Objectives: To determine change in albumin from hospital discharge to 3-month follow-up and the prognostic significance of persistent hypoalbuminemia in older veteran patients.
Design: A prospective cohort study.
Setting: A geriatric rehabilitation unit of a university-affiliated Department of Veterans Affairs hospital.
Participants: The study population consisted of 282 subjects (of 322 randomly selected discharges who were free of cancer and terminal conditions) that completed the 90-day postdischarge assessment; most were older (75.4+/-8.6), white (76%), and male (99%).
Measurements: Each subject completed a comprehensive discharge assessment, had a repeat albumin an average of 94 days later, and was then tracked for 5 years. The strongest predictors of survival were identified using Cox proportional hazards regression analysis.
Results: Between hospital discharge and the 3-month reassessment, albumins improved by more than 2 g/L in 122 subjects (43%), stayed the same in 112 subjects (40%), and deteriorated by more than 2 g/L in the rest. During follow-up, 190 subjects (67%) died. Of the 38 nutritional, functional, demographic, and illness severity variables evaluated, the 3-month postdischarge albumin was the strongest predictor of long-term mortality. Those with albumins less than 35 g/L had a 2.6 times greater mortality than those with albumins of 40 g/L or greater (relative risk=2.6, 95% confidence interval=1.8-3.8). After controlling for 3-month albumin, hospital-discharge albumin was not significantly associated with long-term mortality.
Conclusion: In older people, a low serum albumin 3 months after hospital discharge is associated with a poor long-term prognosis. It is not known whether this represents ongoing inflammation or inadequate nutrition.