Active vs. passive recovery during an aerobic interval training session in well-trained runners

High-intensity interval training (HIIT) is a method that consists of performing repeated bouts (i.e., work intervals) above the anaerobic threshold, interspersed with pauses, thus allowing the accumulation of more time at a targeted intensity during a single training session (Billat et al. 2000; Tuimil et al. 2011; Buchheit and Laursen 2013). Following Buchheit and Laursen (2013), HIIT protocols are often performed at intensities close to the maximum oxygen consumption (V̇O_2max), with shorter bouts (≤ 60 s) typically performed above V̇O_2max, and longer bouts (> 60 s) performed at or below V̇O_2max. Based on the intensity of bouts, HIIT can be performed taxing more the anaerobic (e.g., sprint interval training, SIT) or the aerobic (i.e., aerobic interval training, AIT) metabolism (Buchheit and Laursen 2013; Schoenmakers et al. 2019). In this regard, Hill and Rowell (1997) had previously suggested in the late 90s that bout duration should be, at least, > 60% of the time limit (T_lim) at the velocity associated to V̇O_2max or maximum aerobic speed (MAS) for reaching V̇O_2max in AIT running sessions. Thus, prescription of HIIT sessions includes the manipulation of intensity and duration of bouts and recovery intervals for designing appropriate training workloads when looking for specific metabolic and neuromuscular adaptations in the short and the long term (Buchheit and Laursen 2013).

Pause duration is determined by the duration of the bout and its intensity, to allow a proper recovery for successfully performing the subsequent work interval, and it is based on the work-to-rest ratio or the return of some physiological parameters (e.g., heart rate – HR, muscle oxygen consumption) to a set value (Foster et al. 2017; Schoenmakers et al. 2019; Schoenmakers and Reed 2019; Fennell and Hopker 2021b). While duration of pauses appears not to be important for the total physiological strain of a HIIT session unless the recovery duration was very short (Smilios et al. 2018; Schoenmakers et al. 2019), the intensity of pauses, with selection of active recovery (AR) vs. passive recovery (PR), is an issue which is still under debate. In this regard, Buchheit and Laursen (2013) suggested that AR could be recommended for longer recovery intervals of 3–4 min, while PR could be more beneficial for shorter recovery intervals of < 3 min. These recommendations were based not only on the plausible positive effect of AR on oxygen kinetics, and therefore on the total time spent in the so-called “red zone”, which implies to attain an exercise intensity of at least 90% of V̇O_2max (Midgley et al. 2006; Buchheit and Laursen 2013), but also on its negative effect on performance capacity as AR can lower muscle oxygenation (Dupont et al. 2007; Buchheit et al. 2009) and phosphocreatine re-synthesis (Spencer et al. 2006) when interval recovery time is < 3 min. In fact, recent scientific evidence demonstrates that longer AR durations (i.e., 4 min) reduce blood lactate concentration significantly more than shorter ones (i.e., 2 min), suggesting that a short AR may activate anaerobic glycolysis to a greater extent in the subsequent work intervals as a result of accumulated fatigue (Smilios et al. 2018). Paradoxically, these previous recommendations were mostly based on studies performing SIT (Dupont et al. 2007; Buchheit et al. 2009) and not typical AIT sessions. Thus, several studies have investigated the effects of AR and PR on running performance (Tardieu-Berger et al. 2004; Abderrahman et al. 2013), accumulated time in the “red zone” (i.e. > 90% V̇O_2max) (Thevenet et al. 2007a, b) and on metabolic and hormonal parameters (e.g., plasma glucose and gluco-regulatory hormones, such as adrenaline and noradrenaline) (Abderrahmane et al. 2013; Abderrahman et al. 2018), during SIT-based running sessions (e.g., 2 × 8–15 × 30 s at 100–110% of MAS, interspersed with 30 s AR at 50% of MAS or PR). Taking together, these previous studies suggest that time spent in the “red zone” did not significantly differ between recovery modes, whereas PR may lead to greater running performance (i.e., longer T_lim or more repetitions performed) and AR can improve some metabolic and hormonal parameters.

More recently, it has been suggested that PR leads to a greater performance during HIIT sessions in both SIT and AIT modalities (Perrier-Melo et al. 2020). Specifically, a recent study with cyclists showed that PR during AIT sessions may facilitate a greater external training load for similar physiological responses while lowering perceived exertion (Fennell and Hopker 2021a). Further, previous evidence has highlighted the important role that perception of effort plays during high-intensity aerobic exercise for exercise tolerance among highly motivated individuals (Marcora et al. 2010). Additionally, AR after self-paced long bouts of cycling exercise (e.g., 4 min or 8 min) may elicit longer time accumulated in the “red zone”, especially when the work-to-rest ratio is 2:1 (Dall’ Agnol et al. 2021).

