Article Text

Is interval training the magic bullet for fat loss? A systematic review and meta-analysis comparing moderate-intensity continuous training with high-intensity interval training (HIIT)
  1. Ricardo Borges Viana1,
  2. João Pedro Araújo Naves1,
  3. Victor Silveira Coswig2,
  4. Claudio Andre Barbosa de Lira1,
  5. James Steele3,
  6. James Peter Fisher3,
  7. Paulo Gentil1
  1. 1 Faculty of Physical Education and Dance, Federal University of Goiás, Goiânia, Goiás, Brazil
  2. 2 Faculty of Physical Education, Federal University of Pará, Castanhal, Pará, Brazil
  3. 3 Centre for Health, Exercise, and Sport Science, School of Sport, Health and Social Sciences, Southampton, Hampshire, UK
  1. Correspondence to Dr Paulo Gentil, Faculdade de Educação Física e Dança, Universidade Federal de Goiás, Goiania 74605-220, Brazil; paulogentil{at}


Objectives To compare the effects of interval training and moderate-intensity continuous training (MOD) on body adiposity in humans, and to perform subgroup analyses that consider the type and duration of interval training in different groups.

Design Systematic review and meta-analysis.

Data sources English-language, Spanish-language and Portuguese-language searches of the electronic databases PubMed and Scopus were conducted from inception to 11 December 2017.

Eligibility criteria for selecting studies Studies that met the following criteria were included: (1) original articles, (2) human trials, (3) minimum exercise training duration of 4 weeks, and (4) directly or indirectly compared interval training with MOD as the primary or secondary aim.

Results Of the 786 studies found, 41 and 36 were included in the qualitative analysis and meta-analysis, respectively. Within-group analyses showed significant reductions in total body fat percentage (%) (interval training: −1.50 [95% CI −2.14 to −0.86, p<0.00001] and MOD: −1.44 [95% CI −2.00 to −0.89, p<0.00001]) and in total absolute fat mass (kg) (interval training: −1.58 [95% CI −2.74 to −0.43, p=0.007] and MOD: −1.13 [95% CI −2.18 to −0.08, p=0.04]), with no significant differences between interval training and MOD for total body fat percentage reduction (−0.23 [95% CI −1.43 to 0.97], p=0.705). However, there was a significant difference between the groups in total absolute fat mass (kg) reduction (−2.28 [95% CI −4.00 to −0.56], p=0.0094). Subgroup analyses comparing sprint interval training (SIT) with MOD protocols favour SIT for loss of total absolute fat mass (kg) (−3.22 [95% CI −5.71 to −0.73], p=0.01). Supervised training, walking/running/jogging, age (<30 years), study quality and intervention duration (<12 weeks) favourably influence the decreases in total absolute fat mass (kg) observed from interval training programmes; however, no significant effect was found on total body fat percentage (%). No effect of sex or body mass index was observed on total absolute fat mass (kg) or total body fat percentage (%).

Conclusion Interval training and MOD both reduce body fat percentage (%). Interval training provided 28.5% greater reductions in total absolute fat mass (kg) than MOD.

Trial registration number CRD42018089427.

  • exercise
  • fat percentage
  • meta-analysis
  • sports and exercise medicine
View Full Text

Statistics from


Whether physical activity affects weight control has been an ongoing topic of controversy.1 2 The majority of physical activity guidelines for the management of obesity recommend high exercise volumes.3 4 Guidelines generally recommend 150–250 min/week, and up to 60 min/day, of moderate-intensity aerobic exercise to prevent weight gain or to reduce body mass a little bit (2–3 kg).4 5 More than an hour of exercise daily (>420 min/week) is recommended to lose more weight (5–7.5 kg)3 and few people meet these guidelines.6 7

Interval training may have the potential to promote weight loss as it has some benefits similar to moderate-intensity continuous training (MOD) while requiring less time.8 9 MOD is typically defined as continuous effort that elicits 55%–70% of the maximal heart rate (HRmax) or promotes oxygen consumption (Embedded Image O2) equivalent to 40%–60% of the maximum Embedded Image O2 (Embedded Image O2max).10 Interval training is an intermittent period of effort interspersed by recovery periods11; the two most common types of interval training are high-intensity interval training (HIIT) and sprint interval training (SIT).7 HIIT requires ‘near maximal’ efforts performed at a heart rate (HR) ≥80% of the HRmax or the equivalent as expressed in the function of the Embedded Image O2max. Even more intense exercise, SITs are efforts performed at intensities equal or superior to the one that elicited a peak Embedded Image O2 on an incremental test (i Embedded Image O2peak), including ‘all-out’ efforts.7 12

