Most of the training for elite middle- and long-distance runners is performed at low intensity (1–3). However, especially in the Olympic middle distance running events, a part of the training is performed at moderate to high intensity (4). In addition to conventional threshold training (5), interval training sessions aimed at increasing maximal oxygen uptake (VO_2max) are still completed to enhance performance. The optimal effectiveness of these training sessions are described by research groups as involving a high time spent near maximal oxygen uptake (>90%–95% of VO_2max) (6, 7), which is also reflected in some international training models (8, 9). Although the optimal intensity range for interval sessions aimed at developing VO_2max has not been conclusively scientifically determined (7, 10–12), and overall training volume and altitude training should also be considered, the use of this intensity range in research and practice indicates high relevance.

Several studies (13–15), as well as reviews by Buchheit and Laursen – across disciplines and performance levels (6–16) and Parmar et al. (12), suggest that in running interval training, the use of short interval durations (<60 s) tends to be less successful in achieving high levels of maximal oxygen uptake and maintaining these levels above 90% or 95% of VO_2max for comparable durations to longer intervals. For example, Cipryan et al. (13) compared the physiological strain of long intervals (4 × 3 min) vs. short intervals (21 × 30 s) at the same intensity (100% velocity at VO_2max; vVO_2max), same recovery intensity (passive), and work-rest ratio (1:1), with the same total duration (21 min) for sixteen highly-trained males, participating in endurance (n = 8; VO_2max: 66.2 ± 5.0 ml/min/kg) or sprint (n = 8; VO_2max: 56.8 ± 5.0 ml/min/kg) events. Averaged over the entire training session, there was an equal relative heart rate, but higher relative oxygen uptake, higher respiratory quotient, and higher subjective exertion in the longer intervals.

In the literature, there are also approaches to optimizing short interval sessions in terms of time spent above 90% VO_2max. Cipryan et al. (13) recommend increasing the intensity within the intervals, increasing the intensity during the recoveries, or reducing the recovery duration to increase time spent near VO_2max when using short interval durations. Millet et al. (17) confirmed in a study with seven highly trained triathletes (VO_2max: 71.2 ± 4.2 ml/min/kg) the hypothesis that increasing intensity in short interval sessions increases the time spent above 90% VO_2max. For example, running an interval session with 105% vVO_2max (30 s work duration; 30 s recovery; 50% vVO_2max recovery intensity), the time above 90% VO_2max is about twice as high as running the intervals at 100% vVO_2max with the same interval duration, recovery duration, recovery intensity, and total training time (46 ± 20% vs. 23 ± 18% of total training time, or 338.1 ± 149.3 s vs. 167.7 ± 131.3 s; three sets of n intervals, with n × 30 s = t_lim; 731 ± 121 s total running time).

However, Thevenet et al. showed with nine endurance-trained male adolescents (VO_2max: 64.9 ± 4.2 ml/min/kg) that under the condition of performing training sessions to maximal perceived exertion, increasing intensity in 30 s intervals (100% maximal aerobic velocity vs. 110% maximal aerobic velocity; 50% vVO_2max recovery intensity) does not lead to a significant increase in time spent above 90% VO_2max (300.0 ± 150.3 s vs. 197.8 ± 148.1 s) or 95% VO_2max (117.8 ± 66.2 s vs. 62.2 ± 74.6 s) (14). Due to the higher anaerobic energy metabolism, the higher intensity results in fewer repetitions (11 vs. 24; 653 ± 187 vs. 1,440 ± 169 s) and therefore a 65% lower time above 90% VO_2max. The comparison with Cipryan et al. (13) illustrates that the active recovery design by Thevenet et al. (50% vVO_2max) and the expectedly slightly higher intensity lead to a significant increase in time spent near VO_2max in short interval sessions with 100% of the maximal aerobic velocity (14). An increase in time spent near VO_2max at higher recovery intensities in short interval sessions, across disciplines and performance levels, was also noted in the review by Buchheit and Laursen (6).

Based on these results, the present study investigated whether a short interval session with higher interval intensity and higher recovery intensity can achieve the same time spent above 90% VO_2max as a long interval session with lower intensity during both intervals and recovery periods. Recent studies (13) suggest that increasing intensity in interval and recovery during short intervals could be a method to increase time above 90% VO_2max. To date, however, there have been no studies in running that have intensified both factors (interval and recovery intensity). The present study addressed this novelty, hypothesising that by intensifying interval and recovery intensity, time above 90% VO_2max during shorts intervals can be achieved at an equally high or even higher level compared to long intervals.

