Cluster sets (CSs) are a popular resistance training (RT) strategy categorised by short rest periods implemented between single or groups of repetitions. However, evidence supporting the effectiveness of CSs on acute intra-session neuromuscular performance is still equivocal.
The objective of this investigation was to determine the efficacy of a single session of CSs to attenuate losses in force, velocity and power compared to traditional set (TS) training.
Screening consisted of a systematic search of EMBASE, Google Scholar, PubMed, Scopus and SPORTDiscus. Inclusion criteria were (1) measured one or more of mean/peak force, velocity or power; (2) implemented CSs in comparison to TSs; (3) an acute design, or part thereof; and (4) published in an English-language, peer-reviewed journal. Raw data (mean ± standard deviation) were extracted from included studies and converted into standardised mean differences (SMDs) and ± 95% confidence intervals (CIs).
Twenty-five studies were used to calculate SMD ± 95% CI. Peak (SMD = 0.815, 95% CI 0.105–1.524, p = 0.024) and mean (SMD = 0.863, 95% CI 0.319–1.406, p = 0.002) velocity, peak (SMD = 0.356, 95% CI 0.057–0.655, p = 0.019) and mean (SMD = 0.692, 95% CI 0.395–0.990, p < 0.001) power, and peak force (SMD = 0.306, 95% CI − 0.028 to 0.584, p = 0.031) favoured CS. Subgroup analyses demonstrated an overall effect for CS across loads (SMD = 0.702, 95% CI 0.548–0.856, p < 0.001), included exercises (SMD = 0.664, 95% CI 0.413–0.916, p < 0.001), experience levels (SMD = 0.790, 95% CI 0.500–1.080, p < 0.001) and CS structures (SMD = 0.731, 95% CI 0.567–0.894, p < 0.001) with no difference within subgroups.
CSs are a useful strategy to attenuate the loss in velocity, power and peak force during RT and should be used to maintain neuromuscular performance, especially when kinetic outcomes are emphasised. However, it remains unclear if the benefits translate to improved performance across all RT exercises, between sexes and across the lifespan.
Cluster set (CS) training is an effective means of attenuating velocity and power loss during a resistance training session.
CSs appear to be most beneficial for moderate- and high-load paradigms where fatigue has the potential to impair performance.
Additional research is needed in order to fully understand the benefits of CSs with additional exercises, between sexes and across the lifespan.
Resistance training (RT) is a fundamental component of athletic development, with the aim of improving performance and minimising injury risk [1,2,3,4]. In particular, the work performed during a RT session provides the necessary stimuli for metabolic, muscular and neuromuscular adaptations to occur and, thus, improve performance over time. Furthermore, it is well-established that specific neuromuscular adaptations occur in response to the training stimuli . As such, the manipulation of mechanical stimuli (e.g. movement velocity and load) is considered to be a key training strategy when focusing on the development of muscular strength and power [6, 7].
In practice, designated training blocks are prescribed to progressively increase physiological stress and, thus, develop specific neuromuscular traits (i.e. hypertrophy, strength or power). Fundamentally, RT prescription has focused on empirically based set and repetition schemes performed in a continuous traditional set (TS) configuration [8, 9], such that during TS training, rest intervals are only implemented after the completion of each set. During the early phase of periodised training, higher-volume hypertrophy-inducing programmes have previously been implemented [7, 10, 11], before progressing to lower-volume, higher-intensity programmes designed to facilitate maximal strength development [10, 12, 13]. During peaking phases, an emphasis on power, i.e. 3–5 repetitions (not to failure), with loads that correspond to 30–80% of 1 repetition maximum (1RM), are employed . However, novel strategies such as cluster sets (CSs) have gathered interest for their proposed ability to maximise neuromuscular adaptations, provide overload, maintain training intensity and minimise overtraining [15, 16]. Although anecdotal evidence dates back to the 1950s, CSs were first reported in the literature by Roll and Omer  in 1987 and later popularised by Siff and Verkhoshansky . CSs are based on the principle of implementing short, intra-set rest periods between groups of repetitions [15, 19,20,21]. For example, a TS approach may consist of 4 × 6 continuous repetitions with typically 1–3 min of inter-set rest, in comparison to a CS comprising 4 × (2 × 3 clusters) with 15–45 s of ‘intra-set’ rest implemented between each cluster in addition to the inter-set rest period . However, this has also extended to inter-repetition rest strategies, whereby a short rest period is implemented after each repetition, rest re-distribution, whereby the total rest time calculated from a TS protocol is interspersed evenly between groups of repetitions, or the rest–pause method [16, 22, 23]. Despite the recent interest in CS paradigms, it remains unclear which method of CS application is superior, with continuing debate over the true definition of a CS.
Despite the growing popularity of CSs, an understanding of the acute performance benefits over a training session remains limited. Emerging evidence has suggested a reduction in fatigue [23,24,25,26,27] and an attenuation of the loss in force, velocity and power with CSs during a RT session [19, 21, 26, 27]. For example, fatigue during a RT session can severely reduce movement kinetics due to a combination of central (neural) and peripheral (muscular) factors [28, 29]. In particular, this may be caused, at least in part, by an increase in blood lactate concentration and reduction of adenosine triphosphate and phosphocreatine stores. Although fatigue was previously thought to be necessary, the benefit of performing RT close to momentary failure (i.e. repetition maximum paradigms) is still debatable for strength adaptation  and may be adverse for power development. Ultimately, this fatigue contributes to the reduction in velocity, power and work output, especially when performed to repetition failure . Thus, intra-set rest should, at least in theory, attenuate fatigue development and allow for a (1) maintenance in force and velocity (power); (2) maintenance of training intensity; and (3) greater overall amount of work to be performed . Conversely, there are several studies demonstrating that structuring training into CSs does not influence force, velocity or power output [32,33,34]. Such discrepancies are likely caused by a lack of methodological consistency between studies (e.g. loading schemes) or variability in the equipment used to capture kinetic data, rendering interpretation within the literature difficult. In particular, it is unclear how factors such as loading intensity, exercise selection and training status are affected by CS. Thus, some conjecture remains about the effectiveness of the CS and its ability to positively impact performance during RT.
