Cross-sectional and longitudinal predictors of serum albumin in hemodialysis patients.BackgroundLower serum albumin concentrations predict increased mortality in hemodialysis (HD) patients. Many demographic, comorbidity, and modifiable treatment-related factors that predict HD patient outcomes may be associated with serum albumin.MethodsCross-sectional predictors of baseline albumin on December 31, 1993 were sought (N = 3981). Additional effects of the same baseline predictors on subsequent trends in albumin over one year were examined in a nested subsample of patients (N = 2245). Wave-1 of the United States Renal Data System Dialysis Morbidity and Mortality special study provided the data.ResultsSignificant associations (P < 0.05) are summarized as older age, female gender, peripheral vascular disease, chronic obstructive pulmonary disease, and cancer predicted a lower baseline albumin and negatively influenced subsequent albumin trends. Baseline albumin was higher for blacks (vs. whites), lower for smoking and diabetes, and lower during the first year of HD treatment (<3 months and 3 to 12 months, vs.> 1 year). Trend analysis showed more positive albumin slopes for patients in their first year on HD and more negative slopes for Native Americans (vs. whites). Baseline albumin was correlated with the type of vascular access being used [arteriovenous (AV) fistulas > AV grafts > permanent catheters > temporary catheters]. Trend analysis predicted more negative albumin slopes for AV grafts and permanent catheters (vs. AV fistula access). Baseline albumin correlated inversely with bicarbonate and directly with hematocrit. Dialysis with unmodified cellulose membranes, without reuse, predicted lower baseline albumin than the other membrane-reuse categories.ConclusionsSeveral exposures, which may be modifiable, were associated with serum albumin. Cross-sectional and longitudinal predictors of serum albumin in hemodialysis patients. Lower serum albumin concentrations predict increased mortality in hemodialysis (HD) patients. Many demographic, comorbidity, and modifiable treatment-related factors that predict HD patient outcomes may be associated with serum albumin. Cross-sectional predictors of baseline albumin on December 31, 1993 were sought (N = 3981). Additional effects of the same baseline predictors on subsequent trends in albumin over one year were examined in a nested subsample of patients (N = 2245). Wave-1 of the United States Renal Data System Dialysis Morbidity and Mortality special study provided the data. Significant associations (P < 0.05) are summarized as older age, female gender, peripheral vascular disease, chronic obstructive pulmonary disease, and cancer predicted a lower baseline albumin and negatively influenced subsequent albumin trends. Baseline albumin was higher for blacks (vs. whites), lower for smoking and diabetes, and lower during the first year of HD treatment (<3 months and 3 to 12 months, vs.> 1 year). Trend analysis showed more positive albumin slopes for patients in their first year on HD and more negative slopes for Native Americans (vs. whites). Baseline albumin was correlated with the type of vascular access being used [arteriovenous (AV) fistulas > AV grafts > permanent catheters > temporary catheters]. Trend analysis predicted more negative albumin slopes for AV grafts and permanent catheters (vs. AV fistula access). Baseline albumin correlated inversely with bicarbonate and directly with hematocrit. Dialysis with unmodified cellulose membranes, without reuse, predicted lower baseline albumin than the other membrane-reuse categories. Several exposures, which may be modifiable, were associated with serum albumin.