On the other hand, previous studies have demonstrated the appropriateness of simple explosive exercises involving the stretch–shortening cycle as jumps (i.e., countermovement jump, CMJ) for assessing lower limbs neuromuscular function in endurance runners of different levels (Boullosa and Tuimil 2009; Boullosa et al. 2011, 2018; García-Pinillos et al. 2015, 2021). In this regard, it has been also suggested that a better potentiation/fatigue balance would be expected in endurance-trained athletes during and after conditioning activities as is the case of interval training running sessions (Boullosa et al. 2018). However, to the best of our knowledge, there is no evidence on the role that intensity during recovery might play for perceptual, physiological and neuromuscular responses during AIT running sessions, particularly when performed in the field by well-trained endurance runners. Previous studies focused on the effects of SIT sessions with different populations rather than endurance runners. This knowledge is important from a practical point of view to better select the mode of recovery depending on the session objectives.

Therefore, the aim of this investigation was to compare the use of AR vs. PR during a typical AIT running session (4 × 2 min at 100% of MAS, with a work-to-rest ratio of 1:1) performed on a 400-m outdoor track on metabolic (i.e., cardiorespiratory and blood lactate [La]), neuromuscular (i.e., jump capacity) and perceptual (i.e., rating of perceived exertion, RPE) responses in well-trained endurance runners. Based on previous literature (Fennell and Hopker 2021a), our hypothesis was that PR would favor a lower internal load (i.e., physiological and neuromuscular effects: V̇O₂, HR, [La], RPE, jump height/jump power) for a similar external training load (i.e., volume × intensity: total covered running distance, running speed).

Table 2 shows VO₂ demands, both in absolute and relative (% V̇O_2max) values, total covered distance, time spent in the “red zone”, and time-to-reach the “red zone” during bouts in AR and PR sessions. Higher mean V̇O₂ values (r = 0.79; 95% CI [0.333,1]) and mean V̇O₂ relative to V̇O_2max (G = 0.479; 95% CI [0.142,0.817]) were recorded during the session with AR, whereas higher peak V̇O₂ values relative to V̇O_2max were obtained during the PR session (G = 0.345; 95% CI [0.039,0.655]). Oxygen kinetics during bouts and recovery intervals from the same runner during both sessions is shown in Fig. 2.

Absolute mean V̇O₂ and mean V̇O₂ expressed as a percentage of V̇O_2max throughout the bouts are represented in Fig. 3.

Recorded HR values during running bouts were significantly higher during AR when compared to PR for mean HR (158.56 ± 7.92 vs 151.81 ± 5.48 bpm, p = 0.001; G = 0.853, 95% CI [0.348, 1.358]) and peak HR (169.28 ± 3.63 vs 166.34 ± 4.59 bpm, p = 0.05; G = 0.636, 95% CI [0.166, 1.106]). A significant effect of bout (p < 0.001; pη² = 0.993), session (p = 0.003; pη² = 0.648) and session × bout interaction (p = 0.01; pη² = 0.510) were found for mean HR throughout bouts. Post hoc comparisons detected higher values in PR in comparison with AR for bout 2 (p = 0.001; G = 1.122, 95% CI [0.565, 1.680]), bout 3 (p = 0.002; G = 0.842, 95% CI [0.339, 1.344]) and bout 4 (p < 0.001; G = 1.134, 95% CI [0.574, 1.694]). Similarly, mean HR during recovery intervals was significantly higher during AR when compared to the PR session (142.64 ± 8.49 vs 114.64 ± 7.53 bpm, p = 0.001; G = 3.194, 95% CI [2.077, 4.311]).

A significant effect of bout (p < 0.001; pη² = 0.883) and session (p = 0.016; pη² = 0.454), with no session × bout interaction, was found for [La] recorded 3 min after finishing the last bout during the sessions, with higher values for PR when compared to AR session (6.93 ± 2.22 vs 6.24 ± 1.93 mmol·L⁻¹, respectively).

Table 3 shows the changes for CMJ parameters during sessions. The ANOVA did not detect any significant session × bout interaction for maximum jump height (p = 0.199), mean jump height (p = 0.729), peak power (p = 0.245), and kvert (p = 0.289). A main effect of bout was only detected for mean jump height (p = 0.004; pη² = 0.443) and peak power (p < 0.001; pη² = 0.518).

Finally, descriptive and ANOVA results regarding RPE are presented in Fig. 4. A significant effect of bout (p = 0.002; pη² = 0.560) and session (p = 0.045; pη² = 0.343), with no session × bout interaction, was revealed with AR showing higher RPE values when compared to PR session (14.95 ± 1.66 vs. 14.05 ± 1.75, respectively).

