Value of body mass index in the detection of severe malnutrition: influence of the pathology and changes in anthropometric parameters

Abstract

Background & aims: We have estimated the prevalence of severe malnutrition in groups of patients hospitalized for different medical causes and assessed the sensitivity of BMI in the diagnosis of severe malnutrition.

Design: A prospective study enrolled 1052 patients: 396 patients with liver cirrhosis including 165 non-ascitic patients (NAP), 124 patients with mild ascites (MAP), 107 patients with tense ascites (TAP), 251 patients after cardiac surgery (SCP), 81 patients with cardiac diseases (MCP), 85 patients with stroke (SP), 36 patients with degenerative neurological diseases (DNP), 68 patients after surgery of a hip fracture (HFP), 91 patients with palliative care for cancer (CP) and 44 elderly patients with medical affections (EP). BMI, mid-arm muscular circumference (MAMC) and triceps skinfold thickness (TST) were measured within 48 h after admission. Patients with MAMC and TST below the 5th percentile of a reference population when aged < or = 74 or the 10th percentile when aged > or = 75 were defined as severely malnourished. Sensitivity of BMI < 20 to detect malnutrition was assessed.

Results: The prevalence of severe malnutrition was the highest in TAP (39.1%) HFP (25.6%) and MAP (24.3%) and the lowest in SCP (4%), SP (4.8%), DNP (5.7%) and MCP (7.4%) (P < 10(-4)). In multivariate analysis, low TST was associated with female gender (P < 10(-4)) mild and tense ascites (P = 0.038, P = 0.0004), low MAMC with male gender (P < 10(-4)), low BMI with female gender (P = 0.0082), hip fracture (P = 0.0407) and cancer (P = 0.0059). The sensitivity of BMI to detect severe malnutrition was the highest in HFP, CP and EP (100%, 80% and 100% respectively) and the lowest in TAP, MCP and SP (40%, 33.3% and 50% respectively). After exclusion of TAP, sensitivity of BMI to detect malnutrition correlated significantly with the coefficient of correlation between MAMC and TST observed in each group (r = 0.821, P = 0.0066).

Conclusion: Ascitic cirrhotic patients and elderly patients after surgery of hip fracture had the highest prevalence of severe malnutrition. BMI had the highest sensitivity when both TST and MAMC were damaged to the same extent. BMI < 20 has a high sensitivity in the diagnosis of severe malnutrition in elderly and cancer patients but not in cirrhotic patients with tense ascites, cardiovascular and neurological patients.