Introduction Malnutrition affects at least one in three hospital inpatients¹. It prolongs hospital stay and affects the recovery process. Early detection and intervention are essential to tackling malnutrition. Current nutritional assessment tools are dependent upon clinician experience and vary in their detection rates of malnutrition. We assess whether bioelectrical impedance analysis (BIA)-derived phase angle (PA), plasma markers such citrulline and pre-albumin could aid assessment and if current starving and resting protocols are required for BIA.

Method Eighty adult in-patients were recruited over a 40 week period. The Malnutrition Universal Screening Tool (MUST) and Subjective Global Assessment (SGA) were performed. Serum was analysed for citrulline and pre-albumin. PA was measured using Bodystat 4000 in three different metabolic states: rested and fasted, post-exercise and post food and exercise. The reference standard for malnutrition was determined using SGA. MUST score >0 indicated malnutrition risk, severe if >2. PA was considered to define malnutrition if lower than reference ranges for sex and age; severe malnutrition was indicated as >2 integers below reference range. Statistical analysis included paired t test and logistical regression.

Results 80 patients took part in the study; 47 (59%) male (20–86 years) and 33 (41%) female (18–87 years). Sixty-six patients completed all arms of the study; Fourteen patients did not consent to a blood test but completed all other aspects.

1. SGA and PA identified 47 (59%) and 50 (63%) patients with malnutrition compared to 9 (11%) by MUST

2. SGA and PA identified 15 (19%) and 10 (13%) patients with severe malnutrition compared to 2 (0.03%) by MUST

3. Compared to SGA, PA had sensitivity of 91% (95% CI 80%>98%) and specificity of 36% (95% CI 11%>69%). Positive predictive value was estimated at 86% (95% CI 80%>91%). Negative predictive value was estimated at 50% (95% CI 23%>77%)

4. There were no food or exercise related difference in PA (p=0.1837; p=0.1344 respectively)

5. Pre-albumin co-related with SGA (p=0.003), citrulline did not (p=0.348)

Conclusion These results show that BIA-derived PA is an objective measurement suitable for use in nutritional assessment. It is sensitive for detecting malnutrition and has a highly significant association with SGA. Furthermore, current fasting and resting protocols surrounding its use were found to be invalid. Pre-albumin is a useful biomarker in nutritional status assessment.

Reference

1. . British Association of Parenteral and Enteral Nutrition (BAPEN): “The cost of malnutrition in England and potential cost savings from nutritional interventions” Accessed: 20 February 2017 at: http://www.bapen.org.uk/pdfs/economic-report-short.pdf

Disclosure of Interest None Declared