RESUMEN
PURPOSE: To test whether the oxygen uptake ([Formula: see text]) plateau at [Formula: see text] is simply a calculation artifact caused by the variability of [Formula: see text] or a clearly identifiable physiological event. METHODS: Forty-six male participants performed an incremental ramp and a [Formula: see text] verification test. Variability of the difference between adjacent sampling intervals (difference) and of the slope of the [Formula: see text]-workload relationship (slope) in the submaximal intensity domain were calculated. Workload defined sampling intervals used for the calculation of the difference and slope were systematically increased from 20 to 100 W until the expected risk of false plateau diagnoses based on the Gaussian distribution function was lower than 5%. Overall, more than 1500 differences and slopes were analyzed. Subsequently, frequencies of plateau diagnoses in the submaximal and maximal intensity domains were compared. RESULTS: Variability of the difference and slope decreased with increasing sampling interval (p < 0.001). At a sampling interval of 50 W, the predefined acceptable risk of false plateau diagnoses (≤ 5%) was achieved. At this sampling interval, the actual frequency (1.4%) of false-positive plateau diagnoses did not differ from the expected frequency in the submaximal intensity domain (1.6%; p = 0.491). In contrast, the actual frequency at maximal intensity (35.7%) was significantly higher compared to the submaximal intensity domain (p < 0.001) and even higher than the expected frequency of false-positive diagnoses (p < 0.001). CONCLUSION: The [Formula: see text] plateau at [Formula: see text] represents a physiological event and no calculation artifact caused by [Formula: see text] variability. However, detecting a [Formula: see text] plateau with sufficient certainty requires large sampling intervals.