HIIT programmes, when compared with MOD, promote greater increases in Embedded Image O2max,13 ventricular and endothelial function,14 greater or comparable improvements in insulin sensitivity15 and blood pressure,16 lower ratings of perceived exertion,17 similar6 or higher levels of enjoyment,17 18 and similar6 or higher adherence18 than MOD, depending on how the programme is designed. In addition, despite lower training volume in SIT programmes, SIT may promote increases in skeletal muscle oxidative capacity,19 specific metabolic adaptations during exercise19 and exercise performance similar to MOD.20

Decreases in body fat may be similar21 or higher22 in interval training than MOD. Interval training may elicit greater weight loss even if the energy expenditure obtained during the interval training is lower23 or equal24 to that during MOD. This may be due to greater resting energy expenditure and fat utilisation immediately following interval training exercise.25 26 However, there are currently many different approaches to performing interval training,27 and there is still no consensus as to which training method (HIIT/SIT vs MOD) is best to reduce body fat. The considerable variability among interval training protocols,27 compared with fairly homogeneous MOD protocols, introduce ‘noise’ in the literature. No study has yet addressed the simple question: Which type of exercise is better for weight loss?

We conducted a systematic review, qualitative appraisal and meta-analysis of studies that directly or indirectly compared the effects of HIIT or SIT with MOD on adiposity. We compared subgroups to test whether (1) the nature of the interval training (HIIT or SIT), (2) sex, (3) baseline body mass index (BMI) or other variables influenced the outcome. We hypothesised that HIIT/SIT would reduce body fat more effectively than MOD.


The results of this systematic review and meta-analysis are presented according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement,28 and was preregistered in the International Prospective Register of Systematic Review (PROSPERO).29

Search strategy

English-language, Spanish-language and Portuguese-language searches of the electronic databases PubMed and Scopus were conducted from inception to 11 December 2017 by two independent researchers (RBV and JPAN). Articles were retrieved from electronic databases using the following search criteria: (interval training OR intermittent training OR high intensity OR sprint interval training OR aerobic interval training HIIT OR HIIE OR high intensity interval training OR high-intensity interval training OR high intensity interval exercise OR high-intensity interval exercise OR high intensity intermittent exercise OR high-intensity intermittent exercise OR high intensity intermittent training OR high-intensity intermittent training) AND (continuous training OR moderate-intensity continuous exercise OR moderate intensity continuous exercise OR moderate-intensity continuous training OR moderate intensity continuous training) AND (body fat OR adiposity OR body composition OR abdominal fat OR visceral fat OR adipose tissue) AND Humans.

Initially, titles and abstracts of identified studies were checked for relevance by two reviewers (RBV and JPAN). Subsequently, the reviewers independently reviewed the full text of potentially eligible studies. Any disagreement for inclusion between the reviewers was resolved by a third researcher (PG). Additional studies were identified via hand-searching and reviewing the reference lists of relevant papers. All these steps were performed for 3 weeks. Figure 1 presents the flow of papers through the study selection process.

Figure 1

Flow diagram of outcomes of review. HIIT, high-intensity interval training; MOD, moderate-intensity continuous training; SIT, sprint interval training.

Inclusion and exclusion criteria: participants, interventions, comparators and outcomes

Studies with participants of all ages and sexes with a minimum exercise training duration of 4 weeks, which directly or indirectly compared HIIT or SIT with MOD as the primary or secondary aim (according to previous definitions), and which evaluated fat change by methods that infer total or regional mass, or total or regional percentage fat, were included. Studies that reported only BMI and that compared HIIT or SIT or MOD with only non-training control groups were not included for analysis. When employing two interval training protocols, both were included, in different analyses, for comparison with MOD. Studies were excluded based on the following article types: letters to the editor, books, book sections, theses, film/broadcasts, opinion articles, observational studies and abstracts without adequate data, or reviews.

