Effects of Continuous and Cluster‐Set Configurations Performed to Failure on Performance and Physiological Responses in the Bench Press and Half Squat

ABSTRACT

This study compared the acute effects of continuous execution (CON) and two cluster‐set configurations performed to momentary failure with high loads (~85% one‐repetition maximum) on mechanical and physiological responses during the bench press and half squat. Twelve recreationally trained men completed three randomized, counterbalanced conditions: two using cluster sets of three repetitions with either 10 s (CS10) or 20 s (CS20) intra‐set rest, and a control (CON) condition. In the bench press, CS10 did not significantly increase the number of repetitions compared with CON, while CS20 increased repetitions from 6.8 ± 1.6 to 11.4 ± 2.9 (p = 0.044). In the half‐squat, the effect of cluster sets on repetitions was evident in both CS10 and CS20 compared with CON (18.3 ± 4.0 and 42.1 ± 7.7 vs. 10.6 ± 3.2 repetitions, respectively, p < 0.001). Set duration increased progressively with longer intra‐set rest, particularly in the half squat, reaching in CS20, 90.7 ± 26.2 s and 344.6 ± 65.3 s (p < 0.001) in the bench press and half squat exercises, respectively. Mean heart rate (HR) in CS20 was higher in the half squat than the bench press (87.2% ± 4.9% vs. 70.4% ± 6.5% of max HR, p < 0.001), while blood lactate responses were also greater in half‐squat compared with bench press (~8.5 vs. 5.5 mmol L⁻¹, p < 0.001), with values in CS20 reaching 10.8 ± 4.5 mmol L⁻¹. In summary, when sets are performed to momentary failure with high loads, CS10 provides only modest performance benefits, whereas CS20 markedly increases training volume, particularly in the half squat. The elevated cardiovascular and metabolic stress observed under CS20 reflects the substantially greater total work performed rather than an inherent effect of the cluster configuration.

1. Introduction

During resistance training, strength and conditioning professionals frequently manipulate set configurations to achieve specific training objectives [1, 2]. Cluster sets (CS) incorporate brief, pre‐planned intra‐set rest intervals inserted between individual repetitions or small groups of repetitions, thereby enhancing repetition performance and attenuating metabolic stress [2, 3, 4]. For example, a set of 8 repetitions performed in a continuous manner‐where repetitions are executed consecutively without planned intra‐set rest‐ can be reorganized into four CSs of two repetitions with 10 s of rest between clusters [2]. In contrast to sets performed with consecutive repetitions, in which accumulating fatigue may negatively influence kinematic and kinetic variables [5, 6] with velocity loss emerging as a practical marker of neuromuscular fatigue and metabolic stress [7], CS configurations have been shown to increase both repetition quantity and repetition quality (i.e., higher movement velocity and power output), particularly under high load conditions [8, 9, 10, 11]. Additionally, CSs may contribute to the maintenance of exercise technique [12, 13, 14] and reduce internal load [10, 15], as reflected by lower ratings of perceived exertion. This is particularly relevant in sports such as weightlifting, CrossFit, and powerlifting, where technical consistency and fatigue management are critical for performance and safety [16, 17, 18].

The increased number of repetitions during CS configurations is likely underpinned by both mechanical and physiological mechanisms. From a bioenergetic perspective, brief intra‐set rest intervals facilitate partial restoration of phosphocreatine, thereby allowing high ATP utilization rates during subsequent repetitions [19, 20, 21]. Short intra‐set intervals between clusters (e.g., 10–30 s) may therefore exploit this rapid recovery component and delay the onset of metabolic perturbations [21]. In addition to preserving high‐energy phosphate availability, reduced accumulation of hydrogen ions and inorganic phosphate may attenuate impairments in excitation–contraction coupling and cross‐bridge cycling, thereby sustaining force and velocity output [22]. Consequently, CS configurations may represent an integrated mechanical and a metabolic strategy that enhances neuromuscular performance and improves energy efficiency during high‐load resistance exercise [23, 24]. Consistent with this mechanistic framework, previous studies report that CSs permit a greater number of repetitions and higher total training volume compared with continuous sets [25, 26, 27], with one investigation demonstrating a 9.3‐fold increase in repetitions under a CS configuration [26]. Such improvements in volume capacity may translate into comparable or potentially enhanced hypertrophic adaptations relative to traditional set structures [28]. Furthermore, CS configurations have been shown to attenuate glycolytic stress, as reflected by lower blood lactate concentrations when total repetition number is matched compared with that of continuous repetition execution [29, 30, 31].

