Endurance exercise training studies frequently show modest changes in VO2max with training and very limited responses in some subjects. By contrast, studies using interval training (IT) or combined IT and continuous training (CT) have reported mean increases in VO2max of up to ~1.0 L · min−1. This raises questions about the role of exercise intensity and the trainability of VO2max. To address this topic we analyzed IT and IT/CT studies published in English from 1965–2012. Inclusion criteria were: 1)≥3 healthy sedentary/recreationally active humans <45 yrs old, 2) training duration 6–13 weeks, 3) ≥3 days/week, 4) ≥10 minutes of high intensity work, 5) ≥1:1 work/rest ratio, and 6) results reported as mean ± SD or SE, ranges of change, or individual data. Due to heterogeneity (I2 value of 70), statistical synthesis of the data used a random effects model. The summary statistic of interest was the change in VO2max. A total of 334 subjects (120 women) from 37 studies were identified. Participants were grouped into 40 distinct training groups, so the unit of analysis was 40 rather than 37. An increase in VO2max of 0.51 L ·min−1 (95% CI: 0.43 to 0.60 L · min−1) was observed. A subset of 9 studies, with 72 subjects, that featured longer intervals showed even larger (~0.8–0.9 L · min−1) changes in VO2max with evidence of a marked response in all subjects. These results suggest that ideas about trainability and VO2max should be further evaluated with standardized IT or IT/CT training programs.
Citation: Bacon AP, Carter RE, Ogle EA, Joyner MJ (2013) VO2max Trainability and High Intensity Interval Training in Humans: A Meta-Analysis. PLoS ONE 8(9): e73182. doi:10.1371/journal.pone.0073182
Editor: Conrad P. Earnest, University of Bath, United Kingdom
Received: April 15, 2013; Accepted: July 18, 2013; Published: September 16, 2013
Copyright: © 2013 Bacon et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This publication was partially supported by Grant Number UL1 TR000135 from the National Center for Advancing Translational Sciences (NCATS), the Mayo Foundation, and Caywood Professorship to MJJ. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The benefits of an active lifestyle are well documented –. Many of these benefits are also associated with higher levels of cardiorespiratory fitness (VO2max) which may exert protective effects that are independent of traditional risk factors , . Additionally, for individuals with low physical fitness, even modest improvements in fitness can have substantial health benefits. However, some individuals may have a limited ability to increase their cardiorespiratory fitness (trainability) in response to endurance exercise training , .
A key study advancing the idea that some humans have limited trainability comes from Bouchard et al. studied 483 sedentary white adults from 99 nuclear families who completed a standardized 20-wk endurance training program , . The subjects trained three times per week on a treadmill. Initially, they trained at a heart rate that correlated to 55% of their baseline VO2max for 30 minutes per session. Every two weeks the intensity and duration of the exercise was progressively increased until each subject was training for 50 minutes at a heart rate associated with 75% of their baseline VO2max. This level of intensity and duration was reached by the 14th week of training and maintained until the conclusion of the study. Using this approach, they found a mean increase in VO2max of ~0.4 L · min−1 with a SD of >0.2 L/min. Additionally, 7% of subjects showed a gain of 0.1 L · min−1 or less while 8% of subjects improved by 0.7 L · min−1 or more.
Based on this distribution of VO2max responses it appears that the “trainability” of at least some subjects is low or non-existent with little or no improvement in cardiorespiratory fitness in spite of 20 weeks of structured exercise training , . These observations are in contrast to reports from smaller studies that have used either interval training (IT), or interval training in combination with continuous training (CT) and shown more robust increases in VO2max with at least some evidence of marked responses in all subjects –.
In this context, we sought to explore the hypothesis that all subjects can show marked improvements in VO2max if training programs that include periods of high intensity (~90% of VO2max) exercise are used. A fundamental rationale underpinning our analysis is that the biology of VO2max trainability needs to be evaluated using regimens designed to maximize physiological adaptations. To test this hypothesis, we evaluated the changes in VO2max in response to interval training (IT) or combined IT and continuous training (CT) reported in 37 studies –. We also sought to gain insight into the idea that shorter periods of IT might be either superior or more time efficient at generating increases in VO2max in comparison to traditional continuous training.
