Beyond Age and Fitness: Modeling Metabolic and Vascular Predictors of VO₂max in Firefighters Using GLM and GAM Approaches

Authors

Yuhang Liu, Texas A&M UniversityFollow
Michael J. Conner, Front Line Mobile HealthFollow
NEWMAN BRIAN, Fort Worth Fire Department, Fort WorthFollow
Lisa C. Colvin, Texas A&M University, College Station, TXFollow
Steven E. Martin, Texas A&M University, College StationFollow
Drew Gonzalez, Sam Houston State University, HuntsvilleFollow

Abstract

While aerobic capacity (VO₂max) is vital for cardiovascular and metabolic health, limited evidence exists on how physical performance, body composition, and biochemical markers jointly predict aerobic fitness in firefighters. Combining these indicators may improve understanding of VO₂max variability and aid strategy development to reduce cardiometabolic risk. PURPOSE: This study investigates how body composition, physical performance, and blood biomarkers relate to VO₂max in firefighters, with the aim of developing predictive models to optimize health and performance. METHODS: 170 firefighters were analyzed. Variance inflation factors (VIFs) were used to check predictor independence. General linear models (GLM) identified linear predictor– VO₂max relationships, while generalized additive models (GAM) explored nonlinear ones. Tenfold cross-validation tested model performance and generalizability, with diagnostic and sensitivity analyses assessing assumptions and robustness. RESULTS: VIF (1.13–5.43; TG = 5.43) indicated no serious multicollinearity; all predictors were retained. The baseline GLM—including age, sex, body composition indices, resting heart rate, grip strength, and push-ups—explained a moderate proportion of VO₂ variance (adjusted R² = 0.446, F (7,96) = 12.85, p < 0.001). Significant predictors were age (β = −0.188, p = 0.002), sex (higher in males; β = 10.66, p = 0.009), and fat mass (β = −0.362, p < 0.001). Adding ePWV, HDL, TG, HbA1c, and AOPP improved model fit (adjusted R² = 0.469, F (12,91) = 8.58, p < 0.001); fat mass (β = −0.359, p < 0.001), sex (β = 12.16, p = 0.004), and HDL (β = 0.113, p = 0.035) remained significant. The GAM demonstrated superior fit (adjusted R² = 0.545, AIC = 625.9) and nonlinear effects for fat mass and ePWV. Cross-validation showed better performance for the GAM (CV-RMSE=5.05, CV-R²=0.26) than the GLM (CV-RMSE=5.27, CV-R²=0.19). A predictive equation was derived: VO₂≈25.68 − 0.135×Age + 12.16×Sex − 0.359×Fatkg + 0.113×HDL. CONCLUSION: Higher fat mass and arterial stiffness are linked to lower aerobic capacity in firefighters; HDL was positively associated with VO₂max. The GAM revealed curvilinear relationships among fat, ePWV, and VO₂max, suggesting metabolic and vascular factors influence VO₂max beyond age and sex. Early interventions to reduce adiposity, improve arterial health, and maintain high HDL could preserve VO₂max and lower cardiovascular risk in firefighters.

Recommended Citation

Liu, Yuhang; Conner, Michael J.; BRIAN, NEWMAN; Colvin, Lisa C.; Martin, Steven E.; and Gonzalez, Drew (2026) "Beyond Age and Fitness: Modeling Metabolic and Vascular Predictors of VO₂max in Firefighters Using GLM and GAM Approaches," International Journal of Exercise Science: Conference Proceedings: Vol. 2: Iss. 18, Article 9.
Available at: https://digitalcommons.wku.edu/ijesab/vol2/iss18/9