end-stage kidney disease (ESKD). About one-third of patients with ESKD due to AKI recover kidney function. However, the inability to accurately predict recovery leads to improper triage of clinical monitoring and impacts the quality of care in ESKD. Using data from the United States Renal Data System from 2005 to 2014 (n=22922), we developed a clinical score to predict kidney recovery within 90 days and within 12 months after dialysis initiation in patients with ESKD due to AKI. Multivariable logistic regressions were used to examine the effect of various covariates on the primary outcome of kidney recovery to develop the scoring system. The resulting logistic parameter estimates were transformed into integer point totals by doubling and rounding the estimates. Internal validation was performed. Twenty-four percent and 34% of patients with ESKD due to AKI recovered kidney function within 90 days and 12 months, respectively. Factors contributing to points in the two scoring systems were similar but not identical, and included age, race/ethnicity, body mass index, congestive heart failure, cancer, amputation, functional status, hemoglobin and prior nephrology care. Three score categories of increasing recovery were formed: low score (0-6), medium score (7-9) and high score (10-12), which exhibited 90-day recovery rates of 12%, 26% and 57%. For the 12-month scores, the low, medium and high groups consisted of scores 0-5, 6-8 and 9-11, with 12-month recovery rates of 16%, 33% and 62%, respectively. The internal validation assessment showed no overfitting of the models. A clinical score derived from information available at incident dialysis predicts renal recovery at 90 days and 12 months in patients with presumed ESKD due to AKI. The score can help triage appropriate monitoring to facilitate recovery and begin planning long-term dialysis care for others.