Study Eligibility Criteria and Literature Search
A search of PubMed was developed and conducted by a reference librarian looking for English language studies occurring between the years of 1968–2010 using the following search terms: interval training, VO2max, interval exercise + high/low intensity, maximum O2, maximum/maximal VO2, maximal oxygen consumption, peak oxygen uptake, maximal aerobic capacity, and the following limits: humans, English. The primary author of each eligible study was searched for in PubMed using the authors name and the search term: exercise. Searches were also run on other authors known by the investigators to have published papers that used a combination of interval and continuous training. The references of all eligible studies were also reviewed to identify other potentially eligible studies that might have been missed using the approaches outlined above. Subsequently, these studies were searched for in PubMed using the specific authors and article reference titles. Only published material was used. Figure 1 is a schematic of our search strategy.
In order to be considered eligible for inclusion, studies had to meet the following criteria:
- Studies had to include a minimum of 3 subjects.
- All subjects were previously healthy and between 18–45 years of age.
- The initial fitness level of the subjects had to fall within the range of untrained/sedentary to recreationally active, defined as 55 ml · kg−1 · min−1 or lower for men and 49.5 ml · kg−1 · min−1 or lower for women. Studies were included with recreationally trained individuals if they were in the minority and if individual data was reported, allowing us to recalculate the pre and post training VO2max values and their associated SDs without the trained individuals. When VO2max was not provided in ml · kg−1 · min−1 a group estimate was calculated using the mean weight and pre VO2max in L · min−1.
- Subjects had to train a minimum of 3 days per week for 6–13 weeks. At least two, and preferably three, training sessions had to be interval training.
- Subjects had to be trained on either a cycle ergometer or by running with a minimum work: relief ratio of 1 min on:1 min off.
- During each interval training session subjects had to have a minimum cumulative time of 10 minutes at high intensity. Due to the heterogeneity of training programs in the various studies we felt it was critical for the “work” period of the exercise sessions to be at intensities that were likely at least 80–85% of VO2max or higher. In some studies the desired intensities were set and monitored rigorously via power output or heart rate and in some cases the authors had to make a reasoned estimate. This was especially the case with some of the running training studies that used timed runs (for example reference 23).
- Results had to be reported in pre and post mean VO2max values ± SD or SE, as ranges of change, or individual data.
- Subjects could not be engaged in high intensity training before the start of the study.
Two independent investigators (M.J.J and A.B.) screened the titles and abstracts of all studies identified by the various search methodologies to determine their potential eligibility. Studies that did not have an abstract, along with studies that were deemed potentially eligible, had their full text reviewed in order to determine if they met the criteria for inclusion in the meta-analysis. In order to avoid double counting of studies that reported different aspects of the physiological responses to training, the study design (training methodology, number of subjects, study duration) and subject characteristics (age, pre and post VO2max) of all accepted studies were compared. When studies with common authors, study design, and subject characteristics were found, the study with the larger number of subjects was used. This was done because it appeared that some training studies generated multiple publications with, for example, cardiorespiratory data the focus of one paper, and the metabolic adaptations to training the focus of a second paper.
Data Extraction and Synthesis
Data extraction was performed by one investigator (A.B.) who used a standardized form that allowed for the extraction of study characteristics (author, title, number of subjects, study duration, age range, work: relief ratio, frequency of training, mean pre and post training VO2max values in L · min−1 and in ml · kg−1 · min−1 along with the associated standard errors of the mean (SEM) and standard deviations (SD)).
When needed, pre and post VO2max values, along with pre and post SD values, were converted into L • min−1 to limit the potentially confounding effects of changes in body weight that might have occurred with training. Pre and post SEM values were converted into SD. The calculated values were included in the data extraction sheet. For several studies mean, SD, and pre post correlation were recalculated from individual data. In the majority of studies, only the pre and post measurements were reported (i.e., the standard deviation of the change or the correlation of the pre and post measurements were not reported). Accordingly, the correlation was conservatively set at 0.5 for these cases. When studies reported treadmill and cycle VO2max values, the treadmill values were used.