Several investigations have examined cluster‐set configurations under volume‐equated conditions to isolate the effects of set structure on physiological and mechanical responses. Ribeiro‐Torres et al. demonstrated that resistance training performed with inter‐repetition rests reduced cardiovascular stress compared with traditional resistance training despite equivalent external work [32]. García‐Ramos et al. reported that cluster configurations with longer intra‐set rest intervals preserved movement velocity and reduced blood lactate accumulation compared with traditional set structures during the bench press exercise [33]. Similarly, González‐Hernández et al. observed reduced mechanical and metabolic fatigue during cluster‐set configurations in the full squat exercise [34]. Collectively, these studies suggest that when total work is controlled, cluster‐set strategies may attenuate physiological stress and preserve mechanical performance. However, whether these advantages persist when cluster configurations are performed to volitional failure—thereby inherently allowing differences in total work and repetition volume—remains unclear.

In general, studies report 35%–50% lower blood lactate concentrations in CS configurations compared with traditional continuous sets when total repetition number matched [29, 30, 31]. However, it remains unclear whether this metabolic advantage persists when CS is performed to volitional failure, a condition characterized by substantial metabolic accumulation and neuromuscular fatigue. Under failure conditions, the extended duration of effort and increased reliance on glycolytic metabolism may attenuate the presumed metabolic benefits of intra‐set rest intervals [22].

While the majority of previous investigations have examined cluster‐set configurations using volume‐matched or submaximal protocols, comparatively little is known about the acute responses when cluster sets are performed to momentary failure. This distinction is practically important for several reasons. First, training to failure remains a common practice in strength and conditioning, particularly among athletes seeking to maximize hypertrophic or strength adaptations within time‐constrained sessions. Second, understanding the upper limits of volume capacity under different cluster configurations provides practitioners with essential information for programming decisions. Third, the interaction between intra‐set rest duration and fatigue accumulation under maximal effort conditions may differ substantially from responses observed in submaximal, volume‐equated protocols. Although cluster training is often implemented to manage fatigue and maintain movement quality during submaximal training, there are practical scenarios in which athletes and practitioners may choose to train to momentary failure. These include: (1) hypertrophy‐focused training phases where metabolic stress and mechanical tension are deliberately maximized; (2) testing or assessment contexts where maximal repetition capacity is evaluated; (3) time‐constrained training sessions where maximizing volume within a single set is prioritized; and (4) sports such as CrossFit and certain powerlifting methodologies where training to failure is a programmatic feature. Understanding how cluster configurations interact with failure‐based training provides practitioners with evidence to inform programming decisions across these varied contexts.

The present study deliberately employed sets performed to momentary failure without volume matching between conditions. This design choice was made to address a specific question that volume‐equated protocols cannot answer: to what extent can different cluster‐set configurations extend repetition capacity when athletes train to failure, and what are the associated physiological costs of this extended effort? While volume‐matched designs are essential for isolating the effects of set structure on acute responses, they do not capture the real‐world scenario in which cluster sets are used precisely because they allow greater volume accumulation. Thus, our approach provides complementary information regarding the practical upper limits of volume capacity and the physiological demands of training to failure under different rest configurations.

With respect to cardiovascular responses, the available evidence suggests that CS configurations impose distinct hemodynamic demands compared with continuous sets. Although the incorporation of brief intra‐set rest intervals is typically associated with lower perceptual and metabolic stress, heart rate (HR) responses appear to depend on the specific CS structure and population studied. For example, a recent study reported that, CS configurations consisting of single‐ or double‐ repetition clusters during the back squat elicited higher HR responses than continuous sets in trained older men (~67 years), despite lower ratings of perceived exertion and reduced fatigue [35]. However, given well‐established age‐related differences in cardiovascular regulation, autonomic function, and neuromuscular recovery capacity, the transferability of these findings to younger populations requires empirical verification. Nevertheless, these findings indicate that intra‐set rest does not necessarily reduce cardiovascular stress but may, under certain conditions, augment cardiac demand, potentially due to the maintenance of higher velocities and power outputs across repetitions. Whether similar cardiovascular responses occur in young, recreationally trained individuals during multi‐exercise protocols performed to volitional failure remains to be determined.

