Objective: We sought to determine the prognostic value of serum albumin for 4-year mortality among high-functioning persons who did or did not have evidence of inflammation as indicated by high interleukin-6 (IL-6) levels.
Design: We used a case-cohort design of healthy, nondisabled older persons who had serum albumin and plasma IL-6 measured at baseline. Crude and multiply adjusted (for sociodemographics and chronic diseases) proportional hazards models were used to identify the effect of baseline levels of serum albumin level on 4-year mortality among those with higher and lower levels of IL-6.
Results: Among subjects without evidence of IL-6-mediated inflammation (IL-6 < 3.20 pg/mL), having a lower (< or = 4.4 g/dL) albumin level was associated with a multiply adjusted relative risk of 2.1 for 4-year mortality compared with those with higher albumin. In the presence of inflammation (IL-6 > or = 3.20 pg/mL), higher and lower serum albumin levels had similar risks (adjusted relative risks 4.0 and 3.8, respectively) compared with the referent group (higher albumin and low IL-6).
Conclusions: High serum albumin has a protective effect in healthy older persons who do not have evidence of cytokine-mediated inflammation. This protective effect is not conferred in presence of inflammation. The mechanisms by which inflammation eliminates the protective effect of high albumin remain to be determined.