The principle summary measure utilized was group difference in pre and post training means for VO2max expressed in L • min−1, calculated with a 95% confidence interval using a random effects model . As a sensitivity analysis to the random effects model, a fixed effects model was also fit to the data to estimate the degree of heterogeneity, indicated by I2. We used L • min−1 to avoid the potential confounding effect of significant weight loss and to better compare our analysis with the data from Bouchard et al. . Mean difference was employed as it allowed us to calculate the effect size, which was then compared to the effect size of Bouchard et al. . A standardized change in VO2max (mean change for each study divided by the standard deviation of change) was also analyzed using the same random effects modeling framework.
In order to evaluate the possibility of publication bias, a funnel plot of the mean differences was constructed and assessed. The funnel plot (Figure 2) indicated the potential for publication bias (e.g., smaller studies had larger effect sizes reported), so an imputed model was considered with fictional studies representing the complementary effect sizes to observed studies. A ‘Trim and Fill’ analysis was conducted to further assess for publication bias by providing an estimated pooled effect with the imputed studies while ‘trimming’ (removing) the most extreme small studies . Comprehensive Meta Analysis Software (Biostat) version 2.2.064 was used for all analyses.
The open circles are observed studies whereas the solid circles are imputed studies to assess the role of publication on the estimated effect. The two estimates, denoted by the open and filled diamonds for the observed and imputed studies, respectively, agree qualitatively with one another suggesting that the findings may be robust in the presence of publication bias.
A total of 211 articles were identified in the initial search. After screening titles and abstracts, 181 articles were discarded as it was apparent that they did not meet the study criteria. Figure 1 is a schematic of our search strategy and displays the criteria for exclusion. The full text of the remaining 30 articles was examined in detail. An additional 70 articles were identified through reference searches. The systematic review resulted in 37 articles for the Meta-analysis. Within these 37 articles were 334 participants (120 women) and 40 unique training groups (i.e., three articles each contained the results of two training programs). In brief, all studies contained between 3 to 19 healthy subjects with an age range of 18 to 42 years and pre training VO2max values ranging from ~26 to 52 ml · kg−1 · min−1.
All subjects trained between 6 and 12 weeks using running, cycling, or a combination of the two modalities with a work relief ratio between 1:1 and 5:2. In all but six studies , , , , ,  it was readily apparent that the subjects had trained at high intensity for 10 minutes or more. Knuttgen and Overend reported their work: relief ratios and the duration of their interval training, which allowed us to estimate the number of intervals performed and the time spent performing high intensity exercise which was determined to be greater than 10 minutes , . Slordahl reported the duration of their interval training but reported their work: relief ratios as a range . This allowed us to roughly estimate that the subjects performed at least 10 minutes of work at high intensity. Burke reported that their subjects trained to exhaustion; therefore it was assumed that they performed at least 10 minutes of work at high intensity [17. Warburton reported that their subjects performed interval training until the amount of work performed equaled the amount of work that would have been performed by the subject if they had been performing continuous training . Since the subjects in the continuous training group worked for 30–48 min/day at 64.3±3.7% of their VO2max we assumed that the subjects performing interval training worked for at least 10 minutes at high intensity. Finally, Roca reported that their subjects trained less than 10 minutes at high intensity on one day of their high intensity training. Additionally, this was the only study in which interval training did not account for at least half of all training days for the test subjects. However, this exception was justified because the continuous component was judged to be strenuous enough to make up for the diminished interval component .
Physiological Responses: Meta-Analysis Results
As a result of the training programs, VO2max increased (p<0.001). The estimated mean change in VO2max over the training programs using the random effects model was 0.51 L · min−1 (95% CI: 0.43 to 0.60 L · min−1) (Figure 2). In terms of standardized effect sizes (estimate divided by standard deviation), the change in VO2max was 0.86 SDs (95% CI: 0.72 to 0.99 SDs).