The main finding of the present study was that intensified 30-s intervals were inferior to traditional 3-min intervals regarding the time spent above 90% VO_2max. An increase in intensity in the short intervals by 5% in both the interval and the recovery period did not result in a higher effect, despite the same total exercise time and work-rest ratio in both interval forms. These data are consistent with the early results of Astrand et al. in the 1960s (24), who described that long intervals (2–3 min) were the most effective form of interval training and that maximal intensities were not necessary to elicit VO_2max. Other studies support this preference for long intervals. The findings of Hill and Rowell (25) also align, showing that the minimum time to reach VO_2max is at 60% t_lim (running time to exhaustion at maximal aerobic speed). Therefore, some research groups suggest that a duration of 50%–60% t_lim (∼2–3 min) is ideal for effective VO_2max training. In competitive middle- and long-distance running, Cipryan et al. concluded that short intervals are cardiopulmonary less effective than long intervals (13). Both interval forms were performed at 100% vVO_2max with passive recovery. A study with eight highly trained triathletes showed that intervals (100% vVO_2max, recovery: 50% vVO_2max) of 30 s (recovery: 30 s) led to less time above 90% VO_2max than a work duration of 60 s (recovery: 30 s) or ½ t_lim (recovery: ½ t_lim∼118 s) with the same intensity (26). These results also show, in agreement with our data, a more positive effect with long intervals (>1 min).

The presented results already indicate that there are several factors influencing the interpretation of the “most effective VO_2max interval form”. Besides the length of the intervals, the intensity design is also an important parameter (6), both during the intervals and the recovery periods. Our results showed no positive effect of increasing intensity during interval and recovery by 5% in 30-s intervals, however compared to a 3-min session. Millet et al. suggest that a certain increase in interval intensity can be beneficial. The time above 90% VO_2max could be increased with 30-s intervals at 105% vVO_2max compared to 30-s intervals at 100% vVO_2max (with 30-s recovery periods at 50% vVO_2max) (17). However, a further increase to 110% vVO_2max (compared to 100% vVO_2max) did not lead to a significant difference in time above 90% VO_2max under an otherwise identical setting (14), indicating that fine-tuning of speeds seems necessary and faster is not always better.

Additionally, the design of the recovery period (duration, intensity) is another important variable. It determines for example how the VO₂ drop looks (VO₂ kinetics) or how lactate metabolism behaves. The 50% vVO_2max intensity during recovery in the 3-min intervals led to an average value of 65% VO_2max in our study, while in the 30-s intervals, the 55% vVO2max intensity during recovery corresponded to an average of 78% VO_2max, considering the time lag in VO₂ increase and decrease. During the 3-min intervals, a significant drop in VO₂ during the recovery period is evident (Figure 1), which is not the case with the 30-s intervals. On average, the interval and recovery VO₂ for the short interval runs are almost identical at approximately 78% VO_2max. Comparing 30-s intervals at 105% vVO_2max to exhaustion, Thevenet et al. (21) showed with eight endurance trained adolescents (VO_2max: 57.4 ± 6.1 ml/min/kg) that active (50% vVO_2max) and passive recovery periods did not result in a significant difference in time above 90% VO_2max. However, passive recovery periods required a significantly longer total training time (until t_lim) to achieve this. Dupont et al. (27) demonstrated with eleven male soccer players (VO_2max: 59.4 ± 4.2 ml/min/kg) with 15 s intervals (15 s at 120% vVO_2max, recovery: 15 s) that passive recovery periods enabled reaching VO_2max and extended t_lim. It should also be pointed out that the “off kinetics” in the recovery periods are not the primary focus of many studies and are therefore not adequately described and recorded in some cases. Currently, the question of the optimal recovery intensity remains unresolved. 50% vVO_2max is commonly used, as it is frequently mentioned in the literature and current practice among athletes, and it is assumed to facilitate effective lactate clearance. Studies indicate that the best lactate elimination occurs at intensities around 52%–63% VO_2max (28, 29). In the present study, the values for the 3-min session were in this range, while the short recovery periods with increased intensity in the 30-s session were significantly higher, suggesting that “optimal lactate elimination” might not be achieved. However, some of the data collected shows that recovery periods lead to a certain amount of lactate removal. In the 30-s session, the mean value after interval 12 was 5.61 ± 1.29 mmol/L. After interval 13 (i.e., a passive recovery period in between) the value was 5.43 ± 1.38 mmol/L. In the 3-min session, the value dropped from 8.43 ± 1.00 mmol/L after interval 2 to 7.86 ± 1.28 mmol/L during the 3-min recovery period (30 s passive recovery, 2:00 min active recovery).

