BACKGROUND An important proportion of critically ill patients who survives their acute illness remains in a critical state requiring intensive care management for weeks to months. Nevertheless, data on risk factors for in-hospital mortality and especially for long-term mortality and functional capacity are scarce. This study investigated outcome and prognostic factors in long-term critically ill patients. METHODS This retrospective observational cohort study was performed at our mixed adult 8-bed cardiologic ICU at a 2200-bed University Hospital. Patient data from our local database connected to an Austrian multicenter program for quality assurance in intensive care were analyzed. Data were collected between March 1(st), 1998 and December 31(st), 2003. Patients with an ICU stay > or =30 days formed the long-term study group. Morbidity and functional capacity were assessed using the Barthel mobility index in telephone interviews. RESULTS Patients spending > or =30 days in the ICU numbered 135 (10%) and occupied 5962 bed-days, representing 40.9% of the total bed-days. Compared with patients with an ICU stay <30 days, patients in the long-term group had a significantly higher SAPS II score during the first 24 hours after ICU admission (54 [IQR 41-65] vs. 38 [IQR 27-56], p < 0.001). There was a trend towards male preponderance in the long-term group (98/135 [82.6%] vs. 782/1215 [64.4%], p = 0.05). Differences in ICU and in-hospital mortality were not significant (28/135 (20.7%) vs. 295/1215 (24.3%), p = 0.620 and 46/135 [34.1%] vs. 360/1215 [29.6%], p = 0.285, respectively). After 12 and 48 months, the overall cumulative rates of death in hospital survivors were 14% and 26%, respectively in the short-term ICU group and 31% and 61% in the long-term group. A log-rank test revealed a significantly higher probability of survival in the short-term group after hospital discharge (log rank = 34.3, p < 0.001). Multivariate analysis of hospital survivors and non-survivors in the long-term group showed that the need for renal replacement therapy during the ICU stay was the sole independent predictor for in-hospital death and death within 1 year after ICU discharge (OR = 2.88; 95%CI 1.12-7.41, p = 0.028 and OR = 3.66, 95%CI 1.36-9.83, p = 0.01, respectively). In 28/31 long-term survivors (90%) in the long-term ICU group, the Barthel index indicated no or only moderate disability during daily activities. CONCLUSIONS Hospital mortality rates in critically ill patients with a stay <30 or > or =30 days were comparable. The necessity for renal replacement therapy was the sole independent predictor for in-hospital and 1-year mortality in long-term ICU patients. Critically ill patients with a stay > or =30 days have a high and ongoing risk of death after hospital discharge; however, a substantial number of these patients are long-term survivors with no or only moderate disability during daily activities.