Mid upper arm circumference and Powell-Tuck and Hennessy's equation correlate with body mass index and can be used sequentially in gastrostomy fed patients

Body Mass Index (BMI) is a simple and widespread method to assess undernutrition. However its use may be limited in bedridden patients. Aims:1) compare BMI, Mid Upper Arm Circumference (MUAC) and Powell-Tuck and Hennessy's regression equation for BMI (BMIPTH) in gastrostomy fed patients, 2) validate its correlation for sequential use and 3) select the best cut-off of MUAC and BMIPTH to predict risk of undernutrition.Prospective study including adult patients who underwent endoscopic gastrostomy (PEG). BMI, MUAC and BMIPTH adjusted for sex and age were determined at the day of gastrostomy (0), first (1) and third (3) months of follow up. Correlations between BMI and MUAC and BMIPTH were calculated at all time-points using Spearman's test. MUAC's Area Under ROC (AUROC) to predict risk of undernutrition (BMI ≤ 18,5 kg/m² in non-elderly (<65 years) and BMI ≤ 22,5 kg/m² in elderly (≥65 years)) was calculated with DeLong method and Youden Index was used to select the best cut-off for this outcome.405 PEG patients were included (69,9% males, median age 62,9 ± 15,3 years). Head and neck cancer and neurological disorders were the main indications for gastrostomy. BMI correlated moderately with BMIPTH0 (ρ = 0.646–0.694), MUAC0 (ρ = 0.669) and MUAC1 (ρ = 0.699). BMI correlated strongly with BMIPTH1 (ρ = 0.764–0.794), BMIPTH3 (ρ = 0.714–0.732) and MUAC3 (ρ = 0.725). MUAC and BMIPTH's accuracy was not significantly different to predict undernutrition neither in elderly (AUROC 0.835 ± 0.033 Vs. 0.836 ± 0.033 respectively, p = 0.319) nor in non-elderly patients (AUROC 0.857 ± 0.027 Vs. 0.888 ± 0.053 respectively, p = 0.256). MUAC <26 cm (positive predictive value (PPV) 83.5%) or BMIPTH <22 kg/m² (PPV 83.5%) in elderly and MUAC <25 cm (PPV 90.7%) or BMIPTH <21 kg/m² (PPV 91.7%) in non-elderly can accurately predict risk of undernutrition in PEG patients.MUAC and BMIPTH correlated with BMI not only at the day of PEG placement but also at the 1st and 3rd month of follow up and were equivalent to predict risk of undernutrition according to the new cut-offs defined for this population.