Diagnostic accuracy and reliability of equations for the calculation of body fat mass percentage in the Colombian military population: a diagnostic test study.

Abstract

Introduction: Body composition is a crucial health indicator, particularly in military settings, where optimal physical fitness directly impacts performance and operational readiness. Accurate assessment of body fat and muscle mass is essential for identifying health risks, such as chronic diseases, and for improving soldiers' physical capabilities.

Methods: Study of diagnostic accuracy and reliability. A total of 947 adults belonging to the military forces in Colombia were included through consecutive sampling. People with the whole spectrum of body composition (normal weight, overweight and obesity) participated, recreating their prevalence in the Colombian population. Sensitivity (Se), specificity (Sp) and accuracy of the US Army (E1), Deurenberg (E2), CUN-BAE (E3) and Woolcott (E4) equations were calculated, using bioelectrical impedance analysis (BIA) as reference standard. For test-retest and inter-rater reliability, kappa and weighted kappa were used.

Results: Of the total participants, 92% were men, with a median age of 35 years in men and 33 years in women. 65.4% were physically active (150 min per week). Diagnostic accuracy demonstrated the following performance in obesity classification for the equation used in the US military (Se: 100%, Sp: 88.1%; 95% CI 85.3% to 90.6%), E2 (Se: 91.6%; 95% CI 87.5% to 94.7%, Sp: 86.6%; 95% CI 83.7% to 89.3%), E3 (Se: 96.8%; 95% CI 93.8% to 98.6% Sp: 67.9%; 95% CI 64% to 71.7%) and E4 (Se: 96.4%; 95% CI 93.3% to 98.3%, Sp: 57.9%; 95% CI 53.9% to 61.9%). For overweight classification, performance was found in E2 (Se: 71.5%; 95% CI 66.3% to 76.3%, Sp: 72.3%; 95% CI 68.2% to 76.1%) and E3 (Se: 47.9%; 95% CI 42.4% to 53.4%, Sp: 73.8%; 95% CI 69.8% to 77.6%). For normweight classification, E2 was found to perform (Se: 50.2%; 95% CI 44.1% to 56.3%, Sp: 95.7%; 95% CI 93.7% to 97.2%) and E3 (Se: 42.8%; 95% CI 36.8% to 48.9%, Sp: 98.8%; 95% CI 97.5% to 99.5%). For all equations, their reliability is excellent.

Conclusions: The US Army equation, when applied to the Colombian military population, presents greater accuracy for classifying obesity when compared to BIA. For the overweight and normal weight category, the Deurenberg equation performs better.