Data extraction

The following study characteristics were extracted: age, sex, body mass, BMI, Embedded Image O2max/peak, total or regional fat mass (kg), percentage total and regional body fat (%), and HIIT/SIT and MOD interventions characteristics. These data were extracted independently by two researchers (RBV and JPAN), with disagreements resolved by a third researcher (PG). When studies provided insufficient data for inclusion in the meta-analysis (five studies), the corresponding authors were contacted via email to determine whether additional data could be provided; however, no corresponding authors responded.

Study quality assessment

Study quality was assessed by two researchers (RBV and JPAN) using a modified Downs and Black checklist.30 Items included the appropriate reporting of the hypotheses, outcomes, interventions, adverse events, participant characteristics (a clear statement on inclusion and exclusion criteria), descriptions of patients lost to follow-up (studies with ≥10% dropout without characteristics reported scored 0), assessment method accuracy, statistical methods, blinding and randomisation procedures. The scale was modified to include criteria for monitoring and reporting of physical activity level (yes=1, no=0) and diet (yes=1, no=0), the supervision of exercise sessions (yes=1, no=0), and information about adherence and/or compliance to exercise interventions (yes=1, no=0). Therefore, the studies that monitored and reported control of diet, habitual activity, supervision, and adherence or compliance scored 1 point in each item. If an item was unable to be determined, it was scored as 0. The highest possible score for quality was 20. In addition, we recorded the strengths/weakness/unknowns of available information from studies to strengthen the quality analysis of the included studies.

Statistical analyses

All analyses were conducted using the R package (V.3.2.4). Meta-analysis was conducted using a random-effects model (DerSimonian and Laird approach) for the individual effects of HIIT/SIT and MOD on total body fat (kg) and body fat percentage. The random-effects model was preferred to a fixed-effect model as certain experimental parameters had wide variation. For the secondary meta-analysis, premeans, postmeans, absolute and relative changes, and SD for each group were collected. Initially, a within-group effect size (ES) was calculated using a random-effects model (DerSimonian and Laird approach) to estimate the change from baseline for each group, given that a random-effects model considers true random errors within a single study and variation in effects occurring from study to study. The statistical heterogeneity of the treatment effect among studies was assessed using Cochran’s Q test and the inconsistency I2 test, in which values above 30% and 50% were considered indicative of moderate and high heterogeneity, respectively.31 Publication bias was assessed with funnel plots and Begg’s test. To improve our results, we conducted several sensitivity analyses to consider the individual influence of each study on the overall results, as well as the type of comparison group (HIIT or SIT), type of modality (walking/jogging/running or cycling), age (<30 and ≥30 years), sex, BMI (<30 kg/m2 and ≥30 kg/m2), intervention duration (<12 and ≥12 weeks), study quality (‘low, middle and high’) and supervision of exercise sessions (yes and no).


Included studies

The search strategy retrieved 786 records. After deduplication and language examination (English, Spanish and Portuguese studies), 24 studies were excluded from the review process and 698 were excluded after title and/or abstract analysis; 64 full-text copies of the remaining studies were obtained and subjected to further evaluation. After reading full-text copies, 23 studies were excluded from this review due to the following reasons: (1) four studies included MOD in combination with the HIIT protocol23 32–34; (2) one study did not perform MOD35; (3) four studies applied an HIIT or MOD intervention combined with other activities36–39; (4) five studies did not use the MOD criteria adopted in this review40–44; (5) five studies stated the use of HIIT or SIT protocols,45–49 but did not match the HIIT and SIT criteria adopted in this review; (6) three studies did not assess body fat50–52; and (7) one study did not present separate body composition data.53 At the end of the process, 41 publications meeting the eligibility criteria were included for qualitative analyses,22 52–91 of which 5 studies provided insufficient data and were excluded from this review due to the following reasons: (1) two studies did not provide SD for change in mean values67 92; (2) two studies reported only regional body fat percentage (%)58 84; and (3) one study provided only skinfold values.55 Therefore, 36 studies provided sufficient data for meta-analysis (35 for total body fat percentage [%] and 15 for total absolute fat mass [kg]) (figure 1).

