A preview of this full-text is provided by Wiley.
Content available from Scandinavian Journal of Medicine and Science in Sports
This content is subject to copyright. Terms and conditions apply.
Scand J Med Sci Sports. 2023;33:99–100.
|
99
wileyonlinelibrary.com/journal/sms
Received: 20 June 2022
|
Revised: 12 September 2022
|
Accepted: 13 September 2022
DOI: 10.1111/sms.14259
LETTER TO THE EDITOR
Critical power: Evidence- based robustness
Recently, Gorostiaga et al.1 published an article entitled
“Over 55 years of critical power: Fact or artifact?” with the
objective of “generating an evidence- based debate” regard-
ing the validity of the critical power (CP) as a physiolog-
ical gold- standard parameter. The researchers critically
examined the concept of the CP model under four main
arguments; “(1) the inconsistency in experimentally
demonstrating the classical definition of CP, (2) the wide
range of relative intensities at which CP was identified
to occur, (3) the choice of duration of exercise trials used
to assess CP, and (4) inadequate choice of the hyperbolic
function model, despite the power function model, to de-
scribe the relationship between velocity and time.” These
arguments were originally presented using the best race
time performances of the top 10 male and female Spanish
runners who completed 1.5km to 10km distance races in
the same outdoor competitive season (Year 2019). In this
way, the critical speed values (CS, analogous to CP) were
determined using the race times up to 5km (CS1.5– 5km, 1.5,
3, and 5km) and 10 km (CS1.5– 10km, 1.5, 3, 5, and 10km).
In addition, the researchers added the half- marathon
and marathon performance times of other athletes to the
model. However, we have identified some serious limita-
tions with the data on which their conclusions are based.
Our objectives are to point out important limitations and
misconceptions in the original article by Gorostiaga et al.,1
which distorted the models and conclusions, and highlight
the robustness of the CP model as the lowest intensity
(speed) that represents the severe- intensity domain, when
using a correct and standardized methodology.2 Using the
values (time and distance) provided in the article, we re-
calculated the CS and determined the D′ values (fixed and
limited distance capacity above the CS) through the same
model used (distance– time). CS (CS1.5– 5km, 5.45 ± 0.38 m/s
vs. CS1.5– 10km, 5.19 ± 0.39 m/s; p= 0.02) and D′ (CS1.5– 5km,
140.6 ± 60.1m vs. CS1.5– 10km, 272.2 ± 110.2m; p=0.001) val-
ues were significantly different (paired t- test). In fact, the
influence of duration / distance of predictive trials on the
CP/CS parameters has been previously deomonstrated.3
Interestingly, during modeling we noticed that the av-
erage speed of four athletes (TN, MM, CC, and CV) in the
3km race was lower than their average speed in the 5km
race. This is interesting because the hyperbolic relationship
between speed and time dictates that speed should decrease
with increasing distance up to the CS. Yet, in the best perfor-
mance of these athletes, they ran at a faster average speed
in a race with greater distance. In this case, several factors,
such as “pacing,” may have prevented the complete deple-
tion of the D′ above the CS, thus, confounding the CS1.5- 5km
model prediction. In fact, it was not possible to predict the
3 km running performance for these athletes (i.e., TN, MM,
CC, and CV) using CS1.5– 5km parameters. In the 10- km race,
only two athletes (DP and MJ) ran at average speeds above
CS1.5- 5Km, showing that all other athletes were at intensities
below CS, justifying the running time (average, 31.54 min-
utes). The potential overestimation of D′ may explain why
the actual times (Mean ± SD: 1.5 km, 247.5 ± 22.8, 3 km,
533.2 ± 44.0, 5 km, 893.0 ± 67.4, and 10 km, 1892.3 ± 144.6s)
were statistically (t- test, p=0.05– 0.0002) different from the
predicted times (Mean ± SD: 299.2 ± 59.26, 566,0.9 ± 75.1,
652.3 ± 131, and 2215.7 ± 304.9 s, respectively) when
used CS1.5– 10km model. In this sense, the inclusion of ex-
ecution times at intensities slightly below the CS (10km,
half- marathon, and marathon) introduces errors in the
mathematical modeling, which do not represent real
physiological consequences on performance, possibly by
reaching domains of different intensity. In addition, due to
numerous influences during a competitive sporting season
(e.g., time of day, environmental conditions, training status,
and performance level of competing athletes),4,5 it does not
seem correct and valid to use different sporting events for
CS modeling, particularly paced race times.6
There is evidence that the time trial protocol, similar
to the time to exhaustion protocol, is valid for determin-
ing the CP/CS model, as long as it induces exhaustion to
a maximal of ~15 min, aiming both to deplete D′ and to
attain VO2max.7 Indeed, several key parameters of aerobic
fitness (VO2max/VO2peak, blood lactate response to exer-
cise, running economy, and oxygen uptake kinetics) are
also protocol- dependent, that is, data obtained by different
protocols should not be used interchangeably.
Thus, it is important to emphasize that (1) Gorostiaga
et al.1 used severe- intensity race efforts that were per-
formed submaximally in several cases (40% of athletes in
the 3 km race); (2) 80% of the athletes were below the CS in
the 10km race (according to the model with 3 distances);
© 2022 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.