The intra‐set rest intervals of 10 and 20 s were selected based on the known time course of phosphocreatine resynthesis, which exhibits a rapid initial component within the first 10–20 s of recovery. These durations represent practically relevant intervals that are commonly employed in cluster‐set training and allow examination of whether minimal (10 s) versus moderate (20 s) recovery periods differentially affect performance sustainability under high‐load failure conditions.

Therefore, the purpose of the present study was to examine the acute effects of continuous (CON) and two CS configurations incorporating 10 and 20 s of intra‐set rest, performed with high loads (~85% of one‐repetition maximum; 1‐RM) on movement velocity, HR, and metabolic responses, as reflected by blood lactate concentration. In addition, these responses were compared between a representative upper‐body exercise (bench press) and a lower‐body exercise (half‐squat), both performed on a Smith‐machine.

Based on the known time course of phosphocreatine resynthesis and previous findings from volume‐matched studies, we hypothesized that: (1) both cluster‐set configurations would enable a greater number of repetitions compared with continuous execution, with CS20 producing larger increases than CS10; (2) cluster configurations would attenuate velocity loss across the set compared with continuous execution; (3) the effects of cluster sets on repetition capacity would be more pronounced in the half squat than the bench press due to the greater muscle mass and metabolic reserve of the lower extremities; and (4) physiological responses (heart rate and blood lactate) would increase in proportion to total work performed, with CS20 eliciting the highest responses when performed to failure.

2. Materials and Methods

2.1. Participants

Twelve recreationally trained men took part in the study (age: 23.7 ± 4.1 years, height: 1.75 ± 0.06 m, body mass: 78.4 ± 8.8 kg, body fat: 10.1% ± 3.7%, 1‐RM bench press strength: 96.2 ± 16.9 kg or 1.23 ± 0.17 kg kg⁻¹ body mass, 1‐RM half squat strength: 181.9 ± 23.6 kg or 2.33 ± 0.30 kg kg⁻¹ body mass). The participants were involved in team and individual sports (soccer, basketball and cross‐fit) and had strength training experience of at least 3 years. As an inclusion criterion, participants were required to demonstrate a minimum relative strength of at least 1.0 times body mass in the bench press and 1.5 times body mass in the half‐squat exercise, ensuring and adequate strength level. Furthermore, participants reported no use of medications or nutritional supplements and had been free from musculoskeletal injuries for at least 1 year prior to the study. Each participant provided written informed consent after receiving a detailed explanation of the experimental procedures, the potential risks involved, and their right to withdraw from the study at any time without penalty. The study was approved by the local Institutional Review Board (Approval no. 1779/20‐03‐2025), and all procedures were in conducted in accordance with the ethical standards of the World Medical Association as outlined in the Declaration of Helsinki of (1964, as revised in 2024).

Although an a priori power analysis was not conducted, a sensitivity analysis was performed using G*Power (version 3.1.9.7) to determine the minimum detectable effect size given the final sample. For a repeated‐measures ANOVA with three conditions, 12 participants, α = 0.05, power = 0.80, and a conservative correlation among repeated measures of r = 0.5, the analysis indicated that the study was adequately powered to detect effect sizes of f ≥ 0.35 (corresponding to η ² ≥ 0.11, i.e., medium‐to‐large effects). Given that the primary outcomes of interest (repetition number, heart rate, blood lactate) yielded large effect sizes (η ² ranging from 0.33 to 0.93), the sample size was sufficient to detect the observed differences. Nevertheless, we acknowledge that the sample size may have limited statistical power for detecting smaller effects.