To the best of our knowledge, this is the first study simultaneously analyzing cardiorespiratory, metabolic, neuromuscular, and perceptual responses during an AIT running session with AR vs. PR intervals. The main findings of the current study are: (1) as expected, the AR elicited higher mean V̇O₂, mean HR and peak HR levels during running bouts; (2) the peak V̇O₂ relative to V̇O_2max during bouts was higher with PR; (3) the mean V̇O₂ and HR values during recovery intervals were lower with PR; 3) the peak lactate concentration was globally lower with AR; (4) the jump capacity was potentiated, independently from recovery modes; (5) the perceived effort was higher with AR; 6) total covered distance was greater with AR.

The finding of greater mean V̇O₂ values in the session with AR is not surprising and agrees with previous literature. However, it should be noted that most previous studies used shorter supramaximal bouts (i.e., SIT) and not all of them were performed with runners. This is an important consideration as the training session selected, in terms of intensity and volume, could be considered more appropriate for the level of the recreational runners of our study (i.e., 10 km personal best of ⁓35 min). In this regard, O’Brien et al. (2008) demonstrated that a typical AIT running session consisting of 5 bouts of 2 min at MAS with 2 min AR at 50% of MAS leads to a higher mean V̇O₂ (3200 ± 661 mL/min) and more time spent in the “red zone” (4.47 ± 3.55 min) during the session than 1-min bouts (3076 ± 604 mL/min and 1.32 ± 3.94 min, respectively) or continuous running (2909 ± 584 mL/min) of equivalent external load. On the other hand, Seiler and Sjursen (2004) reported similar mean V̇O₂ values relative to V̇O_2max (⁓61% VO_2max) when performing 12 self-paced bouts of 2 min interspersed with 2-min self-paced recovery intervals compared to 1-min bouts in a sample of well-trained runners. The higher mean V̇O₂ values attained during bouts (ranging from 81 to 84% V̇O_2max in both conditions) could be explained by the differences in volume and intensity of our protocol in comparison to that used by Seiler and Sjursen (2004) (i.e., more bouts and self-paced). Moreover, peak V̇O₂ values relative to V̇O_2max reported by Seiler and Sjursen (2004) (92% V̇O_2max) were similar to those obtained in our study (ranging from 95 to 98% VO_2max in both conditions).

Taking together, this idea supports the previous recommendations of Hill and Rowell (1997) of using longer bouts during AIT sessions for maximizing the time spent in the “red zone” and therefore, improving endurance performance (Hill and Rowell 1997; Buchheit and Laursen 2013). Further, this would be reinforced in our study considering that the duration of prescribed bouts for both AR and PR sessions lasted longer (i.e., 120 s) than 60% of the T_lim at 105% of MAS (⁓106 s) during the verification phase. In this sense, we found non-significant differences between experimental sessions regarding TT90% V̇O_2max, TTpeak V̇O₂, TTR90% V̇O_2max and TTRpeak V̇O₂. However, and although these variables were quite similar for both conditions, we found longer TTpeak V̇O₂ in the PR session with confidence interval of the effect size not including null effect (p = 0.056; G = 0.341, 95% CI [0.022,0.660]). This phenomenon could be explained by a more pronounced increase in V̇O₂ rate for the PR condition due to a greater V̇O₂ amplitude (i.e., the difference regarding V̇O₂ demands between the start of a bout and the attainment of the “red zone”), resulting in a steeper V̇O₂ slope (i.e., better V̇O₂ kinetics; see Fig. 2 as an example) since the time-to-reach the “red zone” was similar for both conditions. Our results and rationale are in accordance with Fennell and Hopker (2021a), who also found a non-significant difference in time spent in the “red zone” between conditions in a recent investigation performed with well-trained cyclists. These authors argued that PR may reduce V̇O₂ demands at the start of the subsequent bout during AIT sessions, eliciting a higher V̇O₂ amplitude and reducing the time-to-reach not only a V̇O₂ plateau but also the “red zone”, thus suggesting better V̇O₂ kinetics (Fennell and Hopker 2021a). Further, it was previously demonstrated that PR facilitates a greater interval performance (e.g., capability of performing more bouts or maintenance of a higher speed or power in subsequent bouts) with a similar physiological stress when compared to AR in both SIT-based (Tardieu-Berger et al. 2004; Thevenet et al. 2007b; Abderrahman et al. 2013; Perrier-Melo et al. 2020) and AIT-based (Perrier-Melo et al. 2020; Fennell and Hopker 2021a) HIIT sessions. The fact that our runners completed only 4 bouts of 2 min at 100% of MAS, experiencing a significantly lower mean V̇O₂, mean HR, peak HR and RPE for a similar external load with PR, may suggest that they could be able to perform more bouts in this condition and, therefore, accumulate more time in the “red zone”. In this regard, it is also worthy to mention that total covered distance was significantly greater during the session with AR (i.e., ⁓ 1.2 km). However, this increment in total running distance should be cautiously interpreted, since it was performed at low intensities (i.e., 80% of vVT₂ during recovery intervals). In this sense, one must consider that total external training load (i.e., including bouts and rest intervals) will be greater for AR when compared to PR sessions. However, targeted training intensities during work intervals were identical for both conditions. Thus, from a practical point of view, the distance accumulated during AIT sessions with AR should be considered in the computed training volume. From the current results, it can be suggested the use of PR when targeting for accumulating a greater volume of high-intensity work intervals, whereas AR could be recommended for maximizing physiological stress during AIT running sessions of moderate volume. This suggestion is based on the lower mean HR and V̇O₂ values recorded during the PR intervals when compared to the active ones. Meanwhile, the correspondence between V̇O₂ and HR responses during running bouts does confirm the validity of HR as a simple and valid monitoring tool to verify the purported physiological adaptations during HIIT sessions.