Participant characteristics

Participants’ characteristics are summarised in table 1. Overall, 1115 participants were included in the qualitative analysis and 1012 in the meta-analysis. The number of participants in the studies varied from 768 to 90.78 Fourteen studies examined exclusively males,55 59 66 67 69 72 73 75 77 79–81 88 92 nine exclusively females,24 63 65 74 82 83 87 90 93 one did not report the number of males and females used to present the body composition results,84 while the remaining studies (n=17) assessed a mixed-sex sample.44 54 56–58 60–62 64 68 70 71 76 78 85 86 89 91 In total, 576 males and 522 females participated in the studies. Two studies used the same sample,61 62 and one performed a double-blind, randomised, crossover investigation with seven athletes (five males and two females).68 The mean age of study participants ranged from 10.464 to 70.1 years.87 The training status of the participants ranged from sedentary73 to high-level athletes.68 The mean BMI ranged from 18.4 kg/m2 91 to 36.7 kg/m2.56

Table 1

Characteristics of the participants

Intervention characteristics

The interval training and MOD programmes are summarised in online supplementary table S1. According to the criteria of HIIT and SIT adopted in this review,7 12 2554–59 63–68 70 73 76–78 81 82 85–87 90 92 93 and 1524 60–62 69 72 74 75 79 80 83 84 88 89 91 of the 41 included studies employed HIIT and SIT interventions, respectively. Only one study71 employed both HIIT and SIT interventions, and one study included two HIIT interventions.86 Of the 41 studies, 20 used cycling,24 55 57 58 69 70 72 73 75 78–81 83 84 87–90 93 16 used walking/jogging/running,54 56 59–67 71 77 82 91 92 1 used a synchronous arm and leg air-braked ergometer,85 1 offered a choice between the two (cycle ergometer or walking/running) depending on orthopaedic limitations,86 1 used swimming,74 1 used boxing drills for the HIIT protocol and walking for MOD,76 and 1 used a rower ergometer.68 Intervention duration ranged from 468 73 75 81 to 16 weeks,86 87 with 12 weeks being the most common (~44%; n=18)54 56 57 59 63 64 66 67 70 71 76 80 82–84 89 92 93 (online supplementary table S1). The most widely used HIIT (n=8) protocol consisted of alternating 4 min at high intensity followed by 3 min of recovery.56 65 71 81 82 85 86 93 The most widely used SIT protocols consisted of alternating 30 s ‘all-out’ efforts followed by 4 min of recovery,60 79 88 and protocols that alternate 8 s ‘all-out’ efforts followed by 12 s of recovery.24 84 The MOD protocols used lasted from 1024 to 60 min,59 74 with 40–45 min (n=6)63 65 70 76 83 87 and 29–35 min (n=6)71 77 82 85 86 90 being the most used protocols.

Twenty-two HIIT protocols used active recovery,54–56 58 63–65 67 68 70 76–78 81 82 84–87 89 90 92 one used passive recovery,93 and five did not report clearly what type of recovery was used.57 59 66 73 91 Eight SIT protocols used active recovery,54 60 69 75 79 80 83 88 one used passive recovery,74 and three did not report clearly what type of recovery was used.61 62 72 The only study to employ HIIT versus SIT versus MOD protocols71 used active recovery in both the interval training protocols.

The intensity of effort for HIIT protocols was prescribed by the percentage of i Embedded Image O2max66 67 77 or i Embedded Image O2peak,58 percentage of Embedded Image O2max55 63 73 or Embedded Image O2peak,70 percentage of HRmax54 56 59 71 78 87 90 92 93 or peak heart rate (HRpeak),64 65 81 82 85 86 rating of perceived exertion,76 HR corresponding to 20% above the HR at ventilatory threshold,57 and 90% of 4 min maximal power.68 The intensity of effort in most SIT protocols (n=13) was prescribed by ‘all-out’ efforts,24 60–62 69 71 72 74 83 84 88 89 91 percentage of i Embedded Image O2peak,80 percentage of maximal power output75 and percentage of anaerobic power.79

More than half of the protocols (~63%; n=26) were performed three times per week.24 54–57 59–62 65 66 70–74 77 80 81 83 84 88–92 Four protocols were performed four times per week,58 76 82 85 two protocols were performed three to four times per week,67 93 three protocols were performed twice a week,64 68 87 one protocol was performed five times per week,63 and five MOD protocols had a frequency greater (five times per week) than the interval training protocols (three times per week).69 75 78 79 86

Diet and physical activity control

Almost half of the studies (~42%; n=17) instructed participants to maintain both normal diet and physical activity.24 60 61 63 69 70 72 73 77 78 81 84 85 87 90 91 93 Twenty-three (~56%) and 19 (~46%) studies reported a diet24 55 58 61 62 64 65 67–71 73 80 82 84 86–88 90–93 and physical activity control,56 58 62 67 68 70 71 73 78 80–82 84–88 91 93 respectively (online supplementary table S2). One study provided a 1-hour diet education session per week.58 In addition, one study employed a caloric reduction of 500 kcal/day based on participants’ normal intake.92 Online supplementary table S1 shows additional information about diet and physical activity of the participants of the included studies.

