2.2. Protocols

All tests were conducted on a wind resistance braked rowing ergometer (Concept II model C, Morrisville, VT, USA). Performing the experiments on a stationary ergometer facilitated monitoring of the different measured parameters. The participants were accustomed to the use of the apparatus. The computer of the ergometer continuously displayed power output, stroke rate and distance traveled.

Each athlete performed three trials, except for one who performed only two because of an injury. A summary of the protocols and measured parameters is available in Figure 1A. The protocol consisted in performing a 1,500 m on rowing ergometer in three different conditions at least 48 h apart (excepted once where only 24 h separated two consecutive trials). The three visits were carried out as far as possible, at a similar time of the day, in the same prandial state, and in the same energetic status (diet). The order in which the conditions were carried out was randomly assigned.

Open in a separate window

Figure 1

Protocol design (A), mean power on each 500 m (B) and Ppeak time-courses (C). Values are mean ± SD. The symbols *, , and + indicate a significant difference between degr-prog, prog-stab, and stab-degr, respectively. The number of symbols (one, two, or three) indicate the p-value (<0.05, <0.01, and <0.001, respectively).

Each trial consisted of performing 1,500 m on rowing ergometer divided in three 500 m with a degressive (degr), progressive (prog), and stable (stab) profile/design in terms of mechanical power output performed during the test. Every 500 m was associated with a target power output determined as follows for each athlete and strategy:

• For degr : P₅₀₀ = 107%P_ref; P₁₀₀₀ = P_ref; P₁₅₀₀ = 93%P_ref.

• For prog : P₅₀₀ = 93%P_ref; P₁₀₀₀ = P_ref; P₁₅₀₀ = 107%P_ref.

• For stab : P₅₀₀ = P₁₀₀₀ = P₁₅₀₀ = P_ref.

where P_ref = 97.5% of the mean power output sustained during the best 2,000 m all-out performance test of the season. P₅₀₀, P₁₀₀₀, and P₁₅₀₀ correspond to the target power output of the first, second and third 500 m.

At the beginning of each visit, a hyperemic cream was applied to the earlobe for arterialization of the capillary blood. The athletes then performed a self-paced warm-up on land (flexibility), on a cycling ergometer, on a rowing ergometer, or a combination of these possibilities. On rowing ergometer, the warm-up included accelerations and peak strokes. The rowers self-selected their drag factor (identical drag factor for all tests/visits for a given athlete). Subjects were then equipped with the measuring instruments for the following parameters: gas exchanges, heart rate and peripheral oxygen saturation (see Supplementary Materials for details). The hyperemic cream was removed from their earlobe and a blood lactate sample was taken to measure the pre-trial blood lactate concentration.

For the first 500 m, subjects were instructed to reach the target power output (typically in 3–5 strokes) as quickly as possible. Once stabilized at the target power output, the rowers had to perform one maximal stroke (typically the fifth or sixth stroke) to determine Ppeak. Then rowers turned back to the target power. Right at 250 m and just before reaching 500 m, the athletes performed a maximal stroke to determine Ppeak, then returning to the target power output or stopped, respectively. The same procedure was repeated thereafter for the second and third 500 m. A total of 7 Ppeak were thus performed during the 1,500 m trial. A rest period between two consecutive 500 m exercise bouts allowed to take a blood lactate sample at the earlobe and the recording of the rate of perceived exertion (RPE, 1–10 scale). The duration of the rest periods was strictly limited to the time necessary for the measurements (15±4 s). A post-trial blood lactate sample was taken 3 min after exercise completion.

2.3. Measurements

Detailed protocols for the measurements of the mechanical (force, velocity, and power output), physiological [oxygen uptake ($\dot{V}{O}_2$), time-constant of the $\dot{V}{O}_2$ response during the first 500 m (τ), $\dot{V}{O}_2$ at steady state ($\dot{V}{O}_{2}SS$), oxidative contribution to energy supply (E_oxi), heart rate (HR), heart rate at steady state (HRSS), blood lactate concentration ([La]_blood), non-oxidative glycolytic contribution to energy supply (E_non−oxi), lactate area under the curve (AUC_La), peripheral oxygen saturation (SpO₂), and efficiency (ϵ)] and psychological [rating of perceived exertion (RPE), RPE area under the curve (AUC_RPE) and grading of protocols by athletes] parameters are available in Supplementary Materials.