There was some indication of publication bias in the funnel plot (Figure 2). Specifically, the estimated effect using the ‘Trim and Fill’ approach was 0.37 L · min−1 (95% CI: 0.28 to 0.46 L · min−1), but the number of negative studies required to overturn the observed finding was estimated to be over 4000 (Rosenthal's fail-safe N = 4531) . Thus, while Figure 2 shows a slight attenuation of the result with hypothetical studies imputed to account for potential publication bias, the results are not meaningfully changed. Finally, a sensitivity analysis was also conducted under a fixed effects model. There was significant heterogeneity in the exercise groups (I2 = 70), but the results of this analysis agreed qualitatively with the primary random effects model. Specifically, under the fixed effects model, the estimated change was 0.45 L · min−1 (95% CI: 0.41 to 0.49 L · min−1; p<0.001).
To better illustrate the distribution of effects observed through the systematic review and demonstrated in Figure 3, a weighted histogram and density plot of the observed effects was created (Figure 4). Each of the 40 observations included in the analysis was weighted by the sample size to produce the distribution.
The estimates reported for each study are the means and 95% confidence intervals for the change in VO2max in L · min−1. The summary diamonds at the bottom of the plot represent the summarized effects using fixed and random effects models, where the random effects estimates are considered the primary findings for this study due to heterogeneity (I2 = 70). The estimated increase in VO2max as a result of interval training was 0.51 L · min−1 (95% CI: 0.43 to 0.60 L · min−1; p<0.001).
Each of the 40 observations included in the analysis was weighted by the sample size to produce the distribution with the estimate of the percent of subjects with a given effect size noted on the Y axis.
In a supplemental analysis, we also evaluated the nine studies that reported the largest – and nine studies that reported the smallest , , , , , , – mean changes in VO2max. These studies are the ones that would be located in the tails of the empirical distribution. Specifically, the mean ± SD increase in VO2max was 0.27±0.05 vs 0.87±0.15 L · min−1 (p<0.001) in this comparison. Weeks of training (6.9±1.4 vs 9.7±1.8; p = 0.004), duration per week (123±67 vs 209±90 min; p = 0.06), and the duration of the intervals for the entire study (479±246 vs 696±264 min; p = 0.12) showed general patterns that suggested the nine studies showing the smallest increases in VO2max were of shorter duration in terms of both weekly training and number of weeks. They also appear to have been less intense (fewer total high intensity minutes) than the nine studies showing the largest increases, although statistical significance was not achieved for all comparisons. Finally, many of the interval training studies reporting large increases in VO2max also used longer (3–5 min) duration intervals.
The main finding of this meta-analysis is that interval training produces improvements in VO2max slightly greater than those typically reported with what might be described as adult fitness based continuous training even though many of the studies were of short duration with limited training sessions per week. While the observation that more intense training results in greater increases in cardiorespiratory fitness is not surprising, our analysis suggests that longer intervals combined with high intensity continuous training can generate marked increases in VO2max in almost all relatively young adults.
As noted in the introduction, the Heritage Study demonstrated a mean increase in VO2max of 0.4 L · min−1 in response to 20 weeks of standardized adult fitness style training and marked individual variation was observed . This value is also relatively high in comparison to other large adult fitness style training programs that show increases in VO2max on the order of 0.2–0.3 L • min−1 . By contrast, the studies we evaluated that used interval training alone or in combination with continuous training estimated a mean increase in VO2max of 0.5 L · min−1. Figure 4 demonstrates an estimated distribution of training evoked changes in VO2max for the studies we evaluated. This figure presents a distribution of means whereas the Heritage data is of population values. It is worth noting, however, that the distribution of study effects presented in Figure 4 would appear to be shifted right in comparison to the Heritage data in spite of the fact that some of the studies were of limited duration (6–8 weeks); used intervals that were only one minute long and the time spent during the “work” portion of the interval training session was as low as 10 minutes. Thus, assuming the individual subject data used to generate the summary statistics in this meta-analysis follow roughly a normal distribution, the distribution of individual changes would be comparable, if not slightly better, than the Heritage data even though the overall duration (weeks spent training) were shorter in the studies we evaluated.
