Baseline Physiological Biomarkers Associated With Training Completion in a Comparative Observational Study of Extreme Military Training.

INTRODUCTION: The Brazilian Army Commandos Action Course (CAC) is an intense 16-week elite military program that exposes volunteers to extreme physical, psychological, and environmental stressors across diverse biomes. Although multi-domain biomarker panels are increasingly recognized for monitoring operational stress, baseline predictive models to forecast training completion in elite military courses remain scarce. The main study objective was to characterize and compare baseline biochemical, hematological, and bioimpedance-derived physiological markers between candidates who completed (Finishers) and those who failed to complete (Non-Finishers) the CAC, and to evaluate integrated multivariable logistic regression models to predict course completion. MATERIALS AND METHODS: In this 4-year observational study (2012-2015; STROBE-compliant), 243 male active-duty military candidates were evaluated at baseline under standardized fasting conditions before CAC entry. Assessments included serum biochemistry (enzymes, electrolytes, metabolic markers), calculated osmolality (OSMc), estimated glomerular filtration rate (eGFR), complete blood counts, and multi-frequency bioelectrical impedance analysis (BIA). Data were analyzed using univariate independent-samples t-tests, principal component analysis (PCA) for dimensionality reduction, and stepwise multivariable logistic regression models evaluated via the Akaike Information Criterion (AIC). RESULTS: Of the 243 participants, 66 (27.2%) completed the course (Finishers) and 177 (72.8%) did not (Non-Finishers). Univariate comparisons revealed that Finishers possessed significantly higher baseline skeletal muscle mass (SMM: 38.70 ± 4.41 kg vs. 37.59 ± 4.38 kg; P = .04), intracellular water (ICW: 31.21 ± 3.38 L vs. 30.36 ± 3.36 L; P = .04), and extracellular water (ECW: 18.36 ± 2.11 L vs. 17.86 ± 2.11 L; P = .05). PCA identified 3 primary components that accounted for 44.65% of the total variance: body composition and fluid distribution (PC1), hydration and electrolyte regulation (PC2), and adiposity (PC3). An integrated multi-domain logistic regression model (Model 3) achieved superior goodness of fit and parsimony (AIC = 284.94; Hosmer-Lemeshow P = .330) compared with single-domain models. Serum sodium (Na, OR = 1.51), hemoglobin (HGB, OR = 0.218), and hematocrit (HCT, OR = 1.815) emerged as significant independent multivariate predictors of course outcomes (P < .05), while CK was retained (OR = 0.998, P = .084). CONCLUSIONS: Baseline physiological preparedness-specifically the integrated balance of muscular reserve, cellular hydration, body composition, and oxygen transport capacity-influences the successful completion of extreme military training. Composite, multi-domain physiological profiles explain outcomes better than isolated biomarkers, highlighting the advantages of higher lean mass and the dual mechanical and thermoregulatory burdens of excess adiposity. Although these exploratory models have modest predictive capacity and warrant cautious interpretation, they offer a promising framework for utilizing routine screening (e.g., bioimpedance, basic blood tests) to guide targeted prehabilitation and proactively mitigate attrition in elite forces.