Therefore, the aim of this investigation was to collate and analyse the available CS literature investigating acute neuromuscular performance. We have systematically and meta-analytically reviewed the data to (1) determine the acute neuromuscular responses (i.e. strength, power and velocity) following an acute CS session; (2) make a direct comparison to TS training; and (3) investigate potential differences between exercise selection, loading strategy, experience level and CS structure. These findings will provide clarity regarding the effectiveness of CS training to attenuate the loss of force, velocity and power across a RT session. It is intended that the findings will help better inform strength and conditioning professionals on effective programme design to maximise neuromuscular stimuli and inform future research areas within the field.
The aim of this investigation was to systematically review and present the results of a meta-analysis regarding the effects of CS training on acute neuromuscular performance (i.e. force, velocity and power), with moderators consisting of exercise selection, loading intensity, training experience of the individual and CS structure.
2.1 Research Question and Registration
This systematic review and meta-analysis conformed to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines.
The research questions were defined by the PICOS model in accordance with PRISMA guidelines, as follows:
Population: Males and females with or without RT experience.
Intervention: An acute RT session which incorporated a ‘CS’ design.
Comparator: Acute neuromuscular responses compared to TS.
Outcomes: Peak and/or average force, velocity and/or power.
Study design: Randomised controlled designs, counterbalanced crossover or repeated measure designs that investigated the acute mechanical/neuromuscular responses from CS training.
2.2 Literature Search
Searches for this review were performed using the EMBASE, Google Scholar, PubMed, Scopus and SPORTDiscus electronic databases without any year restriction. The following words were combined and used for the searches through article title, abstract and keyword screening: (‘cluster-set*’ OR ‘cluster loading’ OR ‘cluster-type’ OR ‘inter-set rest’ OR ‘rest redistribution’ OR ‘rest-loading’ OR ‘rest-pause’ OR ‘traditional set’ OR ‘intra set’ OR ‘inter rep*’ OR ‘work-to-rest ratio’) AND (‘power’ OR ‘strength’ OR ‘displacement’ OR ‘neur*’ OR ‘repetition’ OR ‘velocity’ OR ‘endurance’ OR ‘performance’ OR ‘volume’ OR ‘work’ OR ‘hypertroph*’ OR ‘fatigue’ OR ‘force’ OR ‘perceived exertion’). After the removal of duplicates, the title and abstract of each article was initially screened for suitability. Full-text articles were retrieved in order to determine inclusion or exclusion. In each full text, the reference lists were screened for additional articles. In addition, the list of articles that cited the included studies (i.e. forward citation tracking) were screened. Two authors (CL and GH) performed the search independently. In the case of any selection bias, a third assessor (W-PT) was included. The search was conducted throughout September of 2017 and updated in August of 2018.
2.3 Dependent Variables
Dependent variables were grouped into force (maximal/peak and/or average from isometric or dynamic movements), velocity (maximal/peak and/or average of the movement, bar speed or body during acceleration) and power (maximal/peak and/or average calculated in watts, or determined from jump performance).
2.4 Inclusion and Exclusion Criteria
Studies were included in this review if they met the following criteria: (1) measured one or more of peak or average force, power and velocity; (2) implemented CS in comparison to TS; (3) the study had an acute design or part thereof; and (4) was published in an English-language peer-reviewed journal. Data (mean ± standard deviation [SD]) from studies that only reported the results in graphical form were extracted using plot digitising software (PlotDigitizer; https://automeris.io/WebPlotDigitizer/). If this method was not suitable, the author(s) of the studies were contacted to obtain original raw data and subsequently excluded if sufficient data for the analysis of the standardised mean difference (SMD) was unavailable or the authors could not be contacted. Articles that did not include a TS condition as a comparator were also excluded from the analysis.
2.5 Data Extraction
For all included articles, the following data were extracted: (1) study characteristics (author, year, sample size and study design); (2) participant demographics (age, sex and RT experience); (3) RT protocols (CS and TS structure [i.e. rest period, repetitions, number of sets, CS configuration, exercise selection and intensity]); and (4) outcome measures (maximal/peak and/or average force, velocity and power). Quantitative data (mean and SD) from pre- and post-training session, first and last repetition or, where necessary, first and last set were extracted from text, tables and figures if required. Where multiple post-training timepoints were reported, the timepoint immediately following the RT session was used. Where the standard error was reported, this was converted post hoc to SD. To increase reliability, data were extracted by two independent assessors (CL and GGH), and in the case of a discrepancy a third assessor (KK) was used as a moderator.
2.6 Statistical Analysis
As systematic influences and random errors were predicted to be present between study-level ES, random effects meta-analyses were conducted for each of performance variables (i.e., force, velocity and power). All performance variable outcomes were presented as averaged SMD ± 95% confidence interval (CI) values. For each study, SMD was computed such that positive values indicate that the intervention group (i.e. CS training) was superior to the control group (i.e. TS training) . Subgroup analyses were agreed upon a priori to assess the influence of moderator variables of RT on physical performance. Where studies had more than one outcome measure in a particular subgroup, they were combined into a single effect size for analysis . This was done to limit the risk of bias of the aggregated effect of comparing the same dataset within the same meta-analysis. Moderator variables in this study included the following:
Training load: power (optimal load determined for power development regardless of relative value to 1RM), and low (≤ 60% 1RM), moderate (60–79% 1RM) or heavy (defined as either ≥ 80% 1RM or ≥ 6RM load), irrespective of optimal load for power development.
