Serum albumin (SA) levels and 10-year mortality in a community-dwelling 70-year-old population.

Although an association between SA concentrations and mortality in the patients is well known, this association is not conclusive in elderly community-dwelling populations. We therefore attempted to determine whether this association could be extended to a Japanese 70-year-old community-dwelling population. Seventy-year-old subjects residing in Niigata City, Japan participated. Baseline examinations including a determination of SA concentrations were performed in June 1998 in 600 participants, and they were followed for 10 years. During the 10-year follow up, 80 subjects died. Albumin levels were divided into four groups (highest>45 g/l, higher 43-44 g/l, lower 41-42 g/l, lowest <40 g/l). The survival rate using Kaplan-Meier methods was longer in the highest and higher albumin groups than in the lowest and lower groups. No differences were found between the four groups in the mortalities from cancer, cardiovascular disease (CVD) or pneumonia. Individuals with the lowest albumin levels had 2.7 or 2.9 times higher total mortality, whether adjusted with confounding factors or not, using Cox regression analysis. The mortality hazard ratio (HR) found in the lowest group was compared to the highest group. In an elderly 70-year-old Japanese community-dwelling population, lower SA concentrations were an independent predictor of total mortality, but not mortality due to cancer, CVD, or pneumonia.