2.2. Design and Procedures

A repeated measures design was used to compare the acute effects of CON and CS configurations performed with high loads (~85% of one‐repetition maximum; 1‐RM) on movement velocity, heart rate, and metabolic responses, as reflected by blood lactate concentration, during the bench press and the half‐squat exercises. A schematic representation of the experimental design is shown in Figure 1. Participants completed two preliminary sessions and three experimental conditions, separated by 5–7 days. Each condition was performed on a Smith‐machine and included the bench press and the half‐squat exercise to volitional failure. The load used was 84.9% ± 0.8% for bench press and 84.0% ± 2.8% for half squat. Condition order was randomized and counterbalanced, while exercise order was fixed (bench press followed by half‐squat after 20 min rest). Exercise order was fixed, with the bench press always performed before the half squat. This decision was made for several reasons: (1) to prioritize participant safety by performing the more systemically demanding lower‐body exercise after participants were thoroughly prepared and warmed; (2) to ensure consistency across all experimental conditions, thereby reducing within‐subject variability in exercise sequencing; and (3) to reflect common training practice where upper‐body exercises often precede lower‐body exercises within a session. However, we acknowledge that this fixed order introduces potential carryover effects, and comparisons between exercises should be interpreted with this caveat in mind (see Limitations). Importantly, the present study did not employ volume matching between conditions. This design permits cluster‐set configurations to express their full volume‐expanding potential but means that observed differences in physiological responses cannot be attributed to set structure independent of total work.

FIGURE 1.

Schematic overview of the experimental protocol for the bench press and half squat. CON: continuous set to momentary failure without intra‐set rest. CS10, clusters of three repetitions with 10‐s intra‐set rest to failure; CS20, clusters of three repetitions with 20‐s intra‐set rest to failure.

In both exercises, repetitions were performed with maximal intended movement velocity (movement tempo X:0:X:0) and without an imposed pause at the lower position. During the half squat, squat depth was standardized using a mechanical stop (safety pins positioned at ~85° knee flexion), and the load used during the experimental sessions corresponded to ~85% of Smith‐machine half‐squat 1‐RM determined under the same movement constraints. The eccentric phase was immediately followed by concentric execution.

The experimental conditions were as follows:

1. a CON set to momentary failure: the participants performed continuous repetitions to volitional failure.

2. a CS with 10‐s intra‐set rest intervals (CS10): participants performed repeated clusters of three repetitions to volitional failure, with 10 s of rest between clusters.

3. Cluster set with 20‐s intra‐set rest intervals (CS20): participants performed repeated clusters of three repetitions to volitional failure, with 20 s of rest between clusters.

The dependent variables were the total number of repetitions, mean bar velocity (MV), time under tension (TUT), set duration (SET_tot), percentage of velocity loss (%VL), blood lactate concentration (BL), mean relative HR (HR_mean), and peak relative heart rate (HR_peak).

2.3. Preliminary Sessions

The participants attended two preliminary sessions. During the first preliminary visit, anthropometric data were collected and maximum strength (1‐RM) was assessed for both exercises, with a standardized 30‐min rest interval between the bench press and half‐squat tests. During the second preliminary session, participants were familiarized with the experimental protocols, including the bench press and the half‐squat exercises. All repetitions during familiarization were performed with maximum intended velocity. Following a standardized general and exercise‐specific warm‐up, participants were instructed to move the barbell as fast as possible from the very first repetition. For the bench press exercise, participants performed one set of 8 repetitions at 40% of 1‐RM, one set of 6 repetitions at 65% of 1‐RM, and three sets of 3 repetitions at 85% of 1‐RM with 5 min of rest between sets. After a 20‐min rest interval, the same procedure was repeated for the half‐squat exercise. Participants first completed a lower‐limb‐specific warm‐up, followed by one set of 8 repetitions at 40% of 1‐RM, one set of 6 repetitions at 65% of 1‐RM, and three sets of 3 repetitions at 85% of 1‐RM again with 5 min of recovery between sets.