Exercise type: strength training (compound or isolated task) versus weightlifting (WL) versus strength + WL versus power.
Training experience: athletic (State-level or above athletes) versus experienced (> 12 months’ RT experience or could squat 1.5 × body weight) versus recreational (physically active and/or < 12 months’ RT experience).
CS structure: inter-repetition rest versus intra-set rest versus rest–pause.
Heterogeneity was measured using the I2 statistic, which indicates the percentage of variance between studies, with cutoff points corresponding to low (0–25%), moderate (26–50%) and high (51–100%) heterogeneity . Funnel plots were used to assess publication bias using Egger’s regression tests where non-significant asymmetry indicated no bias  (Electronic Supplementary Material [ESM] Figure 1). All statistical analyses were performed using Comprehensive Meta-Analysis (version 3.0; Biostat, Englewood, NJ, USA). An α level of p < 0.05 was used to determine statistical significance.
2.7 Methodological Quality and Bias
The methodological quality for each study was evaluated using a modified 11-point Physiotherapy Evidence Database (PEDro) scale; the quality of each study was assessed independently by two authors (CL and KK). Given that it is not possible to blind the participants and investigators in supervised exercise interventions, items 5–7 from the scale, which are specific to blinding, were removed. This approach has been used in previous systematic reviews in the area of RT [39, 40]. With the removal of these items, the maximum result on the modified ‘PEDro 8-point’ scale was 7 because the first item, related to eligibility criteria, is not included in the total score. The qualitative methodology ratings were adjusted similarly to those used in previous exercise-related systematic reviews [39, 40] and were as follows: 6–7 = ‘excellent’; 5 = ‘good’; 4 = ‘moderate’; and 0–3 = ‘poor’. Two assessors (CL and KK) also assessed the bias of included studies using the Cochrane risk of bias assessment tool . The Cochrane risk of bias tool evaluates each study based on the following criteria: sequence allocation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other sources of bias. A third reviewer (GGH) acted a moderator if there were discrepancies in the interpretation of the PEDro or Cochrane risk of bias scales.
3.1 Search Results
The search and screening process is presented as a flowchart in Fig. 1. The initial search identified 2923 potentially relevant articles, with 2386 remaining after the removal of duplicates. An additional 2262 articles were excluded following title and abstract screening, and 124 full-text articles were then assessed for eligibility. Based on the selection criteria, a total of 25 were included in the meta-analysis with a total participant sample size of n = 317. General examples of the TS and CS paradigms employed in the literature can be found in Fig. 2.
3.2 Methodological Quality and Bias
The PEDro scores for the studies in this review ranged from 5 to 6 (mean = 5.7 ± 0.5) (ESM Table 1). Therefore, this result indicates that the evidence used in this review comes from studies with a ‘good’ methodological quality. The Cochrane risk of bias scores indicate a low risk of bias for four of the seven domains (ESM Table 2). Given that allocation concealment, blinding of participants/personnel and outcomes was not feasible in the included studies, we conclude that the generally low risk of bias does not seriously alter the results within or between studies.
3.3 Meta-Analytical Results
A summary of the methods and findings from individual studies is shown in Table 1.
3.3.1 Kinetic Variables
Power was the most assessed outcome (16 individual studies, n = 181 individuals) (peak power: SMD = 0.356, 95% CI 0.057–0.655, p = 0.019; mean power: SMD = 0.692, 95% CI 0.395–0.990, p < 0.001) [19, 21,22,23,24,25, 41, 42, 44,45,46,47,48,49, 51, 59], followed by velocity (14 individual studies, n = 170 individuals) (peak velocity: SMD = 0.815, 95% CI 0.105–1.524, p = 0.024; mean velocity: SMD = 0.863, 95% CI 0.319–1.406, p = 0.002) [21, 25, 43, 46,47,48,49,50,51,52, 55,56,57,58] and then force (11 individual studies, n = 123 individuals) (peak force: SMD = 0.306, 95% CI − 0.028 to 0.584, p = 0.031; mean force: SMD = 0.572, 95% CI − 0.157 to 1.301, p = 0.124) [21, 25, 41, 45, 46, 48, 49, 51, 53,54,55]. The individual study, subgroup analyses and overall SMD ± 95% CI for kinetic variables can be found in Fig. 3a–c.
3.3.2 Exercise Selection
A total of 20 studies included in the meta-analysis used a strength-based exercise, of which 15 used a back squat or half-squat exercise [23,24,25, 41,42,43, 46,47,48,49,50, 52, 53, 55, 59], three used the bench press exercise [22, 56, 58], one used the deadlift  and one used an isometric knee extension exercise . Two studies assessed a WL task (i.e. clean pulls or power clean) [19, 21], one study used a jump squat (power) , one study used the bench press throw  and one study combined strength and WL exercises . An overall effect for exercise selection was observed (SMD = 0.664, 95% CI 0.413–0.916, p < 0.001), but no differences were detected between strength, WL, power and strength/WL exercises (Q = 2.561, p = 0.431). The individual study, subgroup analysis and overall SMD ± 95% CI for exercise selection can be found in Fig. 4.
A total of 15 studies included in the meta-analysis used a heavy loading scheme [19, 21, 22, 25, 41,42,43, 45, 48,49,50, 52, 53, 55, 59], ten used a moderate loading scheme [23,24,25, 44, 46,47,48, 50, 56, 58], two used a low loading scheme [42, 54] and three used a load considered optimal for power development [42, 51, 57]. It should be noted that three studies used more than one loading scheme [42, 48, 50]. An overall effect for loading intensity was observed (SMD = 0.702, 95% CI 0.548–0.856, p < 0.001), but no differences were detected between low, moderate, heavy and power loading schemes (Q = 2.376, p = 0.301). The individual study, subgroup analysis and overall SMD ± 95% CI for loading intensity can be found in Fig. 4b.
