Markers of malnutrition – which should we use?

L Wells, R Wolfenden, S Jorgensen, C Jones

York Hospital

Problem: Malnutrition is common in patients on haemodialysis. Identifying patients at risk from malnutrition and monitoring nutritional status is an important aspect of multidisciplinary care. However detailed nutritional assessment requires expertise and is a time consuming exercise. A simple test is needed to identify poor nutrition. The UK Renal Association standards recommend measurement of weight, height, BMI and serum albumin as a minimum for nutrition screening.

Purpose: To examine the relationships between a number of potential nutritional markers and assessment methods, in order to establish acceptable tools for the identification of malnutrition and to allow audit of nutritional outcomes.

Design: All patients on our haemodialysis unit have a nutritional assessment annually. Data was collected on weight, BMI, 3‑and 7‑point subjective global assessment (SGA), triceps skinfold thickness, mid arm muscle circumference (MAMC), 3 day food intake, a composite nutritional score (comprising both subjective and objective assessment), dialysis adequacy results, protein catabolic rate and biochemistry, including CRP, haemoglobin, sodium, potassium, phosphate, albumin, urea and creatinine. Results were compared to recognised nutritional standards.

Findings: Using 3‑point SGA as the gold standard, 75 patients had normal nutrition (score of A) and 26 were malnourished (25 score of B, 1 score C). However 8 patients had a BMI < 20 kg/m², 19 patients a %IBW < 90%, and 46 patients a serum albumin < 37 g/I. Patients with an SGA of B had significantly lower results for %IBW, BMI, MAC, MAMC and serum creatinine and higher results for composite nutritional score. However there was considerable overlap of values for %IBW, BMI, triceps skinfold centile and composite nutritional score between subjects scored as SGA A vs B. MAMC centile gave the best discrimination between normal and malnourished subjects, with the 95% confidence intervals for SGA B falling below the median value for SGA A.

Conclusion: Traditional markers of nutritional state (BMI, %IBW and serum albumin) give widely differing prevalence of malnutrition. A composite nutritional assessment will provide the most comprehensive picture of nutritional status. If a single marker is going to be used, then MAMC centile gives the greatest discrimination between normal and abnormal nutrition.

Relevance: Renal dietetic time is often limited leading to infrequency of nutrition assessment in large dialysis populations. However, if a quick, easy single marker is needed to predict nutrition risk, MAMC centile may be the most reliable.