The conventional wisdom is that intervals of 3–5 minutes are especially effective in evoking increases in exercise capacity. Consistent with this idea, the nine studies that generated the biggest increases in VO2max (~0.85/min) generally used intervals of 3–5 minutes and high intensity continuous training –. Additionally, many of these studies presented either individual data or ranges for VO2max values pre and post training, and inspection of this data suggests that a marked training response was seen in all subjects. Thus, given the large increases in VO2max seen in these studies, we believe that caution must be used in concluding that at least some humans are incapable (perhaps for genetic reasons) of increasing their VO2max in response to endurance exercise training.
Along these lines, many of the studies showing the largest increases in VO2max appeared to follow a pattern similar to the so-called “Hickson protocol” . This protocol includes 10 weeks of training 6 days/week with interval and continuous training on alternate days. Interval training consists of six 5 minute sessions on a cycle ergometer at a work rate approaching the subjects' VO2max. These work periods are separated by 2 minutes of active rest. As the subjects' power output increases during training, the exercise intensity is increased as needed. On the non-interval days, continuous training consists of running as fast as possible for 30 min/day during the 1st week, 35 min/day during the second week, and 40 min/day or longer thereafter.
The original study of Hickson and colleagues noted that VO2max was continuing to increase at the end of the 10 week training program . However, their subjects declined to continue to train due to the extremely arduous nature of the program. This observation emphasizes that it may be unrealistic to expect significant segments of the population to participate in an exercise training program that permits them to reach their individual upper limit for VO2max. It should also be noted that the high intensity continuous running almost certainly contributed to the large increases in VO2max seen in studies using the Hickson protocol but it is not possible to determine the relative role of each type of training.
Since there is no genetic data on the subjects from the studies included in our analysis it is not possible to use them to retrospectively evaluate the genetic determinants of trainability that have emerged from Heritage . It is well accepted that the increases in VO2max with training are due to increases in cardiac output and peripheral oxygen extraction. However, the contribution of changes in stroke volume, blood volume, capillary density, muscle mitochondrial content and many other factors associated with training induced increases in VO2max might vary on both an individual basis and perhaps via interaction with specific elements of a given training program –.
In addition to the well-known limitations associated with retrospective analysis of data reported previously, there are specific limitations to our analysis. First, as detailed above there was wide variability in training programs used in the studies we evaluated. Second, about two thirds of the subjects included in the studies we evaluated were (likely Caucasian) young men and this could explain at least some of the increased responsiveness that we saw. However, sex and age generally do not have major effects on the VO2max responses to training , . We also did not include studies in patient populations that have shown marked improvements in exercise capacity with interval training in groups previously thought to have limited trainability . Nor did we include a body of work showing that very short bouts of high intensity exercise followed by longer periods of recovery can have profound effects on exercise capacity . The findings from these studies along with the data reported in our paper indicate that some of the peripheral metabolic adaptations to training require can be elicited by very short periods of high intensity exercise, but that longer intervals are required to see large changes in cardiac output and VO2max. The analysis also suggested some effect of publication bias, particularly with studies with small sample sizes and small effects. While the estimated change in VO2max remained statistically significant with various sensitivity analyses, a clear causal association is inherently limited with this data.
In summary, our analysis indicates that in addition to studies using training programs consistent with various public health guidelines, the basic biology of trainability needs to be evaluated using regimens designed to generate the most robust possible physiological adaptations.
We would like to thank Ann Farrell from the Mayo Library for her help in the literature search. Additionally, Janet Beckman assisted with manuscript preparation and Jennifer Taylor helped with the figures. Their help is greatly appreciated.
Conceived and designed the experiments: MJJ. Performed the experiments: APB REC EAO MJJ. Analyzed the data: APB REC EAO MJJ. Contributed reagents/materials/analysis tools: APB REC EAO MJJ. Wrote the paper: APB REC MJJ.