2.4. General and Specific Warm‐Up During the Experimental Conditions

Prior to all experimental conditions, participants completed a standardized general warm‐up consisting of 5 min of light jogging on a treadmill (7–8 km h⁻¹), followed by and 5 min of dynamic stretching targeting the upper‐body musculature [36]. Following the general warm‐up, a bench press specific warm‐up was performed, which consisted of one set of 8 repetitions at 40% of 1‐RM and, one set of 6 repetitions at 65% of 1‐RM. Upon completion of the bench press exercise under each experimental condition, participants rested for 5 min and subsequently walked on a treadmill for 5 min at 5 km h⁻¹. To prepare for the half‐squat, participants then performed a lower body specific warm‐up, consisting of: (1) a series of bodyweight drills including 8 repetitions each of the following exercises: bodyweight lunges, bodyweight squats, leg swings in hip flexion‐extension, leg swings in hip adduction–abduction, and single‐leg deadlifts, followed by (2) a half‐squat specific loaded warm‐up consisting of 1 set of 8 repetitions at 40% of 1RM and 1 set of 6 repetitions at 65% of 1RM, as shown in Figure 1.

2.5. Definition of Momentary Failure and Technique Monitoring

Momentary failure was operationally defined as the point at which the participant could no longer complete a repetition through the full range of motion despite maximal voluntary effort or when the supervising investigators determined that technique had deteriorated to a degree that would compromise safety. Two experienced strength and conditioning coaches supervised all testing sessions, providing standardized verbal encouragement and monitoring movement quality throughout each set.

For the half squat, squat depth was strictly controlled using a mechanical stop (safety pins positioned to correspond to ~85° knee flexion), which the barbell was required to contact on each repetition before initiating the concentric phase. Repetitions that did not achieve the required depth or that exhibited substantial technical breakdown (e.g., excessive forward lean, knee valgus, or loss of neutral spine position) were not counted toward the total. No repetitions were retrospectively excluded from analysis; rather, real‐time monitoring ensured that only valid repetitions were recorded.

2.6. Measurements

2.6.1. Anthropometric Measurements

During the first preliminary visit, anthropometric assessments were conducted. Body height was measured to the nearest 0.1 cm using a portable stadiometer (Charder HM‐200P Portstad). Body mass was measured to the nearest 0.1 kg using a calibrated digital scale (TBF‐300A Body Composition Analyzer‐Tanita), and body fat percentage was estimated from seven skinfold thicknesses using a Harpenden skinfold caliper (British Indicators Ltd., Herts, England).

2.6.2. Maximum Dynamic Strength (1‐RM) in the Bench Press and in the Half‐Squat

Maximum dynamic strength (one‐repetition maximum; 1‐RM) in the bench press and half squat was assessed on a Smith machine, in accordance with the procedures recommended by the National Strength and Conditioning Association (NSCA) [37]. During the bench press exercise, participants lay supine on the bench with the head, shoulders, and hips supported, while both feet were placed flat on the floor with the knees flexed at approximately 90°. Participants used a closed, pronated grip slightly wider than shoulder‐width. Two experienced strength & conditioning coaches served as spotters, providing verbal encouragement and ensuring participant safety throughout the testing procedure. The intraclass coefficient correlation (ICC) for the bench press 1‐RM assessment in our laboratory is 0.92 [36].

During the half squat, participants began from an upright position with the knees and hips fully extended and a stance approximately shoulder‐width apart, with both feet flat on the floor. The barbell was positioned across the upper back at the level of the acromion. Stance width and foot placement were individually adjusted and consistently replicated across all lifts. Participants were instructed not to leave the ground at any point, and the barbell was required to remain in continuous contact with the back and shoulders. From the starting position, participants descended until the barbell contacted a squat stand positioned to standardize squat depth (knee angle: 84.7° ± 3.5°). Immediately upon contact, participants performed the upward (concentric) movement without pausing at the bottom position. The intraclass coefficient correlation (ICC) for the half‐squat 1‐RM assessment in our laboratory is 0.99.

2.6.3. Mean Barbell Velocity

Mean barbell velocity was measured using a linear position transducer (Tendo Power Analyzer System v. 314, TENDO Sports Machines, Trencin, Slovak Republic) attached vertically to the barbell. To ensure accurate assessment of vertical displacement, the position of the transducer was standardized prior to each experimental condition by suspending a small weight vertically from the barbell to the floor, thereby minimizing the influence of any horizontal movement on velocity measurements. Mean barbell velocity was recorded for each repetition across all experimental conditions in both exercises. The intraclass correlation coefficient (ICC) for mean velocity in our laboratory is 0.983 [36]. Velocity loss was also calculated as the percentage reduction in mean barbell velocity from the fastest repetition of each set to the last repetition [5].