Body composition assessments

Most studies (~56%; n=23) used only dual-energy X-ray absorptiometry (DXA) to determine total and/or android, trunk and gynoid body composition.24 56 58 59 63 67–71 74 75 77 79–81 84–88 92 93 Others used bioelectrical impedance,57 64 65 78 89 91 hydrodensitometry,54 air displacement plethysmography60 83 or skinfold measurements.61 62 66 72 76 90 Three studies used two body composition assessments methods, such as DXA and CT,87 bioelectrical impedance and CT,82 and bioelectrical impedance and MRI.73 Online supplementary table S3 shows additional information about body composition assessment methods used in the included studies.

Quality assessment

A modified Downs and Black checklist30 assessment determined that the quality of studies had a mean score of 13.5±2.3 (ranging from 954 to 1970; online supplementary table S1). All included studies specified their main findings and outcomes, participant characteristics, statistical tests and accurate measures. Only one study83 did not report variability estimates. No studies blinded participants to exercise intervention, and only eight (~20%) blinded assessors to group allocation.67 68 70 71 76 82 85 86 Most studies (~88%; n=36) randomised participants to groups.24 54–58 60–64 66–71 73–76 78–82 84–93 Twenty-five studies (~61%) reported adherence or compliance.56 59 61 62 64 67 68 70–72 74 76–78 80–86 88–90 93 Nineteen studies (~46%) adequately reported adverse events.57 59–64 70–72 76 77 79 81 82 85 88 89 93 Four studies (~10%) did not provide information about supervision of exercise sessions.72 73 75 76


The within-group analysis found that interval training (−1.50 [95% CI −2.14 to −0.86, p=0.00001]) and MOD (−1.44 [95% CI −2.00 to −0.89, p<0.0001]) resulted in significant improvements in total body fat percentage (%) (online supplementary figures S1A and S2A, respectively). Significant improvements also were found in total absolute fat mass (kg) for HIIT/SIT (−1.58 [95% CI −2.74 to −0.43, p=0.007]) and MOD (−1.13 [95% CI −2.18 to −0.08, p=0.04]) (online supplementary figures S1B and S2B, respectively).

Primary analysis

The between-group analyses on the effects of interval training versus MOD on total body fat percentage (%) and total absolute fat mass (kg) are presented in figure 2 and figure 3, respectively. Overall, there was no difference between groups in total body fat percentage (%) (p=0.705), with evidence of significant heterogeneity in the meta-analysis of total body fat percentage (%) (I2=75.4%, p<0.0001). However, there was a significant difference between groups in total absolute fat mass (kg) (p=0.0094), favouring interval training, with evidence of significant heterogeneity in the meta-analysis of total absolute fat mass (kg) (I2=48.4%, p=0.0184).

Figure 2

Main effects of HIIT/SIT versus MOD on total body fat (%). 4HIIT, high-intensity interval training (four bouts); HIIT, high-intensity interval training; MD, mean of differences; MOD, moderate-intensity interval training; SIT, sprint interval training.

Figure 3

Main effects of HIIT/SIT versus MOD on total absolute fat mass (kg). HIIT, high-intensity interval training; MD, mean of differences; MOD, moderate-intensity interval training; SIT, sprint interval training.

Subgroup analyses demonstrated a significant effect of interval training mode (SIT vs MOD), modality of exercise (walking/jogging/running vs cycling), supervision (yes vs no), study quality (low vs middle vs high), age (<30 vs ≥30 years) and intervention duration (<12 vs ≥12 weeks) on total absolute fat mass (kg) (online supplementary figures S3–S14); however, no significant effect was found on total body fat percentage (%). No effect of sex or BMI was observed on total absolute fat mass (kg) or total body fat percentage (%) (online supplementary figures S15–S18). Table 2 shows a synthesis of these results.