3.3.4 Resistance Training (RT) Experience
Twelve studies included in the meta-analysis used experienced individuals [21, 23,24,25,26, 43, 44, 46, 49, 53, 55, 58], nine studies used recreational individuals [41, 42, 45, 50, 51, 54, 56, 57, 59], while three used athletic individuals [19, 22, 52]. One study  used a combination of recreational and experienced individuals. An overall effect for RT experience was observed (SMD = 0.790, 95% CI 0.500–1.080, p < 0.001), but no differences were detected between recreational, experienced, athletic and mixed experience levels (Q = 4.008, p = 0.332). The individual study, subgroup analysis and overall SMD ± 95% CI for RT experience can be found in Fig. 4c.
3.3.5 Cluster Set (CS) Structure
Fifteen studies included in the meta-analysis used the inter-repetition rest method [19, 21, 22, 41,42,43, 45, 49, 50, 52, 55,56,57,58,59], 11 studies used the intra-set rest method [22,23,24,25, 44,45,46,47,48, 51, 54], while only one study used the rest–pause technique . Two studies [22, 45] used both inter-repetition and intra-set rest in their study designs. An overall effect for CS structure was observed (SMD = 0.731, 95% CI 0.567–0.894, p < 0.001), but no differences were detected between the inter-repetition rest, intra-set rest and rest–pause method (Q = 2.675, p = 0.367). The individual study, and overall SMD ± 95% CI for CS structure can be found in Fig. 4d.
This is the first meta-analytical investigation comparing the acute neuromuscular effects of CS versus TS in RT. Specifically, the results of this investigation demonstrate that velocity and power benefit from the use of CS strategies, with the overall magnitude considered statistically significant. Force was not different between CS and TS strategies. Additionally, the benefit of using CS during an acute bout of RT extends across strength and WL tasks, individual experience levels (i.e. recreational, experienced and athletic) and moderate or heavy loading strategies. No differences were observed between subgroup categories. Thus, strength and conditioning professionals should consider using CS as an efficacious strategy during acute RT sessions. Specifically, CS should be used when kinetic variables are emphasised, such as those targeting the optimisation of velocity and power outcomes regardless of training experience.
4.1 Exercise Selection
The use of CS paradigms demonstrated a collective benefit for strength and WL exercises. Given that it is common to utilise a combination of, or all, exercises (e.g. squat, deadlift, bench press and power clean) concurrently during a RT session, and at various stages of a periodised plan, the findings suggest that CS strategies can be used across multiple exercises to optimise acute performance. Moreover, only one study, Rio-Rodriguez et al.  used a single joint task. Given programmes emphasising power give precedence to multi-joint movements, implementing CSs for isolated tasks is unlikely to offer the same benefit for athletic performance. Moreover, it is important to note that the majority of evidence stems from lower- or full-body tasks, with only three studies [22, 56, 58] investigating the bench press exercise. Lawton et al.  and García-Ramos et al.  demonstrated a significant effect (SMD = 1.442, p = 0.001 and SMD = 4.606, p < 0.001, respectively), despite a non-significant result observed in the study by Mayo et al.  (SMD = 0.302, p = 0.548). Thus, the limited evidence from upper-body investigations makes it difficult to draw conclusions about the overall effectiveness of CSs between upper- and lower-limb tasks. In particular, some evidence suggests that the development of fatigue  and level of perceived exertion  differs between the upper and lower limbs. Specifically, Vernillo et al.  demonstrated that maximal leg exercise induces a greater magnitude of fatigue, approximately 12% more than an equivalent time-equated upper-body task. Thus, it can be speculated that the CS intra-set rest period required for upper-limb tasks may be different than for lower-limb tasks to maintain or attenuate the loss in performance. For example, Mayo et al.  used an inter-repetition rest of 27.4 s, with an improvement observed for the bench press but not back squat exercise when compared to TS. A lower perceived exertion was also reported for the bench press than for squat exercise. Additionally, Lawton et al.  demonstrated that mean power was reduced by 53.8 Watts (W), 66.9 W and 57.0 W with inter-repetition rest of 23 s and intra-set rests of 56 s and 109 s, respectively, during a bench press task. Therefore, although the intra-set or inter-repetition rest intervals in the included studies ranged from 6.0 to 45.4 s for lower- and full-body exercises, the lack of a direct comparison to an upper body-specific task limits the generalisation of these findings. Hence, further evidence is required from research investigating upper-limb tasks, which may be particularly important for sports requiring upper-body strength and power to fully understand the benefits of CS training.
