Top international artistic gymnastics is characterized by an ongoing increase in the level of difficulty of the routines and, as a result, an increase in the duration and intensity of the routines (1–3). Therefore, beyond a high technical skill level, simultaneous development of well-developed gymnastics-specific muscular strength and endurance are essential to meet the increased demands of the competition (4). Regarding gymnastics-specific endurance, there are clear indications that the demands on the cardiorespiratory and metabolic system are higher and more complex than previously assumed (5). In this context, previous studies based on routine duration and blood lactate concentrations (BLC) estimated that the aerobic metabolism is utilized by 3%–8% during pommel horse, still rings, and bar routines and 20%–30% during floor routines (6, 7). Recent studies on the physiological and energetic demands on the floor, pommel horse, and still rings contradict these assumptions (8–10). According to these studies, the percentage aerobic energy supply is 23.3 ± 4.1% on pommel horse (8), 28.6 ± 4.8% on still rings (9), and 54.4 ± 6.8% on floor (10). These results are consistent with studies confirming a notable aerobic energy supply (between 43% and 49%) in other sports with similar routine durations and intensity profiles (i.e., indoor climbing, 60 s judo matches, or rhythmic gymnastics) (11–13). Due to these differences between the earlier estimates and recent findings, there is still no consensus regarding the specific energetic demands or the most effective training method for gymnastics-specific endurance performance. While some authors do not attribute any importance to aerobic metabolism (6), other authors point out that aerobic performance training should not be neglected (14–16). Currently, training regimes to improve the gymnastics-specific endurance performance focused either on strength training to improve maximal strength capabilities (17), or on high-intensity interval training (HIIT) to increase the endurance capabilities (i.e., VO_2max, muscle endurance) (18–21). However, it is currently unclear whether HIIT improves specific endurance performance in artistic gymnastics and which physiological factors are important for achieving this kind of improvement.

As mentioned previously, aerobic metabolism plays a significant role in the energy supply in artistic gymnastics. An improvement in aerobic performance can lead to a modification in energy supply and thus create an energetic power reserve at a given routine load. Increasing the proportion of aerobic metabolism could potentially save the limited high-energy creatine phosphate (PCr) and reduces lactate and H⁺ production and the breakdown of muscle glycogen (22). Reduced lactate and H⁺ production minimizes the reduction in muscle pH and thus the energetic factors of muscular fatigue caused by anaerobic lactacid metabolism, as the reduction in muscle pH interferes with biochemical and physiological processes (23). To achieve an increase in the aerobic metabolic rate, it is important to speed up the aerobic metabolism, which is determined by the oxygen uptake (VO₂) kinetics (24). Faster VO₂ kinetics means a smaller O₂ deficit and therefore less PCr depletion and less production of lactate and H⁺ by the anaerobic lactic metabolism, which is associated with an increase in exercise tolerance (22, 24). The rate of VO₂ kinetics is mainly limited by the intramuscular part (enzyme activity and content and mitochondrial volume and content) of VO₂ (25). Notably, HIIT appears to be an effective method of speeding up VO₂ kinetics (22, 24). Studies have shown a 17%–34% reduction in the time constant (τ₁) of VO₂ kinetics during moderate and heavy exercise for both upper and lower body exercise (22, 25–28). However, it is unclear how HIIT influences VO₂ kinetics during gymnastic specific endurance tests and whether these effects can be transferred to achieve the desired modification of energy supply during artistic gymnastic routines.

The aim of the study was therefore to analyze the effectiveness of HIIT on gymnastics-specific endurance performance and the change in energy supply using floor and pommel horse routines as examples. Based on the theoretical positions, it was hypothesized (H1) that HIIT integrated into gymnastics training would lead to a greater improvement in performance in gymnastics-specific endurance tests compared to normal gymnastics training. The second hypothesis (H2) was that HIIT integrated into gymnastics training would lead to a greater increase in the relative aerobic energy component and a greater reduction in the anaerobic energy component of gymnastics routines on the floor and pommel horse compared to normal gymnastics training. In order to examine the possible factors of the effectiveness of HIIT, it was examined whether there are correlations between physiological factors (e.g., O₂ deficit) and an improvement in endurance performance. In this context, it was hypothesized (H3) that the O₂ deficit correlates with the performance parameters in the gymnastics-specific endurance tests. The final hypothesis (H4) tested was that the O₂ deficit correlates with the aerobic and anaerobic energy contribution of the floor and pommel horse routines.