- 1. Blair SN, Morris JN (2009) Healthy hearts–and the universal benefits of being physically active: physical activity and health. Ann Epidemiol 19: 253–256. doi: 10.1016/j.annepidem.2009.01.019
- 2. Moore SC, Patel AV, Matthews CE, Berrington de Gonzalez A, Park Y, et al. (2012) Leisure time physical activity of moderate to vigorous intensity and mortality: a large pooled cohort analysis. PLoS Med 9: e1001335. doi: 10.1371/journal.pmed.1001335
- 3. Joyner MJ, Green DJ (2009) Exercise protects the cardiovascular system: effects beyond traditional risk factors. J Physiol 587: 5551–5558. doi: 10.1113/jphysiol.2009.179432
- 4. Lee DC, Artero EG, Sui X, Blair SN (2010) Mortality trends in the general population: the importance of cardiorespiratory fitness. J Psychopharmacol 24: 27–35. doi: 10.1177/1359786810382057
- 5. Bouchard C, Sarzynski MA, Rice TK, Kraus WE, Church TS, et al. (2011) Genomic predictors of the maximal O2 uptake response to standardized exercise training programs. J Appl Physiol 110: 1160–1170. doi: 10.1152/japplphysiol.00973.2010
- 6. Skinner JS, Jaskolski A, Jaskolska A, Krasnoff J, Gagnon J, et al. (2001) HERITAGE Family Study. Age, sex, race, initial fitness, and response to training: the HERITAGE Family Study. J Appl Physiol 90: 1770–1776.
- 7. Hickson RC, Bomze HA, Holloszy JO (1997) Linear increase in aerobic power induced by a strenuous program of endurance exercise. J Appl Physiol: Respirat Environ Exercise Physiol 42: 372–376.
- 8. Coggan AR, Habash DL, Mendenhall LA, Swanson SC, Kien CL (1993) Isotopic estimation of CO2 production during exercise before and after endurance training. J Appl Physiol 75: 70–75. doi: 10.1249/00005768-199205001-00844
- 9. Cox ML, Bennett JB III, Dudley GA (1986) Exercise training-induced alterations of cardiac morphology. J Appl Physiol 61: 926–931.
- 10. Roca J, Agusti AGN, Alonso A, Poole DC, Viegas C, et al. (1992) Effects of training on muscle O2 transport on VO2max. J Appl Physiol 73: 1067–1076.
- 11. Hurley BF, Nemeth PM, Martin WH III, Hagberg JM, Dalsky GP, et al. (1986) Muscle triglyceride utilization during exercise: effect of training. J Appl Physiol 60: 562–567.
- 12. Proctor DN, Miller JD, Dietz NM, Minson CT, Joyner MJ (2001) Reduced submaximal leg blood flow after high-intensity aerobic training. J Appl Physiol 91: 2619–2627.
- 13. Winder WW, Hickson RC, Hagberg JM, Ehsani AA, McLane JA (1979) Training-induced changes in hormonal and metabolic responses to submaximal exercise. J Appl Physiol: Respirat Environ Exercise Physiol 46: 766–771.
- 14. Warburton DER, Haykowsky MJ, Quinney HA, Blackmore D, Teo KK, et al. (2004) Blood volume expansion and cardiorespiratory function: effects of training modality. Med Sci Sports Exerc 36: 991–1000. doi: 10.1249/01.mss.0000128163.88298.cb
- 15. Knuttgen HG, Mordesjo L-O, Ollander B, Saltin B (1973) Physical conditioning through interval training with young male adults. Med Sci Sports 5: 220–226. doi: 10.1249/00005768-197300540-00002
- 16. Berger NJ, Tolfrey K, Williams A, Jones A (2006) Influence of continuous and interval training on oxygen uptake on-kinetics. Med Sci Sports Exerc 38: 504–512. doi: 10.1249/01.mss.0000191418.37709.81
- 17. Burke J, Thayer R, Belcamino M (1994) Comparison of effects of two interval-training programmes on lactate and ventilatory thresholds. Br J Sports Med 28: 18–21. doi: 10.1136/bjsm.28.1.18
- 18. Coggan AR, Spina RJ, Kohrt WM, Bier DM, Holloszy JO (1990) Endurance training decreases plasma glucose turnover and oxidation during moderate-intensity exercise in men. J Appl Physiol 68: 990–996.