There are also studies in cycling indicating that short intervals elicit better training adaptations than long intervals. Rønnestad et al. demonstrated this in highly trained cyclists (30). The study by Appelhans in cycling with 12 cyclists and triathletes (VO_2max: 68.0 ± 6.3 ml/min/kg) shows that the time spent near VO_2max can be similar in short interval sessions (3 × 13 × 30 s; recovery 15 s) and long interval sessions (6 × 5 min; recovery: 2:30 min) (31). The results of Rønnestad et al. and Appelhans could highlight the importance of a higher work-rest ratio while using short intervals. Otherwise, these studies raise the question of whether the response to short interval training depends on the specific sport.

In the overall context of VO_2max training, a “target time at or near VO_2max” is often referenced for athletes during a session. This refers to the time spent above 90% VO_2max in a single workout. Buchheit and Laursen (6) suggest a minimum of 7 min for team sports athletes and 10 min for long-distance runners or endurance athletes. Based on our results, it becomes clear that most athletes were unable to meet this criterion. On average, the time achieved was about 5.5 min during the 3-min intervals and just over 3 min during the 30-s intervals. However, it is important to note the wide individual range. Some athletes managed to exceed 8 min over 90% VO_2max in the 3-min intervals, and one athlete even completed nearly 15 min in the 30-s setting (Figure 2). Moreover, it is important to acknowledge that several other studies have also fallen short of the proposed minimum 7-10 min duration above 90% VO_2max. Millet et al. (17) achieved approximately 3 min and 6 min, respectively, in their 30-s interval setting (100% vVO_2max vs. 105% vVO_2max). Similarly, Thevenet et al. (14) reported 5 min and just over 3 min (100% vVO_2max vs. 110% vVO_2max). Only one study by Thevenet et al. (21) achieved durations above 90% VO_2max of around 12 min and 9 min in their comparison of active and passive recovery during 30-s intervals. However, their athletes ran until exhaustion and completed significantly longer total durations (almost 36 min on average during the passive recovery setting), which is not comparable to our setting (24 min total duration). Overall, it can be concluded that the achieved time above 90% VO_2max must also be considered in relation to the total volume of the session. Additionally, the performance level and the associated VO₂ kinetics of the athletes play a significant role (17). In further analyses, we found no significant correlations between time spent above 90% and 95% VO_2max and maximal performance diagnostic parameters (VO_2max, both relative and absolute; vVO_2max) or submaximal parameters (fixed lactate threshold of 3 mmol/L; %VO_2max at the fixed lactate threshold of 3 mmol/L). Thus, it can be concluded that these parameters cannot be used to infer the effectiveness of a program in terms of time spent above 90% or 95% VO_2max. It is likely that VO₂ kinetics and the physiological profile of the athletes (e.g., muscle fiber distribution, different neuromuscular profiles, etc.) play a more significant role (6). One of the highest durations above 90% VO_2max was reported in another study by Millet et al. (26). With 60–30 s intervals (100%–50% vVO_2max), they reached nearly 9 min on average in the “VO_2max zone” (26). Nevertheless, in the context of running, the minimum target of 10 min above 90% VO_2max per session seems to be an ambitious effectiveness criterion, which has rarely been met in studies and should be reconsidered in future research.

Another interesting finding of the present study was that the time above 90% HR_max showed an opposite trend to the time above 90% VO_2max. The time above 90% HR_max was significantly longer in the 30-s session than in the 3-min session (820 ± 249 s vs. 545 ± 131 s), despite the time above 90% VO_2max being shorter. This opposing trend was not observed in previous studies by Millet et al. (26). In the comparison of 30–30 s, 60–30 s, and 1/2 t_lim–t_lim (approximately 118 s), the settings with the longest time above 90% VO_2max also recorded the longest times above 90% HR_max. Additionally, in another study on increased intensity (30–30 s with 100% vs. 105% vVO_2max), longer durations above 90% VO_2max were accompanied by longer durations above 90% HR_max (17). It could be speculated that the more intense recovery period may have influenced the development of % HR_max. Nonetheless, the data indicate that utilization of HR_max does not automatically translate to utilization of VO_2max. This is relevant for the practical work of athletes and coaches, as HR is often used as a key factor in training control. Furthermore, it supports the observations of Laursen and Buchheit (6), who highlight the problems of HR control in high-intensity training. It might also be more appropriate to use the time above 95% HR_max as a measure for VO_2max load. Interestingly, the time above 95% HR_max for the 3-min intervals was, on average, longer than for the 30-s intervals (244 ± 109 s vs. 197 ± 213 s) and might be seen more as the “HR threshold” for high-intensity training in our study. This would align with newer zone models by Haugen et al. (32), which indicate the “HR threshold” for running in classical VO_2max training (Zone 5) at 93% HR_max.