Intense exercise causes a reduction in neuromuscular performance due to the development of central and peripheral fatigue [28, 29]. Previous evidence has suggested that high-intensity, low-volume exercise causes greater central fatigue, while higher-volume loading schemes cause perturbations at the muscular level . Regardless, the development of fatigue, whether central or peripheral in origin, is considered adverse to the development of force and power due to reductions in neural drive and/or disturbances to intramuscular homeostasis [62, 63]. When grouped by loading intensity, the results of this meta-analysis revealed that CSs were beneficial for optimising acute neuromuscular performance for moderate and heavy loads. Interestingly, despite the known differences in peripheral fatigue development between moderate and heavy load RT schemes , no significant effect was found between the included studies. Moreover, the study by García-Ramos et al.  demonstrated that CSs were better than TSs across low, high and optimal loads at attenuating power loss. Likewise, the reduction in velocity was less for all loads between 60 and 80% of 1RM for the back squat in the study by Mora-Custodio et al. , with a benefit also demonstrated by Tufano et al.  using either 75% or 80% of 1RM. This observation warrants some discussion given that the studies utilising moderate loads generally had a higher overall volume/number of performed repetitions [24, 25, 44, 47, 48, 56]. Thus, it could be theorised that the increase in blood lactate concentration and reduction of adenosine triphosphate and phosphocreatine stores  as well as alterations in other biomarkers such as cortisol during higher-volume fatiguing TS protocols  may be attenuated by CS paradigms. In particular, Haff et al.  suggested that the inclusion of short 15–30 s rest intervals may attenuate these changes, which have previously been associated with a reduction in force and velocity during a RT session [6, 66, 67]. However, the results of this meta-analysis do not substantiate these reports. Moreover, it is worth mentioning that it is not clear whether fatigue is required for neuromuscular adaptation to occur . Thus, achieving the same volume load with minimal fatigue development may be a more favourable approach. It should also be noted that biochemical correlates of fatigue were only reported in a handful of the studies [23, 44, 46, 55] examined in this meta-analysis, suggesting that further work in this area is warranted.
Although no significant effect was observed for low load paradigms, this should be interpreted with caution due to the inclusion of only two studies in this subgroup analysis. Although the inclusion of further studies may provide support for CS use with low load paradigms, the results of the study by Rio-Rodriguez et al.  require some consideration in itself. Firstly, Rio-Rodriguez et al.  used a single-joint isometric knee extension task, which makes it challenging to translate the results of this study to exercises typically used in the preparation of athletes. It should also be noted that the findings from the Rio-Rodriguez et al.  study are based on maximal force production and did not consider how the CSs impacted velocity or power. Conversely, although a significant effect was observed for optimal power loading schemes (SMD = 1.030, 95% CI − 0.629 to 1.432), the inclusion of only three studies [42, 51, 57], and the highly significant result from García-Ramos et al. , suggests that further research in this area is required before a confident conclusion can be drawn. However, it can be speculated that as power training programmes are not designed to induce large amounts of fatigue, CSs may not be as effective as high-intensity or high-volume protocols that are performed to muscular failure [28, 29, 31].
4.3 RT Experience
CSs offer an additional level of programming complexity by allowing for the manipulation of the rest periods between clusters of repetitions or after each individual repetitions within a set. Furthermore, RT programmes emphasising power development are commonly used for more experienced individuals, or during the later stages of periodised programmes . The results of this meta-analysis did not reveal any significant difference between recreational, experienced and athletic individuals. It should be noted that only three studies used athletic [19, 22, 52] individuals and, likewise, only one study included both recreational and experienced individuals but a subgroup analysis was not reported . However, Oliver et al.  made no comparison between experience levels, and thus caution should be used when interpreting these results. Nonetheless, the available evidence suggests that CSs are an efficacious tool for all individuals, regardless of experience, where the emphasis is on maximising kinetic variables during RT.
4.4 CS Structure
As the popularity of CS expands, research continues to investigate the manipulation of the within-set rest periods in an attempt to optimise performance. For example, inter-repetition rest, intra-set rest and the rest–pause method are commonly referred to as a ‘cluster set’. However, the differences in each structure and the subsequent effect on acute neuromuscular performance warrant some discussion.
The results of this meta-analysis revealed a significant benefit for both the inter-repetition rest and intra-set rest CS structures, with less evidence available for the rest–pause method. Specifically, the results of the two studies that included both inter-repetition and intra-set rest in the same investigation [22, 45] did not report any differences between each CS structure. Thus, the evidence from Moir et al. , Lawton et al.  and the collective evidence presented in this meta-analysis suggests that both inter-repetition and intra-set rest schemes provide an effective means of optimising acute neuromuscular performance. Although no significant effect was observed for the rest–pause method, the fact that only one study, Marshall et al. , was able to be included in this subgroup analysis limits the ability to draw confident conclusions regarding this technique. However, as the sets in the study by Marshall et al.  were performed until momentary failure, the effectiveness of introducing short rest intervals may be diminished due to accumulated fatigue prior to the implementation of the rest period. Furthermore, initial force and power outputs may differ between set structures (i.e. higher volume vs. lower volume) and, thus, the relative decrease across a set or relative difference between TS and CS should also be considered when interpreting the literature. Therefore, future research investigations are warranted to determine the effectiveness of each CS structure across independent variables (i.e. exercise selection, loading parameters and experience level) in RT.
4.5 Research Recommendations
Given the growing use of CS in applied settings and the gaps highlighted in this meta-analytic review, we suggest several future directions for research in this space. First and foremost, it is clear that there is a paucity of research examining the efficacy of using CS with female cohorts. Although there are known sex differences in the development of exercise induced fatigue [68, 69], it is currently unclear how CSs, which attenuate fatigue development, modulate acute performance in female cohorts. Specifically, given the importance of kinetic variables in athletic performance, distinguishing the effect of CSs on intra-session force, velocity and power characteristics between males and females is warranted. Secondly, the included studies are based on a demographic of young, healthy adults. It has also been established that fatigue differences exist across the lifespan (e.g. fatigue resistance and power development)  and, thus, the acute neuromuscular responses to CSs likely differ between the young and old. In particular, power may be of more importance than maximal strength in functional tasks, which likely holds greater relevance in aging populations. For example, recent evidence has supported the use of CS RT interventions to improve functionality in elderly individuals . Furthermore, this review has also highlighted that a relatively large percentage of the evidence stems from lower- or full-body RT exercises, especially the back squat. Thus, future research should also seek to further investigate non-stretch–shorten cycle multi-joint tasks (i.e. deadlift) and applications to strength and power resistance exercises in the upper limbs. Of further interest is that CSs did not have an effect on mean force but may potentially attenuate losses in peak force. However, as suggested in previous work [25, 47], movement velocity, rather than force (especially mean), is considered to be the main factor influencing power output. Based on the available evidence from the literature, we cannot say for certain whether other factors such as a change in impulse or movement strategy (i.e. that which affects range of motion) also contributed at least partly to this observation. Lastly, given biochemical correlates were only investigated in a handful of studies [23, 44, 46, 55], further work should seek to understand the effect of volume-matched TS or CS RT on endocrine and other physiological responses and provide a comprehensive profile of fatigue and subsequent recovery following advanced RT paradigms.