On the other hand, peak lactate after the experimental sessions was significantly lower when AR was performed at 80% of vVT₂. These results are in accordance with previous studies which also demonstrated a better lactate clearance during AIT running sessions when active recovering at velocities close to the anaerobic threshold (Menzies et al. 2010).

One interesting finding was the post-activation performance enhancement evidenced by higher mean jump height and peak power values recorded at the end of the last bouts in both conditions. This finding is in agreement with previous reports with endurance runners of different levels and sex after different running exercises. Previously, García-Pinillos et al. (2015) observed that some runners (i.e., responders) exhibited jump potentiation but others did not (i.e., non-responders) during a 4 × 3 × 400 m intervals with 1 min of PR between bouts and 3 min between sets. While differences between sessions and used methods make comparisons difficult, our results show that the expected greater neuromuscular fatigue associated to AR was not observed as both sessions elicited similar responses. Further, the jumping height potentiation was accompanied by a preserved vertical stiffness which confirm a reduced neuromuscular fatigue during the sessions. This may suggest that the completion of additional bouts by our subjects would be feasible. Meanwhile, the regular use of vertical jump evaluations during HIIT sessions can be recommended to monitor the neuromuscular impact of different HIIT sessions to simultaneously evaluate the acute and chronic effects of different HIIT schemes (Boullosa et al. 2018; García-Pinillos et al. 2021).

Finally, we observed lower RPE scores for the PR condition at different time points. Moreover, recorded RPE scores constantly increased during bouts for both conditions. This is in accordance with previous studies since it was suggested that perception of effort increase linearly during AIT sessions and will be felt “hard” (i.e., RPE 15–16) initially and perceived as “very hard” (i.e., RPE 17–18) by the end of the workout (Seiler and Sjursen 2004). In addition, considering the lower RPE values throughout the session with PR, it would be hypothesized that runners in this condition would be able to complete more bouts until exhaustion, as perception of effort plays an important role in exercise tolerance during high-intensity aerobic exercise (Marcora et al. 2010). Further studies should be conducted to elucidate if other recovery work-to-rest ratios would influence the perceptual responses that can be easily recorded by practitioners during HIIT sessions. In this regard, it would be interesting to see how these RPE values would result in different session RPE (sRPE) values.

The present study has some limitations. Although participants were instructed not to change their habitually nutrition patterns, we did not control the diet during the course of the investigation. Another point to mention is the heterogeneity level presented by our runners, since 10 km personal best ranged from ⁓ 32 min to 38 min 30 s. The extrapolation of our results might not be appropriate for athletes with different characteristics (e.g., level, sex, etc.) and other AIT running sessions with different schemes and work-to-rest ratios.

For the first time, we have concurrently analyzed the cardio-metabolic, neuromuscular and perceptual response of AIT running sessions completed on a 400-m outdoor track at 100% of MAS but differing in recovery mode. The practical applications from the current results for coaches, sport scientist and athletes are different regarding training objectives. Recreational runners as those of our study could benefit from AR during AIT running sessions when looking for a higher physiological stress. In contrast, PR may elicit a lower physiological stress for a similar external training load and, therefore, could be preferable for high-volume AIT running sessions. However, it is still to be solved if the same picture would be evident with a longer session with more running bouts. Further, longitudinal studies are guaranteed to better understand the adaptations in the long-term to these recovery modes during AIT running sessions.