Dear Editor-in-Chief: In studies not using the CPX/D, the gross efficiency (GE) of some humans has been reported to clearly surpass 22.0% during cycling (2,6–8). The accuracy of any metabolic cart largely depends on the users’ expertise (adequate maintenance and calibration procedures, request of reliable technical support by manufacturers). Previous studies have shown that, compared with the Douglas bag method, the CPX/D system is valid to measure exercise V̇O2 (1,9,10), e.g., mean variations of 1.8–3.6%, within the 4% standard range of most metabolic systems (1,10). During the same period of the season (January–February), six subjects (4) were also tested in another laboratory with a different metabolic system (Oxycon Champion, Jaeger, Germany) while performing a ramp test (workload increases of 25 W·min−1). Differences in mean V̇O2 values between the CPX/D and the Oxycon system were consistently within the 3.2% range across all workloads (data are courtesy of Dr. Rabadán). Further, if our CPX/D system had underestimated V̇O2 data by ∼12%, as Jeukendrup et al. suggest, the “actual” V̇O2max of subjects 1 and 4 (Table 1 of reference 4) would have been ∼92 mL·kg−1·min−1. Such high V̇O2max values, not the GE we reported, are beyond human limits during cycling. The V̇O2-W (W) relationship is attenuated at high workloads (>300 W) in top-level professional cyclists (5,6). Thus, linear regression equations are, most likely, not suitable to estimate V̇O2 at heavy intensities in these athletes. Thus, we do not have any reason to suspect a methodological error when calculating GE. The V̇O2 cost of ∼4.51 L·min−1 that Jeukendrup et al. have adequately estimated from GE data (Table 1 of reference 4) is an average for the 20-min bouts. These were performed at a constant power output similar to that eliciting subjects’ 80%V̇O2max during the previous ramp test. The V̇O2 slow component (i.e., the continuous rise in V̇O2 that inevitably occurs during heavy constant-load exercise) (12) was responsible for the 20-min mean V̇O2 to surpass 80%V̇O2max. Jeukendrup et al. mentioned some studies with elite cyclists, e.g., reference 3. In a well-conducted and controlled study, Lee et al. (3) tested Australian professional riders with mean V̇O2max (5.4 L·min−1) comparable to our subjects (∼5.3 L·min−1) (4). Their GE and blood lactate levels averaged ∼22% and 6.1 mM during a 30-min bout at a mean power output of 370 W, i.e., ∼10% lower and 52% higher, respectively, than those of our subjects while generating 385 W for 20 min (24.5% and 2.9 mM). Does the unusually low blood lactate concentration of our cyclists (not reported before in this remarkable type of effort) reflect a systematic error (−52%) of our lactate analyzer? The uniqueness of the population group (e.g., a two-time world champion and some top-10 finishers and stage winners of the Tour), not a methodological error, is the most likely explanation for the unusually high GE and/or low lactate levels at heavy workloads found by us compared with previous valuable studies with elite/professional cyclists, yet of lower performance level. In fact, we have recently reported GE and lactate values to average ∼22% and 8–9 mM at 373 W in cyclists who are professionals but unable to finish the Tour in top positions (11). The superior performance of the best distance runners (i.e., Africans) is largely attributable to their greater economy/efficiency. Why couldn’t a similar phenomenon occur in the best cyclists? Alejandro Lucia Jesús Hoyos Margarita Pérez Alfredo Santalla Manuel Rabadán José L. Chicharro