Association Between Serum Albumin and Mortality in Dialysis Patients Is Partly Explained by Inflammation, and Not by Malnutrition

Renée de Mutsert,Diana C. Grootendorst,Fleur Indemans,Elisabeth W. Boeschoten,Raymond T. Krediet,Friedo W. Dekker,for the Netherlands Cooperative Study on the Adequacy of Dialysis-II Study Group
DOI: https://doi.org/10.1053/j.jrn.2008.08.003
IF: 4.354
2009-03-01
Journal of Renal Nutrition
Abstract:OBJECTIVE: We investigated the effects of inflammatory and nutritional status on the association between serum albumin and mortality in hemodialysis (HD) and peritoneal dialysis (PD) patients.DESIGN AND PATIENTS: This was a prospective cohort study of incident dialysis patients starting HD or PD. Inflammation (C-reactive protein >or=5 or >or=10 mg/L), malnutrition (1 to 5 on the 7-point subjective global assessment [SGA]), and low protein intake (normalized protein equivalent of nitrogen appearance [nPNA] <0.99 g/kg/day) were measured at 3 months after the start of dialysis.SETTING: The study involved 38 dialysis centers in The Netherlands.MAIN OUTCOME MEASURE: We ascertained all-cause mortality during the first 2 years after the start of dialysis.RESULTS: In total, 700 patients were included (mean SD age, 59 [+/-15] years; serum albumin, 3.3 (0.7) g/dL; 60% men; 454 starting HD, and 246 starting PD). The 2-year mortality was 21%. In HD patients, the mortality (hazard ratio [HR], with 95% confidence interval [95% CI]) per unit decrease in serum albumin (g/dL) was 1.47 (95% CI, 1.07 to 2.00). Adjustment for SGA did not decrease this risk, whereas adjustment for nPNA decreased the HR to 1.45 (95% CI, 1.06 to 1.97). The mortality risk decreased to 1.30 (95% CI, 0.95 to 1.78) after adjustment for inflammation, and did not further decrease after additional adjustment for SGA and nPNA. Additional adjustments for age, sex, and comorbidity decreased the HR to 1.09 (95% CI, 0.79 to 1.51). In PD patients, the effects of adjustments on the mortality risk of serum albumin (1.38; 95% CI, 0.87 to 2.20) were similar.CONCLUSION: In dialysis patients, a 1-g/dL decrease in serum albumin was associated with an increased mortality risk of 47% in HD patients and 38% in PD patients. These mortality risks were in part explained by the inflammatory pathway. The mortality risks associated with serum albumin were not a consequence of malnutrition, as measured with SGA and nPNA. These findings imply that nutritional status cannot be assessed with precision by the measurement of serum albumin in dialysis patients.
nutrition & dietetics,urology & nephrology
What problem does this paper attempt to address?