A Comparison of Results From Two Bioimpedance Analyzers: HIV Infected V.S. Non-infected Adults

Abstract

LEARNING OUTCOME: To determine whether there would be greater variability in body composition (BC) results obtained from two different bioimpedance analyzers (BIA) used at the same time and under the same conditions with HIV infected (HIV+) v.s. non-infected (HIV-) adults.The accurate determination of BC in HIV+ individuals is a critical adjunct to the provision of timely and appropriate nutrition intervention. Yet its measurement may be complicated by the somewhat unexpected BC changes that are experienced by individuals with this disease. Subjects (S's) included 24 relatively healthy HIV+ adults (5♂,♀19; X̄age=42.7; X̄BMI= 24.9) and 29 healthy HIV- volunteers (22♂,7♀ X̄age=40.5; X̄BMI=26.5 ). All S's met four hour pre-test conditions (fasting, no vigorous exercise, no alcohol, no caffeine) and were measured either on the floor or on an exam table without any parts of the body touching each other. Height and weight were measured on all S's, and age and activity level were determined from interview. Sites were cleaned with alcohol, electrodes were placed on hand and foot as specified by manufacturer's instructions, and BC was measured consecutively using two BIAs (BodyStat 1500 [BS] and Maltron [M]). The difference in % body fat (%BF) as measured by the two analyzers was calculated for each subject (DIF = % BF-M minus % BF-BS). Data were analyzed using SPSS for Windows. Percent body fat results were: 1) BS-HIV+: 23.4%; 2) BS-HIV-: 30.3%; 3) M-HIV+: 23.8%; and 4) M-HIV-: 31.2%). Mean DIF was slightly less for the HIV+ group (0.4±4.9 v.s. 1.0±4.7; ns) and slightly greater for men (2.1 vs. -0.7; p<.05). Values for the 2 BIA methods were within ±1.5% of each other for 42% v.s. 21% for HIV+v.s. HIV- S's, respectively. Examination of other potentially important factors indicated no effect of age or activity level on the intra- and inter-group variability. However, S's with BMI >30 were more likely to be HIV- and have positive differences (value from Maltron > Body Stat). From these data it appears that the variability in measures between the two BIA methods is within acceptable limits in an HIV+ group indicating that the simple BIA technology may be adequate to monitor this important parameter of body composition in relatively healthy clients.