Tables 2, 3 show the results of the RJT and CT. The integrated HIIT in the normal gymnastics training led to a significantly greater improvement in Jump_mean (Z = 17.00; p = 0.011; r = −0.511) and Jump₉₀ (Z = 13.00; p = 0.005; r = −0.587) in the RJT. In addition, the intervention group showed a significantly greater reduction in the O₂ deficit (Z = 10.00; p = 0.002; r = −0.644), τ₁ (Z = 11.00; p = 0.002; r = −0.625), and delta lactate (Z = 23.00; p = 0.045; r = −0.399) in the RJT (Table 2). In the CT no significant difference was found between the groups for these parameters (Table 3). There was also no significantly greater improvement in C_max (Z = 48.00; p = 0.575; r = 0.197) with the integrated HIIT compared to normal gymnastics training (Table 3).

Tables 4, 5 present the physiological and energetic parameters of the floor and pommel horse routines. On the floor, the intervention group showed a significantly greater increase in VO_2mean (Z = 18.00; p = 0.027; r = −0.461) and VO_2peak (Z = 16.00; p = 0.017; r = −0.503) (Table 4.). On pommel horse, a significant reduction in HR_mean (Z = 10.00; p = 0.020; r = −0.551) and HR_peak (Z = 9.00; p = 0.009; r = −0.604) were achieved by the intervention group (Table 5). Regarding energy supply, the HIIT leads to a significantly greater increase in aerobic metabolism (Z = 14.00; p = 0.010; r = −0.544) and a significantly greater reduction (Z = 14.00; p = 0.010; r = −0.544) in anaerobic metabolism during floor routine. In addition, a significantly greater increase in aerobic metabolism (t = 9.00; p = 0.007; r = −0.604) and a significantly greater reduction was detected for anaerobic metabolism on the pommel horse (Z = 9.00; p = 0.007; r = −0.604). The HIIT group was able to increase the aerobic energy supply on floor by 2.69 ± 4.13% on average and at the same time reduce the anaerobic energy supply by −2.68 ± 4.13%. On the pommel horse, the aerobic energy supply in the HIIT group increased by 2.78 ± 2.16% on average with a simultaneous reduction in the anaerobic energy supply of −2.77 ± 2.16%. For both floor (Z = 15.00; p = 0.013; r = −0.461) and pommel horse (Z = 13.00; p = 0.025; r = −0.499), the intervention group shows a significant reduction in WPCr.

Looking at the physiological factors for the effectiveness of HIIT, the correlation analyses revealed a significant negative correlation (Rho = −0.73; p =  < 0.001) between the difference in O₂ deficit and the difference in Jump_mean in the RJT (Figure 2). There is also a significant negative correlation for the difference in O₂ deficit and the difference in Jump₉₀ (Rho = −0.58; p = 0.010). The difference in τ₁ also shows a significant negative correlation with the difference in Jump_mean in the RJT (Rho = −0.74; p = < 0.001) as well as with the difference in Jump₉₀ (Rho = −0.67; p = 0.002).

Furthermore, there is a significant negative correlation (Rho = −0.60; p = 0.020) between the difference in O₂ deficit in the RJT and the difference in the relative W_AER on floor (Figure 3). There is also a significant negative correlation (Rho = −0.61; p = 0.019) for the difference in τ₁ in the RJT and the difference in the relative W_AER on floor (Figure 3).

In CT, there is a significant negative correlation for the difference in O₂ deficit and the difference in C_max (Rho = −0.58; p = 0.008) and a significant negative correlation for the difference τ₁ and the difference in C_max (Rho = −0.55; p = 0.013) (Figure 4). No significant correlations were found for the differences in the physiological parameters in the CT and the differences in the relative energy fractions on pommel horse.