Table 2

Summary of HIIT/SIT versus MOD subgroup meta-analysis on body composition

The mean duration of the HIIT, SIT and MOD protocols included in the analyses of total body fat percentage (%) were 28 min, 18 min and 38 min, respectively. The percentage reductions of total body fat percentage (%) for these protocols were, on average, 4.6%, 3.5% and 3.5%, respectively. On average, HIIT, SIT and MOD protocols included in the analysis on total absolute fat mass (kg) lasted 25 min, 23 min and 41 min, respectively. The percentage reductions of total absolute fat mass (kg) were, on average, 6.0%, 6.2% and 3.4%, respectively.

Sensitivity analyses and publication bias

A sensitivity analysis showed that a significant effect (p<0.05) of HIIT/SIT on total absolute fat mass (kg) remained after removal of each one of the included studies, with evidence of significant heterogeneity (p<0.05). Funnel plots and Begg’s tests for all analyses determined no indication of publication bias.


The present study analysed data from studies that compared the effects of interval training and MOD on body adiposity in humans. The analysis combined 41 studies (36 for meta-analysis) involving a total of 1115 participants. Most studies included in the meta-analysis (86.1%) involved a small sample size (<20 participants per intervention); therefore, the lack of statistical power might have prevented the detection of between-group differences in isolated studies. Notwithstanding, by pooling the data, we did not find superiority of either interval training or MOD in the reduction of total body fat percentage (%), as previously reported in an earlier meta-analysis.21 However, when compared with MOD, we found a superiority of interval training in the reduction of absolute total fat mass (kg). Indeed, both interval training and MOD were similarly beneficial in eliciting small improvements in total body fat percentage (%) (HIIT/SIT: −1.50%; MOD: −1.44%) and in total absolute fat mass (kg) (HIIT/SIT: −1.58 kg; MOD: −1.13 kg). However, a significant difference was found between SIT and MOD in total absolute fat mass (kg) (online supplementary figure S3B).

As a result of the sensitivity analysis that removed each study one by one, we noted that the significant difference favouring interval training for total absolute fat mass (kg) reduction remained. To better understand the factors that might influence the results, we critically reviewed individual studies that favoured interval training or MOD. It is noteworthy that the studies were selected based on their impact on our meta-analysis and not necessarily the results reported in the article.

As for the data that supported MOD for a greater reduction in total body fat percentage (%), the study by Buchan et al 61 62 involved adolescents and started with four 30 s ‘all-out’ running bouts interspaced by 30 s of rest, progressing to six bouts with 20 s of rest. This protocol, however, seems unfeasible, since the recommended recovery between bouts in similar protocols is ~8 times the duration of the effort, such that 30 s maximum efforts are usually followed by 4 min of rest.20 As such, it seems unlikely the participants in the HIIT group in the study by Buchan et al 61 62 were able to maintain maximal effort across the exercise bouts. Koubaa et al 66 reported using running intervals of 2 min at 80% of v Embedded Image O2max followed by 1 min of rest for interval training in adolescents. However, they do not report information about the number of bouts nor about rest intervals, which makes it difficult to analyse the protocol. Moreover, neither Buchan et al 61 62 nor Koubaa et al 66 provide data on dietary and physical activity control.

Another study in favour of MOD is by Nybo et al,59 which involved untrained men. During interval training, participants were instructed to exceed 95% of the HRmax at the end of 2 min of running and then rest for 2 min. However, considering that the intensity of effort was controlled only at the end of each bout, it is not possible to be certain about the intensity of effort maintained during each interval. If we consider that HR progressively increases at a constant rate, it might be possible that the participants spent most of the time at an intensity of effort lower than recommended. Moreover, although the participants were oriented to maintain their habitual lifestyle and dietary practices, the authors did not control for this variable.

Some studies favoured interval training for per cent body fat loss. Panissa et al 90 used a 22 min protocol with 1 min at 90% and 30 s at 60% of HRmax. However, it seems again unfeasible for participants to achieve the prescribed intensity of effort based on the percentage of HRmax since both the times taken to increase and decrease HR seem too short. For example, in the study of Ramos et al,86 participants took 2 min to reach a similar intensity of effort, and a previous study showed that HR decreases ~30 beats per minute in the first minute after intense exercise.94 The results of the study by Thomas et al 54 also favoured interval training for reductions in per cent body fat in a mixed sample of men and women. MOD involved running for 3.2 or 6.4 km at 75% of HRmax, while interval training involved eight bouts of running for 1 min at 90% HRmax followed by 3 min of walking. However, a limitation of this study was the absence of diet and physical activity control. Macpherson et al 60 compared the effects of SIT (4–6 ‘all-out’ efforts of 30 s in a manually driven treadmill with 4 min of rest) and MOD (30–60 min running at 60% Embedded Image O2peak) in a mixed sample of physically active men and women. Their data pointed towards greater decreases in per cent body fat for SIT; however, while the authors reported to have encouraged the participants to maintain their physical activity and diet patterns, there were no objective measures of these variables.