From a methodological perspective, the collective body of evidence comes from studies with ‘good’ methodological quality and a low risk of bias. However, it should be noted that seven of the 25 included studies were not, or did not clearly indicate if the conditions were, randomised. Thus, future research needs to consider the sequence order of trials in order to minimise the potential learning or order effects that can be associated when randomisation is not utilised. Although both items relating to ‘blinding’ suggest a high level of bias, we acknowledge when performing RT studies it is not possible to blind participants or personnel to the treatment being administered and therefore this should not be considered to be confounding factor in the field of research.
Collectively, the results of this investigation highlight the benefit of CSs to maximise neuromuscular performance during an acute RT session. In particular, the loss of velocity and power, and potentially peak force, can be attenuated via intra-set, inter-repetition and rest–pause paradigms. Given that mean force was not different between CSs and TSs, and power is a function of force and velocity, it seems logical that velocity should be considered in the primary assessment of CS efficacy. Moreover, strength and conditioning professionals should also consider the use of CSs as a tool for maintaining movement velocity across a RT set, or series or sets. Additionally, it is important to consider the impact of the CS design, including intra-set and total repetitions per set, when aiming to maximise velocity and power. Furthermore, when strength and conditioning professionals decide to implement CSs into their athlete training programmes, it is important to realise that these set structures could be beneficial for strength, WL and tasks where moderate and heavy loading schemes are employed. Ultimately, when training to maximise kinetic variables and maintain high-volume loads in a time-efficient manner, CSs can be employed by individuals with a diverse training background ranging from those with minimal to extensive RT experience. While the current research strongly suggests there are positive benefits from employing CS, there is a need for extensive research into the potential differences between the sexes, across the age span and a wider variety of exercises. Finally, future research examining the impact of employing CSs as part of a long-term training programme are warranted to determine if these acute responses translate into long-term performance gains.
Cronin J, Sleivert G. Challenges in understanding the influence of maximal power training on improving athletic performance. Sports Med. 2005;35(3):213–34.
Faigenbaum AD, Myer GD. Resistance training among young athletes: safety, efficacy and injury prevention effects. Br J Sports Med. 2010;44:56–63.
Suchomel TJ, Nimphius S, Stone MH. The importance of muscular strength in athletic performance. Sports Med. 2016;46:1419–49.
Suchomel TJ, Mimphius S, Bellon CR, Stone MH. The importance of muscular strength: training considerations. Sports Med. 2018;48(4):765–85.
Morrissey MC, Harman EA, Johnson MJ. Resistance training modes: specificity and effectiveness. Med Sci Sports Exerc. 1995;27(5):648–60.
Pereira MI, Gomes PS. Movement velocity in resistance training. Sports Med. 2003;33(6):427–38.
Crewther B, Cronin J, Keogh J. Possible stimuli for strength and power adaptation. Sports Med. 2005;35(11):967–89.
DeLorme TA, Watkins AL. Technics of progressive resistance training. Arch Phys Med Rehabil. 1948;29:263–73.
Todd JS, Shurley JP, Todd TC, Thomas L. DeLorme and the science of progressive resistance exercise. J Strength Cond Res. 2012;26(11):2913–23.
Ratamess NA, Alvar BA, Evetoch TK, Housh TJ, Kibler WB, Kraemer WJ, et al. Progression models in resistance training for healthy adults. Med Sci Sports Exerc. 2009;41(3):687–708.
Schoenfeld BJ. The mechanisms of muscle hypertrophy and their application to resistance training. J Strength Cond Res. 2010;24(10):2857–72.
Schoenfeld BJ, Wilson JM, Lowery RP, Krieger JW. Muscular adaptations in low- versus high-load resistance training: a meta-analysis. Eur J Sport Sci. 2016;16(1):1–10.
Campos GE, Luecke TJ, Wendeln HK, Toma K, Hagerman FC, Murray TF, et al. Muscular adaptations in response to three different resistance-training regimes: specificity of repetition maximum training zones. Eur J Appl Physiol. 2002;88(1–2):50–60.
Bird SP, Tarpenning KM, Marino FE. Designing resistance training programmes to enhance muscular fitness: a review of the acute programme variables. Sports Med. 2005;35(10):841–51.
Haff GG, Hobbs RT, Haff EE, Sands WA, Pierce KC, Stone MH. Cluster training: a novel method for introducing training program variation. Strength Cond J. 2008;30(1):67–76.
Tufano JJ, Brown LE, Haff GG. Theoretical and practical aspects of different cluster set structures: a systematic review. J Strength Cond Res. 2017;31(3):848–67.
Roll F, Omer J. Football: Tulane football winter program. Strength Cond J. 1987;9:34–8.
Siff M, Verkhoshansky YU. Supertraining. 4th ed. Denver: Supertraining International; 1999.
Haff GG, Whitley A, McCoy LB, O’Bryant HS, Kilgore JL, Haff EE, et al. Effects of different set configurations on barbell velocity and displacement during a clean pull. J Strength Cond Res. 2003;17(1):95–103.
