End-stage renal failure patients requiring renal replacement therapy in the intensive care unit: Incidence, clinical features, and outcome

Shigehiko Uchino, Hiroshi Morimatsu, Rinaldo Bellomo, William Silvester, Louise Cole

Research output: Contribution to journalArticleResearchpeer-review

46 Citations (Scopus)


Aims: To study incidence, clinical features, and outcome of critically ill patients with end-stage renal failure (ESRF) requiring renal replacement therapy (RRT) in the intensive care unit (ICU) and to test the validity of severity scoring systems for these patients. Methods: Data for ESRF patients treated with RRT were collected from 81 Australian adult ICUs providing RRT. They were compared with matched controls with acute renal failure. Results: Thirty-eight ESRF patients received RRT in the ICU over 3 months. The mean APACHE II score was 21.8 (predicted mortality: 37%) and the SAPS II score 44.7 (predicted mortality: 37%). The hospital mortality was 34%. Receiver operating characteristic curves showed good discrimination ability for hospital mortality for these two scores (AUC: 0.81 for APACHE II and 0.84 for SAPS II). Using admission diagnosis and SAPS II scores, 32 ESRF patients treated with continuous RRT (CRRT) were matched to 32 acute renal failure patients also treated with CRRT. ICU mortality (22 vs. 38%) and hospital mortality (38 vs. 38%) were comparable between the two groups. Conclusions: ESRF patients requiring RRT in the ICU were relatively frequent. Severity scores could be used to predict the hospital outcome for these patients. Their mortality, when treated with CRRT, was similar to that of diagnosis- and severity-score-matched patients with acute renal failure.

Original languageEnglish
Pages (from-to)170-175
Number of pages6
JournalBlood Purification
Issue number2
Publication statusPublished - 20 Mar 2003
Externally publishedYes


  • Acute renal failure
  • Chronic renal failure
  • Critical illness
  • Epidemiology
  • Illness severity scores
  • Renal replacement therapy

Cite this