2.6.4. Blood Lactate Concentration and Heart Rate

Blood lactate concentration was measured in a capillary blood sample obtained from the fingertip, using a portable blood lactate analyzer (Lactate Scout+, EKF Diagnostics, Cardiff, UK). Measurements were taken at the following time points: (1) before the general warm‐up with participants seated quietly for 5 min (pre‐bench press baseline); (2) 30 s after completion of the bench press set; (3) 3 min after completion of the bench press set; (4) immediately before the half‐squat specific warm‐up (pre‐half squat); (5) 30 s after completion of the half squat set; and (6) 3 min after completion of the half squat set. The pre‐half squat measurement was taken to characterize the physiological baseline prior to the second exercise, which differed from the pre‐bench press baseline due to the preceding exercise bout. Figure 5 displays the pre‐exercise, 30‐s post‐exercise, and 3‐min post‐exercise values for each exercise separately.

FIGURE 5.

Blood lactate concentrations measured before (PRE BP and PRE SQ) and 0.5 and 3 min after the bench press and half‐squat exercises under control (CON), cluster‐10 (CS10), and cluster‐20 (CS20) conditions. PRE BP and PRE SQ represent the pre‐exercise blood lactate concentrations measured immediately before the bench press and half‐squat exercises, respectively. Values are presented as mean ± SD. *p < 0.01 compared with PRE within the same exercise across all conditions; ^† p < 0.01 compared with the bench press at the corresponding time point; ^# p < 0.01 CS20 compared with CON and CS10 at the corresponding time point.

Heart rate was continuously monitored throughout each experimental session using a chest‐strap heart rate sensor (Polar H10; Polar Electro Oy, Kempele, Finland) and recorded via the Polar Beat mobile application (Polar Electro Oy, Kempele, Finland). Mean heart rate (HR_mean) and peak heart rate (HR_peak) during each exercise set were calculated as a percentage of maximal predicted heart rate.

2.6.5. Time Under Tension and Exercise Set Duration

Time under tension (TUT) was defined as the effective lifting time during the concentric and eccentric phases of each repetition, excluding all rest periods. In addition, the exercise set duration (SET_tot) was calculated as the elapsed time from the initiation of the first repetition to the completion of the final repetition.

2.6.6. Statistical Analyses

Statistical analyses were performed using the SPSS Statistics (Version 23; IBM Corp., Armonk, NY, USA). Data are presented as mean ± standard deviation (SD). Tο examine differences in repetitions, MV, %VL, TUT, SET_tot, HR_mean and HR_peak, a two‐way repeated‐measures analysis of variance (ANOVA) was conducted with exercise (bench press and half‐squat) and condition (CON, CS10, CS20) as within‐subject factors. BL responses were examined using a three‐way repeated‐measures analysis ANOVA, with condition (CON, CS10, CS20), exercise (bench press, half‐squat) and time (three time points) as within‐subject factors. When a significant main effect or interaction was observed, post–hoc pairwise comparisons were conducted using the Bonferroni correction to control for familywise error rate. Effect sizes for main effects and interactions were calculated using partial eta squared (η ²) and interpreted as small (0.01 to 0.059), moderate (0.06 to 0.137), or large (> 0.137). For pairwise comparisons, effect sizes were calculated using Hedges' g and interpreted as small (< 0.3), moderate (0.3–0.8), or large (> 0.8). Exploratory associations between performance and physiological variables were examined using repeated‐measures correlation analyses (rmcorr), which account for the non‐independence of repeated observations within individuals. Repeated‐measures correlations were performed to examine associations among repetition number, time under tension (TUT), mean relative heart rate (HRmean), and post‐exercise blood lactate concentration (BL). Repeated‐measures correlation coefficients (rrm), 95% confidence intervals (CI), and corresponding p‐values were reported.

Statistical significance was set at p < 0.05.