These individual studies highlight the difficulty of drawing general conclusions about the application and effects of interval training on body composition. The inconsistent results might be linked to factors such as habitual diet and physical activity behaviours, since only 23 (~56%) and 19 (~46%) of the included studies reported diet and physical activity control, respectively. Another aspect that needs to be considered is the quality of studies performing interval training and MOD. When considering only the included studies with middle quality, intervention duration less than 12 weeks or with participants’ age less than 30 years, our results found a significant reduction in absolute total fat mass (kg) favouring interval training, although no significant difference was found on total body fat percentage (%). This suggests an influence of the methodological quality of the studies and participants’ characteristics on the results. Moreover, other aspects that might influence weight loss, such as hormonal status,95 sleeping patterns96 and mood disorders,97 are not usually analysed in these studies.

With regard to the factors inherent to interval training, the absence of adequate control for supervision, intensity of effort and the effort to rest ratio might be associated with at least some of the inconsistent results. The subgroup meta-analysis demonstrated that improvements in total absolute fat mass (kg) caused by interval training are higher with supervision during interval training protocols, providing evidence that supervision during interval training is an important variable for total absolute fat mass (kg) reduction. This might occur because it can help interval training practitioners to train with higher intensity of effort.98 Considering that supervision might guarantee adherence to the prescribed protocol, the results provided by studies with supervised sessions are probably more reliable. Therefore, it is important that interval training studies consider providing supervision to guarantee accountability.

In this sense, some examples of the inconsistency with regard to the intensity of effort can be obtained from the analysis of individual studies. For example, Keating et al 70 reported that their protocol was based on the study by Little et al,99 which used 10 bouts of 60 s at a load that elicited 90% of HRmax interspaced by 60 s of recovery. However, in the study by Keating et al,70 HIIT was performed at 120% of i Embedded Image O2max with 30–60 s duration and 120–180 s of rest. On the other hand, at 120% of i Embedded Image O2max, a previous study used seven bouts of 30 s interspaced with 15 s of rest.13

In addition, another interesting example of the possible influence of the control of intensity of effort on the results might be found in Ramos et al.86 While the protocol was reported to involve four bouts of 4 min at 85%–95% of HRpeak with 3 min intervals, the participants took 2 min to reach the targeted intensity. Therefore, the protocol seems to have involved 2 min of the actual prescribed intensity of effort. In other words, the time at target intensity of effort was not reached as planned, resulting in a lower effort to rest ratio than intended. We would like to note that these individual observations do not invalidate or question the merit of previous studies. These are only some aspects that might explain the large inconsistency among the results of interval training and which we must consider when analysing and reproducing previous studies.

Our results found that the effect of interval training protocols when performed using walking/jogging/running modalities on total absolute fat mass (kg) is greater than for MOD with the same modalities. However, the number (n=5) of studies included in this analysis was low,54 60 64 69 82 and only the study by Zhang et al 82 monitored and reported to control diet and habitual activity. Moreover, our data also showed an influence of exercise supervision, and a separate analysis showed that interval training resulted in greater loss of total absolute fat mass (kg) than MOD when training was supervised. Possibly, supervision might influence accountability, influencing adherence to the prescribed intensity of effort.98 Indeed, for other exercise modalities, such as resistance training, supervision has been shown to impact significantly on the intensity of effort and outcomes.98

Separate analyses with HIIT and SIT showed that ‘all-out’ SIT promotes greater total absolute fat mass (kg) reduction than MOD. The greater decreases in fat loss promoted by SIT might be due to the increases in postexercise fat oxidation, which seems to be associated with glycogen depletion.25 100 Indeed, vigorous exercise may be mediated by a more pronounced increase in the skeletal muscle oxidative capacity and by a sympathoadrenal stimulation.101 Thus, protocols that rely more on the glycolytic system might be more beneficial to body fat reductions.102 103