The study aimed to examine the effectiveness of HIIT integrated into normal gymnastics training on gymnastics-specific endurance performance and energy supply during floor and pommel horse routines. Contrary to the first hypothesis (H1), the HIIT only leads to a greater improvement in gymnastics-specific endurance performance in the RJT (i.e., Jump_mean), however, the HIIT did not increase the maximum number of circles in the CT. The increase in Jump_mean is in contrast to a previous study that reported no effects of HIIT on mean jump height in a 30-second RJT (seven young female sub-elite aerobic gymnasts) compared to a control group performing normal acrobatic training (41). The reason for the inconsistent results of the studies could be explained by the different loading times of the RJT and thus the different energy supply. The anaerobic part of the energy supply is presumably higher during a 30 s RJT compared to a 90 s RJT, so the effects of HIIT (i.e., increase in aerobic power) on the average jump performance did not fully apply. Improving the mean jump height in an RJT is important as gymnasts must be able to generate a similar amount of net momentum and force throughout the floor routine (30). The aforementioned positive association is strengthened by the study of Marina and Rodríguez (30), where the mean jump height in a 60 s RJT is positively correlated (Rho = 0.98) with the judge's score on the floor in twenty teenage elite artistic gymnasts. A greater improvement in peak jumping performance, expressed in Jump_peak could not be achieved with the integrated HIIT. However, both groups achieved an improvement in Jump_peak in the post-test (HIIT = +1.62 ± 3.92 cm; CON = +1.15 ± 1.83 cm), which can presumably be attributed to the normal gymnastics training in the preparation phase. As the aim of the study was to improve the gymnastics-specific endurance performance and thus Jump_mean, this result is understandable. In a study with aerobic gymnasts, integrating HIIT into normal training also did not result in a significantly greater improvement in peak jumping power performance in a counter-movement jump compared to a control group with only normal acrobatic training (42). However, the same study showed a significantly greater improvement in peak jump power performance for the jump interval training (JIT) group compared to the control group with only normal training (42). So, if the training goal is to achieve a greater improvement in peak jumping performance compared to normal gymnastics or acrobatics training, JIT should be used. HIIT and JIT appear to produce different adaptations that should be used as specifically as possible, especially in high-performance sport. JIT requires explosive force generation, which more effectively stimulates fast-twitch muscle fibers than running intervals, resulting in an enhanced neuromuscular power, which is crucial for improving maximum jump height (42). HIIT, on the other hand, results in adaptations at physiological energetic levels (oxidative adaptations), which appears to be crucial for the increase in mean jumping power during RJT. The present study found a significant correlation between the increase in mean jump height and the reduction in the O₂ deficit as well as between the increase in mean jump height and the reduction in the time constants of VO₂ kinetics, which partially confirms the third hypothesis.

As mentioned previously, a greater increase in C_max through the integrated HIIT was not detected. Possible reasons for this could be related to the used CT or the designed HIIT. In addition to muscular fatigue, there may have been other influences on C_max in the CT. The fatigue at the end of the CT may have resulted in a poor grip and thus impaired balance, which could have resulted in the CT being stopped without complete fatigue being achieved. However, this effect can also be transferred to the control group. The analysis of the HR data from the HIIT shows that the required exercise intensity of 90% of HR_max was mostly not achieved in the upper body exercises. This can be explained by lower muscle mass in the upper body, which leads to differences in metabolic and cardiovascular responses compared to leg training (28). Arm cranking shows slower HR kinetics compared to leg training, probably due to lower sympathetic stimulation of the heart during arm training (43). Achieving a high intensity in HIIT is important as only exercise intensities near VO_2max allow for both large motor unit recruitment (i.e., type II muscle fibres) and attainment of near-to-maximal cardiac output, which, in turn, together represent signals for oxidative muscle fibre adaptation and myocardium enlargement (44). Therefore, it can be assumed that the training stimulus for an adaptation of aerobic performance in the upper body was not sufficient. Another explanation is that circles are too specific movements to be trained by the implemented HIIT. Although one study showed that HIIT can significantly increase maximum numbers of circles, the HIIT carried out in this study also included 40 s circles as one of the five exercises (20). In the present study, circles were not used as a HIIT exercise, as gymnasts are often affected by wrist problems due to pommel horse training. In addition to physiological and test-related factors, the non-randomized design and sample size may also have limited the statistical power to detect changes in C_max.

The second research hypothesis, that semi-specific HIIT increases the relative aerobic energy contribution during floor and pommel horse routines, was confirmed. The greater increase in aerobic metabolism reduced anaerobic metabolism and thus offers the possibility of a) creating energetic performance reserves (conservation of PCr stores) and b) reducing factors of muscular fatigue. This provides the conditional basis for increasing exercise difficulty and ensures that a high exercise quality is maintained throughout the exercise. These aspects highlight the importance of developing specific endurance performance in artistic gymnastics. The modification of energy supply through HIIT is consistent with the results of other studies (45–47). Some of these studies showed that the change in energy supply also resulted in a reduced O₂ deficit (22, 45, 46). In the present study, the O₂ deficit was not calculated for the floor and pommel horse routines. However, there is a correlation between the reduction of the O₂ deficit in the RJT and the modification of the metabolic components on the floor. This confirms the fourth hypothesis. The reduction in the O₂ deficit can be attributed to a faster VO₂ kinetic.