3.2. Physiological Measurements

3.2.1. Oxygen Uptake

Results drawn from $\dot{V}{O}_2$ are reported in Figure 2. The time constant τ (Figure 2A) was significantly lower in degr than in prog. τ in stab did not significantly differ from the other conditions (Figure 2A). $\dot{V}{O}_{2}SS$ was lower in the 2nd and 3rd 500 m for degr than in stab and prog (Figure 2B). E_oxi for each 500 m bout was not different between conditions (Figure 2C).

Open in a separate window

Figure 2

τ (time constant of the primary response of the oxygen uptake on the first 500 m) (A), $\dot{V}{O}_{2}SS$ (B), E_oxi (C), and HRSS (D). Values are mean ± SD. The symbols * and + indicate a significant difference between degr-prog and stab-degr, respectively. The number of symbols (one or two) indicate the p-value (<0.05 and <0.01, respectively).

3.2.3. Blood Lactate Concentrations

Results drawn from [La]_blood are reported in Figure 3. Initial and final values of [La]_blood were similar for all conditions (Figure 3A). However, after the first 500 m, [La]_blood was higher for degr than for stab and prog, while no difference was observed between the two latter conditions. At 1,000 m, [La]_blood were higher in degr, lower in prog and in between in stab.

Open in a separate window

Figure 3

[La]_blood (A), AUC_La (B), E_non−oxi (C), and SpO₂ (D). Values are mean ± SD. The symbols *, , and + indicate a significant difference between degr-prog, prog-stab, and stab-degr, respectively. The number of symbols (one, two, or three) indicate the p-value (<0.05, <0.01, and <0.001, respectively).

AUC_La (Figure 3B) was significantly greater for degr than for the two other conditions. A higher AUC_La corresponds to a longer time spent at higher [La]_blood.

E_non−oxi (Figure 3C) was higher on the first 500 m for degr than for prog and stab. On the second 500 m, E_non−oxi was higher for degr than for prog with no significant difference with stab. Finally, for the last 500 m bout including the first 3 min of recovery, E_non−oxi was greater for prog than for the other conditions.

4.1. Muscle Function

From a mechanical point of view, the 1,500 m trials were tiring (as illustrated by the decrease of Ppeak throughout the trials) but not in the same way according to the trials. Indeed, while initial and final Ppeak were similar, Ppeak in degr was lower than in prog between 500 and 1,250 m and than in stab between 1,000 and 1,250 m. It is interesting to note, that the difference between the trials are the most apparent on the middle 500 m while this is the period where the power output was similar between trials. This suggests that muscle function at 1,000 m is still under the dependence of what has been done during the first 500 m (vide infra).

4.2. Physiological Responses

Pacing strategies affected physiological responses. Interestingly, and similarly to what has already been described for muscle function, the discrepancies between the strategies were most often apparent during and after the second 500 m, while the power output was similar in the different conditions.

The first surprising result was that during the second 500 m, $\dot{V}{O}_{2}SS$ was lower in degr than in stab and prog while the power output were the same in the three conditions. There is very few probabilities that this result is a type I error insofar as $\dot{V}{O}_{2}SS$ during the first 500 m in degr was also significantly (p < 0.05) lower than $\dot{V}{O}_{2}SS$ during the third 500 m in prog (data not shown). The present study does not allow to find a definite explanation for this result. However, taking into account (i) that the lower $\dot{V}{O}_{2}SS$ in degr were accompanied by higher [La]_blood, (ii) that higher [La]_blood are associated with lower pH (acidosis) (Stewart, 1983), and (iii) that acidosis inhibits oxidative phosphorylation in vivo (Jubrias et al., 2003), one cannot exclude that the lower $\dot{V}{O}_{2}SS$ observed in the present study at 1,000 m is the result of a lower muscle pH. However, further studies are necessary to confirm/infirm this speculation/hypothesis.

Taken into account τ and $\dot{V}{O}_{2}SS$, it resulted that oxidative energies were similar among conditions. These results are similar to those obtained by Hettinga et al. (2007) on 1,500 m cycling time trials conducted with three strategies similar to the ones used here, suggesting that pacing strategies have no significant impacts on the overall contribution of oxidative metabolism in energy supply.