Haff GG, Burgess S, Stone MH. Cluster training: theoretical and practical applications for the strength and conditioning professional. Prof Strength Cond. 2008;12:12–7.
Hardee JP, Triplett NT, Utter AC, Zwetsloot KA, McBride JM. Effect of interrepetition rest on power output in the power clean. J Strength Cond Res. 2012;26(4):883–9.
Lawton TW, Cronin JB, Lindsell RP. Effect of interrepetition rest intervals on weight training repetition power output. J Strength Cond Res. 2006;20(1):172–6.
Oliver JM, Kreutzer A, Jenke S, Phillips MD, Mitchell JB, Jones MT. Acute responses to cluster sets in trained and untrained men. Eur J Appl Physiol. 2015;115:2383–93.
Joy JM, Oliver JM, McCleary SA, Lowery RP, Wilson JM. Power output and electromyography activity of the back squat exercise with cluster sets. J Sports Sci. 2013;1:37–45.
Tufano JJ, Conlon JA, Mimphius S, Brown LE, Seitz LB, Williamson BD, et al. Maintenance of velocity and power with cluster sets during high volume back squats. Int J Sports Physiol Perform. 2016;11:885–92.
Tufano JJ, Conlon JA, Nimphius S, Brown LE, Petkovic A, Frick J, et al. Effects of cluster sets and rest-redistribution on mechanical responses to back squats in trained men. J Hum Kinet. 2017;58:35–43.
Tufano JJ, Conlon JA, Nimphius S, Oliver JM, Kreutzer A, Haff GG. Different cluster sets result in similar metabolic, endocrine, and perceptual responses in trained men. J Strength Cond Res. 2019;33(2):346–54.
Zajac A, Chalimoniuk M, Maszczyk A, Golas A, Lngfort J. Central and peripheral fatigue during resistance exercise—a critical review. J Hum Kinet. 2015;49:159–69.
Enoka RM, Duchateau J. Translating fatigue to human performance. Med Sci Sports Exerc. 2016;48(11):2228–38.
Nobrega SR, Libardi CA. Is resistance training to muscular failure necessary? Front Physiol. 2016;7:10.
Sanchez-Medina L, Gonzalez-Badillo J. Velocity loss as an indicator of neuromuscular fatigue during resistance training. Med Sci Sports Exerc. 2011;43(9):1725–34.
Asadi A, Ramirez-Campillo R. Effects of cluster vs traditional plyometric training sets on maximal-intensity exercise performance. Medicina (Kaunas). 2016;52:41–5.
Nicholson G, Ispoglou T, Bissas A. The impact of repetition mechanics on the adaptations resulting from strength-, hypertrophy- and cluster-type resistance training. Eur J Appl Physiol. 2016;116:1875–88.
Yazdani S, Aminaei M, Amirseifadini M. Effects of plyometric and cluster resistance training on explosive power and maximum strength in karate players. Int J Appl Exerc Physiol. 2017;6(2):34–44.
Schmid JE, Koch GG, LaVange LM. An overview of statistical issues and methods of meta-analysis. J Biopharm Stat. 1991;1(1):103–20.
Wood JA. Methodology for dealing with duplicate study effects in a meta-analysis. Org Res Methods. 2008;11:79–95.
Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539–58.
Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–34.
Kummel J, Kramer A, Giboin LS, Gruber M. Specificity of balance training in healthy individuals: a systematic review and meta-analysis. Sports Med. 2016;46:1261–71.
Schoenfeld BJ, Grgic J, Ogborn D, Krieger JW. Strength and hypertrophy adaptations between low- vs. high-load resistance training; a systematic review and meta-analysis. J Strength Cond Res. 2017;31(12):3508–23.
Boullosa DA, Abreu L, Beltrame LG, Behm DG. The acute effect of different half squat configurations on jump potentiation. J Strength Cond Res. 2013;27(8):2059–66.
García-Ramos A, Nebot V, Padial P, Valverde-Esteve T, Pablos-Monzó A, Feriche B. Effects of short inter-repetition rest periods on power output losses during the half squat exercise. Isokinet Exerc Sci. 2016;1:1–8.
Iglesias-Soler E, Carballeira E, Sánchez-Otero T, Mayo X, Fernández-del-Olmo M. Performance of maximum number of repetitions with cluster-set configuration. Int J Sports Physiol Perform. 2014;9:637–42.
Girman JC, Jones MT, Matthews TD, Wood RJ. Acute effects of a cluster-set protocol on hormonal, metabolic and performance measures in resistance-trained males. Eur J Sport Sci. 2014;14(2):151–9.
Moir GL, Graham BW, Davis SE, Guers JJ, Witmer CA. Effect of cluster set configurations on the mechanical variables during the deadlift exercise. J Hum Kinet. 2013;39:15–23.
Oliver JM, Jenke SC, Mata JD, Kreutzer A, Jones MT. Acute effect of cluster and traditional set configurations on myokines associated with hypertrophy. Int J Sports Med. 2016;37:1019–24.
Oliver JM, Kreutzer A, Jenke SC, Phillips MD, Mitchell JB, Jones MT. Velocity drives greater power observed during back squat using cluster sets. J Strength Cond Res. 2016;30(1):235–43.
Tufano JJ, Conion JA, Nimpius S, Brown LE, Banyard HG, Williamson BD, et al. Cluster sets: permitting greater mechanical stress without decreasing relative velocity. Int J Sports Physiol Perform. 2017;12:463–9.
Wagle JP, Cunanan AJ, Carroll KM, Sama ML, Wetmore A, Bingham GE, et al. Accentuated eccentric loading and cluster set configurations in the back squat: a kinetic and kinematic analysis. J Strength Cond Res. 2018. https://doi.org/10.1519/JSC.0000000000002677(Epub 2018 Jun 20).
