Comparison of prognostic value of two kinds of severity scoring systems for hospital mortality prediction of elderly patients with acute kidney injury

Aging Clin Exp Res. 2012 Feb;24(1):74-8. doi: 10.3275/7624. Epub 2011 Apr 4.


Background and aims: Outcome prediction is important in clinical practice. Despite significant improvements in therapeutics, the mortality associated with acute kidney injury (AKI) in elderly patients remains high. Several severity scoring systems have been used in hospital mortality prediction of patients, but little is known of their significance in elderly patients with AKI. The aim of this study is to evaluate the ability of version II of Acute Physiology and Chronic Health Evaluation (APACHE II) and Acute Tubular Necrosis Individual Severity Index (ATN-ISI) on predicting the hospital mortality of elderly patients with AKI.

Methods: A consecutive sample of 99 elderly patients (age≥65) with AKI in a university hospital was enrolled. Receiver operating characteristic analyses were used to assess the discriminative power for hospital mortality prediction. The McNemar and Kappa tests were also applied.

Results: The areas under the receiver operating characteristic curve of APACHE II and ATN-ISI were 0.895 (95% CI 0.829-0.960) and 0.858 (95% CI 0.783-0.934), respectively. The sensitivity of the hospital mortality prediction of the two scoring systems was 87.72% and 89.47%, respectively, and the specificity of hospital mortality prediction was 76.19% and 66.67%, respectively. No significant differences were found between the predicted and real mortality rates.

Conclusions: APACHE II and ATN-ISI scoring systems can predict the hospital mortality of elderly AKI patients. However, APACHE II performs better than ATN-ISI.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Acute Kidney Injury / mortality*
  • Age Distribution
  • Aged
  • Aged, 80 and over
  • Female
  • Geriatric Assessment / statistics & numerical data
  • Hospital Mortality / trends*
  • Humans
  • Male
  • Predictive Value of Tests
  • Prognosis
  • ROC Curve
  • Sensitivity and Specificity
  • Severity of Illness Index*
  • Survivors / statistics & numerical data