During the first 500 m, [La]_blood increased faster in degr than in prog. This result indicates that the higher power output was not compensated by the faster increase in $\dot{V}{O}_2$ (i.e., the lower τ) in degr, resulting in higher [La]_blood. The increase of the differences in [La]_blood between the conditions after the second 500 m, while the power outputs were similar in the three conditions, may be explained by the delay of muscle lactate to reach the blood. At that point, it is tempting to link the time-course of [La]_blood with those of Ppeak since Ppeak curves mirror those of [La]_blood. In a previous study, Hogan et al. (1995) observed that an increase in [La]_blood induced a decrease of force production (muscle studied in situ). However, a causative link between elevated [La]_blood and low Ppeak is highly speculative, highly controversial, and unlikely in the present case.

In the present study, AUC_La was higher in degr than in stab, this later being itself higher than in prog. This latter result illustrates that the athletes spent a longer time at higher [La]_blood in degr than in prog and stab.

4.3. Psychological Responses

Results for RPE were reminiscent with those for [La]_blood. This is consistent with the association between [La]_blood and the perception of an effort (Borg et al., 1987). Moreover, AUC_RPE was higher in degr than in prog and stab, indicating that athletes spent a longer time at higher RPE in degr than in the two other conditions.

If it is difficult to infer whether there is a link between the RPE profile during the trials and the rating of conditions by the athletes, one may nevertheless report that for the athletes, the worse strategy was degr. At least, degr was the strategy they disliked the most. This result is relatively surprising since this is the most often strategy observed during the first 1,500 m of races (Garland, 2005; Muehlbauer et al., 2010; Chu et al., 2021). Four possible explanations may account for the discrepancy between the personal rating of the rowers and what they actually do during races, and one cannot exclude that the explanation lies in a combination of the following possibilities.

First, starting the race in the lead offers a certain psychological advantage. This situation gives (i) athletes confidence and (ii) a visual control on their opponents, allowing them to adapt and modulate their effort according to possible catching up of the other competitors. The poor rating of the degr strategy in the present study might then be explained by the absence of opposition and visual on this opposition.

Second, degr strategy could be advantageous for the boat speed from a mechanical point of view. This could be investigated through a mechanical model taking into account (i) the propulsive power generated by the athletes and (ii) the resistive properties of the (considered) boat. One of the resistive terms, namely the added mass, is proportional to the acceleration of the boat (Cabrera et al., 2006; Formaggia et al., 2009). Therefore, in a degr strategy, the initial added mass term would be of great magnitude at the beginning of the race. During the following deceleration, this added mass term would result in a propulsive contribution for the boat. However, such a model needs to be further investigated in order to determine and quantify the benefit, if it exists.

Third, there may be a physiological reason to perform higher power outputs at the beginning. As fatigue develops during the race, it may become more and more difficult to perform high power outputs. In other words, the high level of power outputs, generated in the early phase of a race in a degr scenario, might be impossible to produce later, especially during the last 500 m of a race, even if a more “energy-sparing” strategy (e.g., prog) was used at the beginning of the race. In accordance with this possibility, one may note that in the present study, the mean increase in power output in (prog) was of only +5.1 ± 0.2% instead of the required 7%. Indeed, it was difficult for some athletes to increase the power output in the final 500 m of the prog condition.

Last, this strategy may be the simplest for the athletes, insofar as the other strategies (prog or stab) would require rowers to know precisely their physical abilities and their energy reserve at any time of the race. Therefore, starting stronger and letting the deceleration be dictated by the occurrence of fatigue may ensure the athletes to provide a maximum effort, whereas starting with less power output could result in finishing the race without having used all energetic reserves. In favor of this last argument, Hoffmann et al. (2014) have shown that energy management on a 2000 m ergometer race was improved in the presence of a “rowing avatar” that indicated them the pace to follow during the simulated race.

To conclude, a contradiction exists between athletes' preferences and field observations. To understand this contradiction, it would be interesting/necessary to include a “performance” trial in a future study. Indeed, it would be of great interest to quantify the maximum effort achievable in each strategy configuration to be able to conclude on the most optimal strategy (vide infra 4.4 for additional words in that regards).

The authors would like to thank all athletes who volunteered to participate in the study and the Master 2 STAPS (Sport Sciences) students who made the experiments possible: Nikou Péron, Jeremie Aissaoui, Adrien Roge-Pech, Enzo Pisicchio-Brachini, Pauline Stocky, and Nils Laydevant.