A comparative analysis of pre- and post-dialysis albumin as indicators of nutritional and morbi-mortality risks in haemodialysis patients.

Introduction: Pre-dialysis albumin is likely to be falsely low due to a dilution effect, making its usefulness in assessing protein status questionable. OBJEVTIVES: The purpose of this study was to assess whether post-dialysis albumin would be a better marker of malnutrition and risk of mortality, when compared to pre-dialysis albumin.

Methods: We evaluated the correlation between pre- and post-dialysis albumin and the following parameters: body mass index (BMI), adequacy of muscle arm circumference (MAC) and tricipital skinfold (TS) to the 50th percentile (P50), C-reactive protein (CRP), phase angle (PA), protein equivalent of nitrogen appearance (PNA), the Kt/V index of dialysis adequacy, and the hydration status (Pearson's correlation coefficient). Agreement in the nutritional status according to pre- and post-dialysis (post-HD) albumin and PA was estimated according to the Kappa (K) coefficient (Bland-Altman).

Results: A total of 58 haemodialysis (HD) patients were included in this study (30 female; mean age: 49 years). BMI, PA and CRP had a significant correlation with pre- and post-HD albumin, while MAC and PNA correlated only with post-HD albumin. Agreement in the diagnosis of malnutrition according to PA < 5 and pre- and post-HD albumin < 3.2 g/dL was regular (K = 0.432). When using an albumin cut-off value of 3.7 g/dL for malnutrition (mild malnutrition or risk of malnutrition), the diagnosis was concordant only in the post-HD period (K = 0.544).

Conclusions: Post-dialysis albumin levels may be a better marker of protein status and mortality risk in cases of mild malnutrition or risk of malnutrition and in patients with low/medium mortality risk. Pre-dialysis fluid overload may be a confounding factor when evaluating albumin levels.