- 19. Coggan AR, Spina RJ, Kohrt WM, Holloszy JO (1993) Effect of prolonged exercise on muscle citrate concentration before and after endurance training in men. Am J Physiol (Endocrinol Metab 27) 264: E215–E220.
- 20. Cunningham DA, McCrimmon D, Vlach LF (1979) Cardiovascular response to interval and continuous training in women. Eur J Appl Physiol 41: 187–197. doi: 10.1007/bf00430011
- 21. Duffield R, Edge J, Bishop D (2006) Effects of high-intensity interval training on the VO2 response during severe exercise. J Sci Med Sport 9: 249–255. doi: 10.1016/j.jsams.2006.03.014
- 22. Edge J, Bishop D, Goodman C (2006) Effects of chronic NaHCO3 ingestion during interval training on changes to muscle buffer capacity, metabolism, and short-term endurance performance. J Appl Physiol 101: 918–925. doi: 10.1152/japplphysiol.01534.2005
- 23. Fox EL, Bartels RL, Klinzing J, Ragg K (1977) Metabolic responses to interval training programs of high and low power output. Med Sci Sports 9: 191–196. doi: 10.1249/00005768-197621000-00107
- 24. Gaesser GA, Wilson LA (1988) Effects of continuous and interval training on the parameters of the power-endurance time relationship for high-intensity exercise. Int J Sports Med 9: 417–421. doi: 10.1055/s-2007-1025043
- 25. Gurd BJ, Perry CGR, Heigenhauser GJF, Spriet LL, Bonen A (2010) High-intensity interval training increases SIRT1 activity in human skeletal muscle. Appl Physiol 35: 350–357. doi: 10.1139/h10-030
- 26. Gyntelberg F, Rennie MJ, Hickson RC, Holloszy JO (1977) Effect of training on the response of plasma glucagon to exercise. J Appl Physiol: Respirat Environ Exercise Physiol 43: 302–305.
- 27. Hickson RC (1980) Interference of strength development by simultaneously training for strength and endurance. Eur J Appl Physiol 45: 255–263. doi: 10.1007/bf00421333
- 28. Hickson RC, Foster C, Pollock ML, Galassi TM, Rich S (1985) Reduced training intensities and loss of aerobic power, endurance, and cardiac growth. J Appl Physiol 58: 492–499.
- 29. Hickson RC, Hagberg JM, Ehsani AA, Holloszy JO (1981) Time course of the adaptive responses of aerobic power and heart rate to training. Med Sci Sports Exerc 13: 17–20. doi: 10.1249/00005768-198101000-00012
- 30. Hickson RC, Kanakis C Jr, Davis JR, Moore AM, Rich S (1982) Reduced training duration effects on aerobic power, endurance, and cardiac growth. J Appl Physiol: Respirat Environ Exercise Physiol 53: 225–229.
- 31. Hickson RC, Rosenkoetter MA (1981) Reduced training frequencies and maintenance of increased aerobic power. Med Sci Sports Exer 13: 13–16. doi: 10.1249/00005768-198101000-00011
- 32. Overend TJ, Paterson DH, Cunningham DA (1992) The effect of interval and continuous training on the aerobic parameters. Can J Sport Sci 17: 129–134.
- 33. Perry CGR, Heigenhauser GJF, Bonen A, Spriet LL (2008) High-intensity aerobic interval training increases fat and carbohydrate metabolic capacities in human skeletal muscle. Appl Physiol Nutr Metab 33: 1112–1123. doi: 10.1139/h08-097
- 34. Poole DC, Gaesser GA (1985) Response of ventilatory and lactate thresholds to continuous and interval training. J Appl Physiol 58: 1115–1121.
- 35. Poole DC, Ward SA, Whipp BJ (1990) The effects of training on the metabolic and respiratory profile of high-intensity cycle ergometer exercise. Eur J Appl Physiol 59: 421–429. doi: 10.1007/bf02388623
- 36. Ray CA, Cureton KJ (1991) Interactive effects of body posture and exercise training on maximal oxygen uptake. J Appl Physiol 71: 596–600.