In general, although our findings suggest that MOD provides similar benefits to interval training for total body fat percentage (%) reduction, interval training might be an efficacious, ‘time-efficient’ exercise strategy for body fat management, since the MOD protocols examined in the included studies usually had a greater duration than interval training protocols and provide similar reductions in total body fat percentage (%). For example, MOD protocols lasted on average 38 min (average of 35 sessions × 38 min/session=1330 min) and provided a reduction of 3.5% in total body fat percentage (%), while HIIT protocols lasted on average 28 min (average of 33 sessions × 28 min/session=924 min) and provided a reduction of 4.6%, and SIT protocols lasted on average 18 min (average of 29 sessions × 18 min/session=526 min) and provided a reduction of 3.5% in total body fat percentage (%). In other words, MOD protocols provided a reduction of ‘0.0026% per minute’, while HIIT and SIT protocols provided a reduction of ‘0.0050% and 0.0067% per minute’ in total body fat percentage (%), respectively. The analysis showed that interval training promotes greater reductions in total absolute fat mass (kg) than MOD, despite requiring less time to be performed. However, it is important to be aware of the possible risks and caveats associated with higher intensity training. For example, it might increase the risk of injury and impose higher cardiovascular stress. Adherence should also be examined, as higher intensity protocols can result in higher discomfort.

A common criticism of meta-analysis is the combination of largely heterogeneous studies that have important methodological differences, which can influence the reported effects particularly when the number of included studies is low.104 105 According to Grindem et al,106 heterogeneity is a key factor in the decision to pool or not to pool the results of available studies, which makes it a challenging issue in systematic reviews. The included studies presented relatively high heterogeneity in the total meta-analysis of total body fat percentage (%) and total absolute fat mass (kg), and this may be a reflection of the large heterogeneity in exercise protocols used in the included studies. Broadly speaking, the different protocols (HIIT/SIT or MOD) seem similarly effective in modulating body adiposity in humans; however, the varied approaches used make it difficult to draw general conclusions and recommendations about the ‘ideal’ interval training or MOD protocol. Therefore, clinicians must be careful when interpreting these results and applying them to their practice. Future studies must improve their methodological quality, sample size and method of assessment of change in total body fat to provide more compelling evidence in favour of a specific protocol, or to elucidate the principles of protocol design that appear to have the greatest influence on outcomes.


The present systematic review with meta-analysis showed that interval training provides benefits similar to MOD in total body fat percentage (%) reduction; however, interval training provided a greater total absolute fat mass (kg) reduction than MOD. SIT resulted in greater total absolute fat loss when compared with MOD. A number of factors may positively influence the effects of interval training on total absolute fat mass, including supervision of exercise, walking/running/jogging as the exercise of choice, age (<30 years), study quality and intervention duration (<12 weeks). In general, our findings suggest that the ‘signal in the noise’ is the similar effects of interval training and MOD on total body fat percentage (%) management and the superiority of interval training for total absolute fat mass (kg) reduction, yet that these effects can be produced in a ‘time-efficient’ manner when using interval training.

What is already known

  • Physical activity may be a useful tool to reduce body adiposity; however, many people fail to adhere to exercise programmes due to lack of time or lack of results.

  • Interval training is an attractive alternative to address overweight and obesity given its potential to offer benefits similar to moderate-intensity continuous training while requiring less time.

  • There are currently many different approaches to performing interval training, yet there is no consensus as to which training method or protocol is ‘best’ for reducing body adiposity.

What are the new findings

  • Interval training and moderate-intensity continuous training provide similar benefits for body fat percentage reduction; however, interval training provides greater reductions in total absolute fat mass.

  • Supervision, walking/running/jogging, age, study quality and intervention duration seem to favourably influence the decreases in body adiposity observed from interval training programmes.

  • There is great methodological diversity among interval training protocols in the literature, which makes it difficult to generally recommend that one particular protocol is ‘best’ for modulating body adiposity in humans.


View Abstract

Supplementary materials


  • Contributors RBV and JPAN carried out the screenings and reviews. RBV and VSC carried out the analysis of the articles. RBV and PG drafted and revised the manuscript. CABdL, VSC, JS, JPF and PG revised the manuscript. All authors read and approved the final manuscript.

  • Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.