Correlates of Hypoalbuminemia in Community-Dwelling Older Persons
Published in: American Journal of Clinical Nutrition, v. 66, no. 1, July 1997, p. 38-45
Posted on RAND.org on December 31, 1996
To identify easily ascertainable sociodemographic and health characteristics that are associated with hypoalbuminemia in community-dwelling older persons, the authors used data from the first National Health and Nutrition Examination Survey. This population-based stratified probability sample survey included 4728 persons aged 55-74 years. They defined hypoalbuminemia in two ways: less than 35 g/L (1.2% of the sample) or less than or equal to 38 g/l (7.9% of the sample) and used multivariate logistic models to identify independent predictors of hypoalbuminemia. Older age; receiving welfare; a condition interfering with eating; vomiting greater than or equal to 3 d/mo; previous surgery for gastrointestinal tumor; self-reported heart failure; recurring cough attacks; feeling tired or worn-out; edentulous, fair, or poor condition of teeth; little or no exercise; a low-salt diet; trouble chewing meat; self-reported protein albumin, blood, or sugar in urine; and current cigarette smoking were independently associated with albuminemia (less than or equal to 38 g/L) or progressively lower albumin concentrations < 40 g/L. Persons with 3-5 of these factors (51.5% of the sample) had an odds ratio of 2.73 (95% CI: 1.64, 4.54) and those with greater than or equal to 6 factors (9.4% of the sample) had an odds ratio of 6.44 (95% CI: 3.49, 11.86) of albuminemia less than or equal to 38 g/L compared with those with 0-2 risk factors (39.1% of the sample). These findings suggest that several easily assessed sociodemographic, lifestyle, and disease-related factors are associated with hypoalbuminemia and might be valuable items to include on general health surveys to identify older persons who have this marker of poor health status.