Background and purpose: Advances in critical care medicine have increased the chances of survival for patients with severe illness or trauma. However, such patients consume a large proportion of medical resources. This study sought risk factors for mortality that have potential to be modified among patients treated in medical or surgical intensive care units (ICUs).
Methods: This 6-month prospective observational study was conducted in the medical and surgical ICUs of an 1,800-bed university hospital. All adult patients with an expected ICU stay of 48 hours or more were followed up regularly until discharge from the ICUs, or for 10 weeks during their stay in ICUs.
Results: Of 342 patients enrolled, 77 (22.5%) died during a median follow-up period of 5 days (range, 2-70 days). Among a range of variables at the time of ICU entry or developing during stay in ICUs, 17 were associated with higher mortality rate. Multivariate analysis using a logistic regression model demonstrated that the presence of systemic inflammatory response syndrome (SIRS) at the time of ICU entry (adjusted relative risk, ARR, 2.85; 95% confidence interval, CI, 1.16-7.05), Acute Physiological and Chronic Health Evaluation (APACHE) II score on ICU Day 4 (ARR 1.12 with increment of one score; 95% CI 1.01-1.24), Therapeutic Intervention Scoring System (TISS) score on Day 4 (ARR 1.13 with increment of one score; 95% CI 1.05-1.23), parenteral nutrition (ARR 4.97, 95% CI 1.73-14.26), and nosocomial Candida infection (ARR 3.39, 95% CI 1.12-10.23) were independently associated with ICU mortality.
Conclusions: In addition to SIRS and the APACHE II and TISS scores, this study found that nosocomial Candida infection and parenteral nutrition were independently associated with mortality after control for admission conditions, severity of illness scores, and interventions.