Resting Metabolic Rate In Collegiate Male Wrestlers: Cross-validation Of Prediction Equations

PURPOSE: To cross-validate existing RMR prediction equations using laboratory-measured RMR. METHODS: National Collegiate Association Division III male wrestlers participated in this study (n = 177; Age: 19.6 ± 1.5 yrs.; Height: 176.3 ± 6.3 cm; Weight: 76.6 ± 10.9 kg; Body fat: 11.4 ± 3.0 %). Indirect calorimetry was used to measure RMR. Skinfold were measured to estimate body density and body fat percentage (BF%) using the Lohman and Brozek skinfold prediction equations, respectively. RESULTS: Mean measured RMR was 2152 ± 306 kcal·d−1. All RMR prediction equations significantly underestimated the measured RMR, except for the Jagim equation, which overestimated RMR (p < 0.001) (Table 1). The Tinsley fat-free mass (FFM) equation and the Jagim equation had the lowest constant error (CE) values when compared to measured RMR (mean difference ± standard error: 110 ± 22 kcal and -119 ± 23 kcal, respectively). Linear regression indicated the Tinsley FFM (R2 = 0.454, RMSE = 236 kcals, %RMSE = 10.9%) and Cunningham FFM equations had the closest agreement to measured RMR (R2 = 0.454, RMSE = 259 kcals, %RMSE = 12.0%), along with the Jagim equation for a body weight-based prediction equation (R2 = 0.492, RMSE = 260 kcals, %RMSE = 12.1%). The prediction error for predicted RMR ranged from an SEE = 227 kcal·d−1 (Cunningham, Mifflin, Nelson, and Tinsley FFM) to an SEE = 239 kcal·d−1 (DeLorenzo), accounting for 45%, 67%, 45% and 44% of the variance in RMR, respectively. CONCLUSION: For FFM-based RMR prediction equations, the Tinsley and Cunningham equations exhibited acceptable validity. The Jagim body weight-based equation provided acceptable validity for predicting RMR, which provides a useful tool for predicting daily energy requirements when body composition measurements are not available. It is recommended that future research explore the development of a wrestler-specific RMR prediction equation. Table 1. Differences between measured and predicted resting metabolic rate. - Mean ± SD r p (correlation) CE ± SE 95% CI 95% CI p(repeated measures) Measured (kcal·d−1) 2152±306 Cunningham (kcal·d−1) 1993±193 .674 <0.001 159±22 85 232 <0.001 DeLorenzo (kcal·d−1) 1899±157 .665 <0.001 252±24 175 330 <0.001 Mifflin (kcal·d−1) 1750±172 .674 <0.001 402±23 328 476 <0.001 Jagim (kcal·d−1) 2270±208 .702 <0.001 -119±23 -193 -44 <0.001 Nelson (kcal·d−1) 1782±230 .675 <0.001 369±22 296 443 <0.001 Tinsley BW (kcal·d−1) 1915±265 .702 <0.001 237±24 159 314 <0.001 Tinsley FFM (kcal·d−1) 2042±227 .674 <0.001 110±22 37 184 <0.001 Schofield (kcal·d−1) 1849±161 .702 <0.001 303±23 227 379 <0.001