The energy data also show that the integrated HIIT primarily reduces the energy of the anaerobic alactic metabolism by increasing the aerobic metabolism. The anaerobic lactic metabolism showed no significant reduction in the relative energy content. Thus, the results are somewhat contrary to the results of Park and Yang (47), who reported that a 4-week integrated HIIT led to a significant reduction in the relative anaerobic lactic energy content in a high-intensity cardio yoga test (ten physically active individuals). This discrepancy may be due to the different energy demands of the different types of exercise. For example, the relative W_PCr for the floor and pommel horse routines examined in this study is around 30%, while for the high-intensity cardio yoga test it is around 10%. The fact that different types of exercise have different energy demands was also demonstrated by Kaufmann, Hoos (48), who showed that a 30-second RJT depends more on alactic and less on lactic energy contribution than a Wingate test. Because anaerobic alactic capacity is an important component of gymnastic-specific endurance (8, 9), reducing W_PCr is an important training adaptation. In this way, HIIT can be performed to save PCr, which can be used as an energy reserve for increasing exercise difficulty.

Based on the determined correlation, conclusions about the physiological reason for the effectiveness of HIIT in modifying energy supply and improving specific endurance performance on the floor can be made. Notably, an acceleration of VO₂ kinetics is mainly due to an improvement in the intramuscular part (enzyme activity and content, and mitochondrial volume and content) of VO₂. Aerobic power, expressed as maximum VO₂, is made up of the power of the pulmonary, cardiovascular, and muscular systems (49). Therefore, it can be concluded that especially the intramuscular part of aerobic performance is an important component of gymnastic-specific endurance performance. Using this knowledge, future gymnastics-specific endurance training programs and performance tests can be designed more specifically to stimulate or identify corresponding physiological adaptations.

When interpreting the results, however, the following limitations of the study must be taken into account. The study was conducted with twenty-five male junior high-level gymnasts. Due to the small sample size, the generalizability of the results is limited. The selection of exclusively male junior high-level gymnasts limits the transferability to other groups of people (junior female gymnasts or senior male gymnasts), meaning that the validity and significance of the results can only be related to the group of people and performance level studied. Future studies with a larger participant group and an age- and gender-related expansion of the participant group are recommended to increase the transferability of the results. Instead of randomization, group selection in this study was based on the training location, which could not be realized otherwise in professional training practice. Nevertheless, the insufficient randomization leads to a reduction in internal validity and could have caused a selection or location effect. Further limitations of the study include the use of the Fox formula to determine HRmax, which is less accurate in adolescents, and the matching of the total training duration of the groups instead of the training volume. This approach was taken to enable the practicability of such a study in a professional sports context, as, for example, an additional HRmax test could not be performed due to time limitations. The experimental process was routine in a training competition, therefore, the available data should be compared with future studies. When calculating the energy contributions, it was assumed that the replenishment of PCr stores after exercise is entirely attributable to the aerobic system, whereby a possible minor contribution of the glycolytic system to PCr resynthesis was neglected (48). Another point that needs to be addressed is the calculation of the energy supply during the pommel horse routine. For the pommel horse routines, the VO₂ data during the short breaks were neglected as a potential W_PCr contribution, as the recovery time was very short and the athletes jogged during the breaks, which maintained the increased VO₂. Thus, there was no rest period in which the PCr could have been replenished, so these values were considered as oxidative phosphorylation energy (50). Since this study is mainly concerned with the difference between the intervention and control groups, the aspects of energy calculation can be neglected.

It was shown that HIIT resulted in a greater increase in the relative aerobic and a greater reduction in the relative anaerobic energy component during floor and pommel horse routines. The modification of the energy supply can have a positive influence, as it can create energetic performance reserves and reduce factors of muscular fatigue. It was also shown that HIIT increased Jump_mean in a 90-s RJT. The average jump height is an important performance parameter for gymnastic-specific endurance on the floor (30). HIIT fosters faster VO₂ kinetics and thus a reduced O₂ deficit. Based on this, it can be concluded that the muscular part of aerobic performance is an important component of gymnastics-specific endurance performance. Contrary to assumptions before the study, a greater improvement in C_max through HIIT could not be confirmed. This could be due to upper body exercises used in the HIIT, which mostly failed to achieve the desired exercise intensity. For this reason, future studies could examine whether a modified exercise selection or a modified load profile can positively influence endurance performance on the pommel horse.