ABSTRACT
OBJECTIVES: To determine outcomes of difficult-to-wean, ventilator-dependent patients transferred from intensive care units to rehabilitation hospitals and to determine predictors of weaning success in such patients. DESIGN: A retrospective cohort study. SETTING: A rehabilitation facility. PARTICIPANTS: One hundred forty-five difficult-to-wean patients (55.2% men; 83.4% white; mean age +/- standard deviation, 65.8 +/- 16.4y) transferred to a rehabilitation facility between July 1994 and June 1996. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Demographic and clinical data, including variables identified previously as predictive of weaning success among highly selected populations. RESULTS: Patients' Gillespie categories (reason for ventilator dependency) included "other medical conditions" (eg, pneumonia, neurologic) in 42.1% of the cases, postoperative in 24.8%, previous lung disease (eg, chronic obstructive pulmonary disease, interstitial lung disease) in 15.2%, trauma in 11.7%, respiratory failure with multisystem failure in 3.4%, and uncomplicated acute lung injury (acute respiratory distress syndrome) in 2.8%. Of 145 patients, 50.3% were completely weaned, 4.8% were partially weaned, and 44.8% remained ventilator dependent. In a stepwise multivariable logistic regression analysis, significant predictors of weaning success included white race (odds ratio [OR] = 3.4), serum albumin level (OR = 2.1g/dL), and blood urea nitrogen (BUN) level (OR = .97mg/dL); in addition, compared with postoperative patients, patients with "other medical conditions" (OR = .15) or previous lung disease (OR = .08) were less likely to be weaned (area under receiver operating characteristic curve = .76). Among 31 long-term survivors who were interviewed at least 6 months after discharge from the rehabilitation facility, 58.1% rated their health-related quality of life as good or better. CONCLUSIONS: Half of the patients admitted to a rehabilitation facility were weaned from their ventilators. Predictors of weaning success included race, BUN level, albumin level, and reason for ventilator dependency.
