Physiological Demands of a CrossFit® Class: A Comparison Between Athletes and Recreational Participants

Background: The increasing popularity of CrossFit® has resulted in a heterogeneous population ranging from recreational practitioners to athletes. However, most studies have examined isolated benchmark workouts or complete sessions without characterizing cardiorespiratory responses across different session phases or considering training level. Methods: Thirty-nine CrossFit® practitioners (19 athletes and 20 recreational practitioners) completed two evaluation sessions. Maximal cardiorespiratory fitness was assessed in the laboratory, followed by a standardized 60 min CrossFit® class consisting of a 25 min warm-up, the 20 min Cindy workout, and a 15 min cool-down. During Cindy, participants completed as many repetitions as possible, with scaled movements for recreational practitioners when necessary. Cardiorespiratory responses and energy expenditure (EE) were measured throughout the session using indirect calorimetry (K4b2, COSMED, Rome, Italy) and analyzed according to exercise phase. Results: Athletes completed more repetitions than recreational practitioners (females: 597 vs. 342; males: 731 vs. 396; p < 0.001) and exhibited higher maximal oxygen uptake (V˙O2max) (females: 40.3 vs. 31.5; males: 49.2 vs. 31.4 mL·kg−1·min−1; p < 0.001). Cardiorespiratory and metabolic responses increased from warm-up to Cindy and declined during cool-down. During Cindy, athletes demonstrated higher V˙O2peak (females: 34.3 vs. 30.3; males: 40.7 vs. 32.3 mL·kg−1·min−1, p < 0.001), EE and metabolic equivalents than recreational practitioners (p < 0.05). No statistically significant between-group differences were detected in peak heart rate. Conclusions: A standardized CrossFit® class imposes substantial cardiorespiratory and energy demands regardless of training level. Although athletes exhibited higher V˙O2max during Cindy, 30% of recreational practitioners performed a scaled version, so direct comparisons between groups are not possible under equivalent workload.