Phase Angle From Bioelectrical Impedance Analysis: Population Reference Values by Age, Sex, and Body Mass Index

This study investigates whether bioimpedance indexes rather than derived body compartments would be adequate for nutritional assessment. Evidence is provided that the phase angle as determined by conventional tetrapolar whole body bioelectrical impedance analysis at 50 kHz (1) was largely determined by the arms and legs and not the trunk, (2) was higher in control subjects than in hospitalized patients [mean (SD) 6.6 degrees (0.6) degrees vs 4.9 degrees (1.2) degrees, P<0.001], (3) discriminated poorly between cirrhotic patients of different Child-Pugh class, and (4) was positively correlated with muscle mass ( r=0.53) and muscle strength ( r=0.53) in these patients (each P<0.01). In a prospective study of patients with liver cirrhosis Kaplan-Meier and log rank analyses of survival curves demonstrated that patients with phase angles equal to or less than 5.4 degrees had shorter survival times than patients with higher phase angles [6.6 degrees (1.4) degrees ] and that phase angles less than 4.4 degrees were associated with even shorter survival times ( P<0.01). The prognostic roles of the phase angle and standard nutritional parameters such as total body potassium, anthropometric measurements, and impedance derived fat free mass, body cell mass and fat mass were evaluated separately by Cox regression which eliminated all variables except the phase angle as predictors of patient survival time ( P<0.01). We concluded that for the clinical assessment of patients the phase angle may be superior to commonly used body composition information.

To compare the value of body fat mass (%FM) to indirect measures of general (body mass index (BMI)) and central adiposity (waist circumference (WC); waist-to-height ratio (WC/ht)) for the prediction of overweight- and obesity-related metabolic risk in a study population with a high prevalence of metabolic syndrome (MSX). BMI, WC, WC/ht, body composition (by air-displacement plethysmography) and metabolic risk factors: triglycerides, cholesterol, HDL-cholesterol (HDL-C), uric acid, systolic blood pressure (BPsys), insulin resistance by homeostasis model assessment (HOMA-IR) and C-reactive protein (CRP) were measured in 335 adults (191 women, 144 men; mean age 53 +/-13.9 years, prevalence of MSX 30%). When compared with BMI and WC, %FM showed weaker associations with metabolic risk factors, except for CRP and BPsys in men. In women, HDL-C and HOMA-IR showed the closest correlations with BMI. For all other risk factors, WC or WC/ht were the best predictors in both sexes. Differences in the strength of correlations between an obesity index and different risk factors exceeded the differences observed between all obesity indices within one risk factor. In stepwise multiple regression analyses, WC/ht was the main predictor of metabolic risk in both sexes combined. However, analysis of the area under receiver operating characteristic curves for prediction of the prevalence of >or=2 component traits of the MSX revealed a similar accuracy of all obesity indices. At the population level, measurement of body FM has no advantage over BMI and WC in the prediction of obesity-related metabolic risk. Although measures of central adiposity (WC, WC/ht) tended to show closer associations with risk factors than measures of general adiposity, the differences were small and depended on the type of risk factor and sex, suggesting an equivalent value of methods.