Background: Low serum albumin is a strong risk factor for mortality, but its association with low serum bicarbonate and inflammation in the setting of mild to moderately decreased kidney function is uncertain.
Methods: We analyzed data from 15594 subjects over the age of 20 who participated in the Third National Health and Nutrition Examination Survey (NHANES III). Glomerular filtration rate (GFR) in mL/min/1.73 m2 was estimated by the abbreviated Modification of Diet in Renal Disease (MDRD) equation using appropriately calibrated serum creatinine.
Results: The age-adjusted prevalence of hypoalbuminemia (serum albumin <3.8 g/dL) at a GFR of 90, 60, 30, and 15 mL/min/1.73 m2 was 19%, 21%, 38%, and 59%, respectively, while the age-adjusted prevalence of C-reactive protein (CRP) >or= 0.22 mg/dL was 36%, 44%, 69%, and 81%, respectively, both P trend <0.001. Age, female gender, non-Hispanic black compared with non-Hispanic white race, diabetes, hypertension, hepatitis C, urine albumin: creatinine ratio >1 g/g, dietary protein intake, dietary caloric intake, serum bicarbonate, CRP, and GFR category were all significant predictors of hypoalbuminemia on univariate analysis. On simultaneously adjusting for the above variables, hypertension, diabetes, GFR, and dietary protein and caloric intake were no longer significant independent predictors of hypoalbuminemia. The adjusted odds ratio (OR) of serum bicarbonate (by quartile) for hypoalbuminemia was 1.0 for serum bicarbonate >28 mEq/L (reference), 1.25 for 26-28 mEq/L, 1.51 for 23-25 mEq/L, and 1.54 for <or=22 meq/L. The adjusted OR of CRP for hypoalbuminemia was 1.0 for CRP < 0.22 mg/dL (reference), 2.60 for 0.22-1.0 mg/dL, and 5.56 for >1.0 mg/dL.
Conclusion: Elevated CRP and low serum bicarbonate are independently associated with hypoalbuminemia, explaining much of the high prevalence of hypoalbuminemia in chronic kidney disease.