Nutrition assessment and risk prediction in dialysis patients-a new integrative score

We developed a quantitative nutritional score, based on biochemical measures, taken as part of monthly routine care. The score can be accomplished within a short time after routine laboratory results completion and identify a monthly change in nutritional status.

A longitudinal observational cohort study

The Institute of Nephrology, Wolfson Medical Center, Holon, Israel.

A total of 179 hemodialysis patients were followed up for up to 2.5 years after study baseline.

The Integrative Clinical Nutrition Dialysis Score (ICNDS) is based on the biochemical measures of albumin, creatinine, urea, cholesterol, C-reactive protein, dialysis adequacy, and weight change. Each parameter is ranked between 1 and 5, with the higher rank derived from recommended National Kidney Foundation Kidney Disease/Dialysis Outcomes and Quality Initiative values and the lower rank indicating deviation from those values. The final ICNDS is the sum of ranks over 7 parameters.

The Pearson correlation coefficient was calculated for association between subjective global assessment and ICNDS in 63 randomly selected patients. In 179 dialysis patients, the baseline ICNDS, the slope of 3 subsequent monthly ICNDS values, were tested for their correlation with odds of all-cause mortality, hospitalization frequency, length of stay, after 31 months. Spline Cox regression was used to select the best cutoff point, associated with severe mortality risk.

Score results were significantly correlated with nutrition evaluation by subjective global assessment (r = 0.842, P < .01). For a unit increase in baseline score, death odds were significantly decreased (hazard ratio [HR] = 0.929, 95% confidence interval [CI] 0.88-0.974, P < .002). Each unit increase in slope significantly reduced mortality risk (HR = 0.485, 95% CI 0.278-0.847, P < .011). Hospitalization frequency was significantly increased across worsening baseline score (HR = 0.935, 95% CI 0.906-0.964, P < .0001). A 1-unit increase in slope significantly decreased hospitalization (HR = 0.799, 95% CI 0.726-0.881, P < .0001).

Results confirm that ICNDS is a useful prognostic tool that serves to detect nutrition deterioration at its very beginning.