Mora-Custodio R, Rodriguez-Rosell D, Yáñz-Gracía JM, Sánchez-Moreno M, Pareja-Blanco F, González-Badillo JJ. Effect of different inter-repetition rest intervals across four load intensities on velocity loss and blood lactate concentration during full squat exercise. J Sports Sci. 2018;36(24):2856–64. https://doi.org/10.1080/02640414.2018.1480052.
Koefoed N, Lerche M, Jensesn BK, Kjaer P, Dam S, Horslev R, et al. Peak power output in loaded jump squat exercise is affected by set structure. Int J Exerc Sci. 2018;11(1):776–84.
Nickerson BS, Mangine GT, Williams TD, Martinez IA. Effect of cluster set warm-up configurations on sprint performance in collegiate male soccer players. Appl Physiol Nutr Metab. 2018;43:625–30.
Marshall PWM, Robbins DA, Wrightson AWS, Siegler JC. Acute neuromuscular and fatigue responses to the rest-pause method. J Sci Med Sport. 2012;15:153–8.
Rio-Rodriguez D, Iglesias-Soler E, Fernandez-del-Olmo M. Set configuration in resistance exercise: muscle fatigue and cardiovascular effects. PLoS One. 2016;11(3):e0151163.
Iglesias-Soler E, Carballeira E, Sanchez-Otero T, Mayo X, Jimenez A, Chapman ML. Acute effects of distribution of rest between repetitions. Int J Sports Med. 2012;33:351–8.
Mayo X, Iglesias-Soler E, Fernandez-del-Olmo M. Effects of set configuration of resistance exercise on perceived exertion. Percept Mot Skills. 2014;119(3):1–13.
García-Ramos A, Padial P, Haff GG, Arguelles-Cienfuegos J, García-Ramos M, Conde-Pipo J, et al. Effect of different interrepetition rest periods on barbell velocity loss during the ballistic bench press exercise. J Strength Cond Res. 2015;29(9):2388–96.
García-Ramos A, González-Hernández JM, Baños-Pelgrín E, Castaño-Zambudio A, Capelo-Ramírez F, Boullosa D, et al. Mechanical and metabolic responses to traditional and cluster set configurations in the bench press exercise. J Strength Cond Res. 2017. https://doi.org/10.1519/JSC.0000000000002301(Epub 2017 Oct 20).
Nickerson BS, Williams TD, Snarr RL, Park K-S. Individual and combined effect of inter-repetition rest and elastic bands on jumping potentiation in resistance-trained men. J Strength Cond Res. 2019;33(8):2087–93. https://doi.org/10.1519/JSC.00000000000002593.
Vernillo G, Temesi J, Martin M, Millet GY. Mechanisms of fatigue and recovery in upper versus lower limbs in men. Med Sci Sports Exerc. 2018;50(2):334–43.
Walker S, Davis L, Avela J, Hakkinen K. Neuromuscular fatigue during dynamic maximal strength and hypertrophic resistance loadings. J Electromyogr Kinesiol. 2012;22(3):356–62.
Todd G, Taylor JL, Gandevia SC. Measurement of voluntary activation of fresh and fatigued human muscles using transcranial magnetic stimulation. J Physiol. 2003;551(Pt 2):661–71.
Kent-Braun JA. Central and peripheral contributions to muscle fatigue in humans during sustained maximal effort. Eur J Appl Physiol Occup Physiol. 1999;80(1):57–63.
Finsterer J. Biomarkers of peripheral muscle fatigue during exercise. BMC Musculskelet Disord. 2012;13:218.
Gorostiaga EM, Navarro-Amezqueta I, Cusso R, Hellsten Y, Calbet JAL, Guerrero M, et al. Anaerobic energy expenditure and mechanical efficiency during exhaustive leg press exercise. PLoS One. 2010;5(10):e13486.
Gorostiaga EM, Navarro-Amézqueta I, Calbet JA, Sánchez-Medina L, Cusso R, Guerrero M, et al. Blood ammonia and lactate as markers of muscle metabolites during leg press exercise. J Strength Cond Res. 2014;28(10):2775–85.
Gorostiaga EM, Navarro-Amézqueta I, Calbet JA, Hellsten Y, Cusso R, Guerrero M, et al. Energy metabolism during repeated sets of leg press exercise leading to failure or not. PLoS One. 2012;7(7):e40621.
Hunter SK. The relevance of sex differences in performance fatigability. Med Sci Sports Exerc. 2016;48(11):2247–56.
Hunter SK. Sex differences in human fatigability: mechanisms and insight to physiological responses. Act Physiol. 2014;210(4):768–89.
Bazzucchi I, Marchetti M, Rosponi A, Fattorini L, Castellano V, Sbriccoli P, et al. Differences in the force/endurance relationship between young and older men. Eur J Appl Physiol. 2005;93(4):390–7.
Ramirez-Campillo R, Alvarez C, Garcìa-Hermoso A, Celis-Morales C, Ramirez-Velez R, Gentil P, et al. High-speed resistance training in elderly women: effects of cluster training sets on functional performance and quality of life. Exp Gerontol. 2018;110:216–22.
Conflict of interest
Christopher Latella, Wei-Peng Teo, Eric J. Drinkwater, Kristina Kendall and G. Gregory Haff declare that they have no conflict of interest.
No financial support was received for the conduct of this systematic review or for the preparation of this manuscript.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Latella, C., Teo, WP., Drinkwater, E.J. et al. The Acute Neuromuscular Responses to Cluster Set Resistance Training: A Systematic Review and Meta-Analysis. Sports Med 49, 1861–1877 (2019). https://doi.org/10.1007/s40279-019-01172-z