Energy Expenditure Analysis and Prediction in Smith Machine Squats Integrating Mechanical Loads and Sex

Qiao, Z, Guo, B, Gao, Y, Wang, Y, Jiang, M, Yu, J, Yi, L, Yan, B, Qiu, J, and Girard, O. Energy expenditure analysis and prediction in Smith machine squats integrating mechanical loads and sex. J Strength Cond Res 40(5): 519-527, 2026-Calculating energy expenditure (EE) and metabolic equivalents (METs) from oxygen uptake (V̇o2) alone often underestimates resistance training intensity because of its intermittent, brief, and high-intensity nature. This study aims to refine EE and METs estimation during Smith machine squats by integrating mechanical work, glycolytic contribution, and postexercise EE, with particular emphasis on weight-dependent and sex-specific metabolic responses. Fifty-one subjects (31 men) performed weight-bearing squats (4 kg-80% 1 repetition maximum) for 5 repetitions across 3 sets with 4-minute rest intervals. Aerobic, glycolytic, and postexercise EE were estimated from V̇o2, blood lactate, and excess postexercise oxygen consumption, respectively. Mechanical work was measured using a linear-position transducer system. Training load (p < 0.05) and sex (p < 0.01) significantly influenced total EE. Mechanical efficiency was significantly influenced by training load (p < 0.001). Energy expenditure-based METs increased with training weight (p < 0.05) and were higher than V̇o2-based METs (p < 0.001). Training weight and mechanical work strongly correlated with total EE (r ≥ 0.80, p < 0.001). The optimal model, including body weight, training load, peak heart rate, and movement distance, demonstrated high accuracy (R2 = 0.885, cross-validated R2 = 0.803, MAE = 10.067). A simplified model using only body weight and training weight achieved comparable accuracy (R2 = 0.847, cross-validated R2 = 0.750, MAE = 10.883). Total EE increased with training weights, but low-to-moderate-weight squats showed superior mechanical efficiency. Women tended to expend less energy than men during squats, but no difference in mechanical efficiency was observed. V̇o2-based METs underestimated squat intensity. Both the simplified and optimal models accurately estimated squat EE, with the optimal model being more accurate.