In the present study, differences in blood lactate concentration were observed. With an average peak value of 9.7 mmol/L, the 3-min session was higher than the 30-s session, which had an average value of 7.6 mmol/L. There was also a high degree of individual variability, with maximum values of 12.4 and 10.3 mmol/L in the respective sessions. Internal calculations of the energy system contributions confirmed a slightly higher anaerobic component during the 3-min intervals, despite the lower intensity during the work phase. The lactate values in this study are generally consistent with findings from Cipryan et al. (13). Although no specific values were provided in their study, graphical data indicated peak values of 10–11 mmol/L for 3-min intervals at 100% vVO_2max, while the 30-s intervals (also at 100% vVO_2max) showed peak values slightly above 4 mmol/L. It is important to note that Cipryan's study used passive recovery periods, which must be considered when interpreting these results. In contrast, our study used active recovery periods, which is known to stimulate lactate removal and consequently impacts the measured lactate concentrations (33, 34). Other studies investigating 30-s interval sessions report similar lactate values, further corroborating our findings (14, 22). An interesting comparison can be made with the study by Rozenek et al. (35), which examined work-rest intervals of 15–15 s, 30–15 s, and 60–15 s at 100% vVO_2max (work) and 50% vVO_2max (recovery). In this study, the 15–15 s format produced the lowest lactate values (7.3 ± 2.4 mmol/L), while the 30–15 s (11.5 ± 1.8 mmol/L) and 60–15 s (12.5 ± 1.8 mmol/L) intervals did not differ significantly. This result highlights the impact of the work-rest ratio on peak lactate values, with a ratio of ≥2:1 increasing lactate concentrations even without raising exercise intensity. Moreover the lower lactate values of the 15–15 s format indicates a lower contribution of the anaerobic lactic energy system, what is similar to our study comparing short and long interval runs. Interestingly in short intervals these lower anaerobic lactic energy contribution is not compensated by a higher aerobic energy contribution. Despite 5% higher intensity, the VO₂ is lower within the short intervals, leading to a lower aerobic energy contribution. Therefore the lack of energy during the short interval runs would inevitably have to be compensated by the anaerobic alactic energy metabolism and its resynthesis during the recovery intervals (36, 37).

In addition to running velocities, HR, and blood lactate concentrations, perceived exertion is often used as a parameter for load monitoring in training settings. Laursen and Buchheit (6) mentioned a threshold of ≥6 on the CR-10 Borg scale for defining high-intensity training, which was also used in the present study. Both interval formats reached this threshold with average values of 6.8 and 6.5, respectively – yet, as previously noted, this did not result in an equally high VO_2max stimulus in terms of time spent above 90% VO_2max. However, it can be concluded that the intensity was not too low, and a subjectively high exertion was achieved. Direct comparisons to other studies are challenging, as most research groups use the 6–20 Borg scale, and converting between the two scales is not straightforward.

The results of the study offer several insights that are highly relevant from a practical sports perspective. They highlight that increasing the intensity of short VO_2max intervals may not always lead to the desired effect, particularly if achieving a high time spent above 90% VO_2max is the primary training goal. Thus traditional long intervals remains superior in terms of time spent above 90% VO_2max. The organization of intervals (duration, intensity, work-rest ratio, etc.) is crucial in determining the training effects and should be carefully considered by coaches.

The time above 90% HR_max shows an opposite trend to the time above 90% VO_2max and is higher in short intervals. Moreover the perceived exertion do not differ between short and long intervals despite higher time spent above 90% VO_2max in long interval runs. Coaches should be careful not to misinterpret HR and the perceived exertion. In practice, when HR is used as a training monitoring marker, the findings suggest that aiming for HR values exceeding 93%–95% of HR_max is essential for effective VO_2max training, particularly for maintaining time above 90% VO_2max. The blood lactate values measured during training should always be interpreted in relation to the design of the specific interval session.