- 37. Siegel W, Blomqvist G, Mitchell JH (1970) Effects of a quantitated physical training program on middle-aged sedentary men. Circulation 41: 19–29. doi: 10.1161/01.cir.41.1.19
- 38. Slordahl SA, Madslien VOE, Stoylen A, Kjos A, Helgerud J, et al. (2004) Atrioventricular plane displacement in untrained and trained females. Med Sci Sports Exerc 36: 1871–1875. doi: 10.1249/01.mss.0000145444.01292.3d
- 39. Smith AE, Walter AA, Graef JL, Kendall KL, Moon JR, et al. (2009) Effects of β-alanine supplementation and high-intensity interval training on endurance performance and body composition in men; a double-blind trial. J Int Soc Sports Nutr 6: 5. doi: 10.1186/1550-2783-6-5
- 40. Spina RJ, Ogawa T, Coggan AR, Holloszy JO, Ehsani AA (1992) Exercise training improves left ventricular contractile response to β-adrenergic agonist. J Appl Physiol 72: 307–311.
- 41. Walter AA, Smith AE, Kendall KL, Stout JR, Cramer JT (2010) Six weeks of high-intensity interval training with and without β-alanine supplementation for improving cardiovascular fitness in women. J Strength Cond Res 24: 1199–1207. doi: 10.1519/jsc.0b013e3181d82f8b
- 42. Weltman A, Weltman JY, Womack CJ, Davis SE, Blumer JL, et al. (1997) Exercise training decreases the growth hormone (GH) response to acute constant-load exercise. Med Sci Sports Exerc 29: 669–676. doi: 10.1097/00005768-199705000-00013
- 43. Womack CJ, Davis SE, Blumer JL, Barrett E, Weltman AL, et al. (1995) Slow component of O2 uptake during heavy exercise: adaptation to endurance training. J Appl Physiol 79: 838–845.
- 44. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7: 177–88. doi: 10.1016/0197-2456(86)90046-2
- 45. Duval S, Tweedie R (2000) Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics 56: 455–463. doi: 10.1111/j.0006-341x.2000.00455.x
- 46. Rosenthal R (1979) The file drawer problem and tolerance for null results. Psychological Bull 86: 638–641. doi: 10.1037/0033-2909.86.3.638
- 47. Bouchard C, Blair SN, Church TS, Earnest CP, Hagberg JM, et al. (2012) Adverse metabolic response to regular exercise: is it a rare or common occurrence? PLoS One 7: e3787. doi: 10.1371/journal.pone.0037887
- 48. Clausen JP (1977) Effect of physical training on cardiovascular adjustments to exercise in man. Physiol Rev 57: 779–815.
- 49. Holloszy JO, Coyle EF (1984) Adaptations of skeletal muscle to endurance exercise and their metabolic consequences. J Appl Physiol 56: 831–838.
- 50. Laughlin MH, Roseguini B (2008) Mechanisms for exercise training-induced increases in skeletal muscle blood flow capacity: differences with interval sprint training versus aerobic endurance training. J Physiol Pharmacol 59 (suppl 7)71–88.
- 51. Levine BD (2008) VO2max: what do we know, and what do we still need to know? J Physiol 586: 25–34. doi: 10.1113/jphysiol.2007.147629
- 52. Laursen PB, Jenkins DG (2002) The scientific basis for High-intensity interval training. Sports Med 32: 53–73. doi: 10.2165/00007256-200232010-00003
- 53. Kemi OJ, Wisloff U (2010) High-intensity aerobic exercise training improves the heart in health and disease. J Cardiopulm Rehabil Prev 30: 2–11. doi: 10.1097/hcr.0b013e3181c56b89
- 54. Gibala MJ, Little JP, MacDonald MJ, Hawley JA (2012) Physiological adaptations to low-volume, high-intensity interval training in health and disease. J Physiol 590: 1077–1084. doi: 10.1113/jphysiol.2011.224725