Predicting outcomes in critically ill patients with acute kidney injury undergoing intermittent hemodialysis--a retrospective cohort analysis.

Abstract

Despite advances in the management of critically ill patients with acute kidney injury (AKI), the prognosis is poor. The evidence base on risk factors for poor outcomes in these patients is scarce. Our aim was to identify predictors of outcome in AKI patients undergoing intermittent hemodialysis (IHD). We retrospectively analyzed patient records from consecutive, critically ill patients with AKI treated with IHD in one teaching secondary care hospital from 2002 to 2006. We used multivariate Cox proportional hazard regression analysis to identify predictors of mortality, hemodynamical instability during hemodialysis and failing renal recovery. Totally, we included 39 patients with a mean APACHE II score of 20.1 (SD 7.5) who had an average of 5.1 +/- 4.8 hemodialysis sessions. All-cause mortality was 35.9% (14/39 patients). In multivariate analysis, pre-existing cardiac co-morbidity (HR 1.92 [0.58-6.47]), metabolic acidosis (2.40 [0-74-7.74]) and presence of ARDS (1.83 [0.52-6.46]) were the strongest predictors. 7 patients (18%) sustained new hemodynamic instability during hemodialysis, for which ARDS (6.42 [0.64-64.03]) was a strong predictor. Among survivors, 20 patients (80%) had partial or complete renal recovery. Preexisting renal insufficiency (3.13 [0.34-29.13]) and high net ultrafiltration quantities (3.30 [0.40-26.90]) were the strongest predictors for failing renal recovery. As a consequence of the small samples size none of the associations was statistically significant. Presence of ARDS and high net ultrafiltration rates seem to represent key factors affecting prognosis in patients with AKI undergoing IHD. Targeting these risk factors may improve the poor prognosis of these patients.