3. Results

3.1. Total Number of Repetitions

A significant exercise × condition interaction was observed for the total number of repetitions (p < 0.001, η ² = 0.91). Post hoc analysis showed that for the bench press, only CS20 resulted in a greater number of repetitions compared with CON (11.4 ± 2.9 vs. 6.8 ± 1.6 repetitions, p = 0.044, g = 1.88), with no significant difference between CS20 and CS10 (p = 0.46) or between CS10 and CON (p = 1.0). In contrast, for the half squat exercise, the effect of cluster sets was evident in both CS10 and CS20, compared with CON (18.3 ± 4.0 and 42.1 ± 7.7 vs. 10.6 ± 3.2 repetitions, respectively, g = 2.1 and 5.1, p < 0.001; Figure 2), with CS20 performance being approximately 130% higher than CS10 (g = 3.7, p < 0.001). No statistically significant difference was detected between the bench press and half squat exercise under the CON condition (p = 0.18), despite a large effect size (g = 1.4). However, the number of repetitions performed in CS10 and CS20 was greater in the half squat compared with bench press (g = 3.2–5.1, p < 0.001; Figure 2).

FIGURE 2.

Number of repetitions performed during the bench press and the half squat Smith squat under traditional (CON), cluster‐10 (CS10), and cluster‐20 (CS20) conditions. Values are presented as mean ± SD. *p < 0.05 compared with CON within the same exercise; **p < 0.05 compared with CON and CS10 within the same exercise; ^† p < 0.05 compared with the bench press within the corresponding condition.

3.2. Mean Barbell Velocity and Velocity Loss

Α two‐way repeated‐measures ANOVA revealed a significant exercise × condition interaction for MV (p < 0.001, η ² = 0.47). Post hoc analyses indicated that, during the bench press, MV was significantly higher under the CON condition compared with CS20 (0.30 ± 0.05 m s⁻¹ vs. 0.27 ± 0.06 m s⁻¹, g = 0.54, p = 0.014, Figure 3) with no difference between CON and CS10 (p = 0.10) as well as between CS10 and CS20 (p = 1.0). For the half‐squat exercise, MV did not differ between conditions (p > 0.46). Comparisons between exercises showed that half‐squat MV was consistently higher than bench press MV across conditions (g = 0.93–1.8, p < 0.001).

FIGURE 3.

Mean barbell velocity (MV) during the bench press and the half‐squat under traditional (CON), cluster‐10 (CS10), and cluster‐20 (CS20) conditions. Values are presented as mean ± SD. *p < 0.05 compared with CON within the same exercise; ^† p < 0.05 compared with the bench press within the corresponding condition.

No exercise × condition interaction was observed for velocity loss (p = 0.78, η ² = 0.02). However, significant main effects were detected for exercise (p < 0.001, η ² = 0.95) and condition (p < 0.001, η ² = 0.50). Bonferroni‐corrected post hoc tests showed greater velocity loss during the bench press compared with the half‐squat exercise (61.9% ± 9.3% vs. 44.3% ± 7.9%, g = 1.97, p < 0.001). Additionally, the CS20 condition resulted in greater velocity loss compared with CON and CS10 (CS20: 58.1% ± 12.4% vs. CON: 49.0% ± 11.1%, g = 0.75, p < 0.01, CS10: 52.2% ± 12.1%, g = 0.46, p < 0.05). No other significant differences were observed.

3.3. Time Under Tension and Set Duration

Α two‐way repeated‐measures ANOVA revealed a significant exercise × condition interaction for TUT (p < 0.001, η ² = 0.89). Post hoc analyses indicated that, during the bench press, TUT was significantly greater under the CS20 condition compared with CON (27.4 ± 6.7 s vs. 16.1 ± 2.4 s, g = 2.16, p = 0.0023) and CS10 (19.1 ± 2.0, g = 1.63, p = 0.041) with no difference between CON and CS10 (p = 1.0). For the half‐squat exercise, TUT differed between all conditions (CS20 > CS10 > CON, g = 3.7–4.9, p < 0.001). Comparisons between exercises showed that TUT was greater during the half‐squat than bench press across CS10 and CS20 conditions (g = 4.0–4.9, p < 0.01). No difference was observed in CON condition (p = 1.0).

A significant exercise × condition interaction was observed for the SET_tot (p < 0.001, η ² = 0.93). Consequently, follow‐up simple effects analyses were conducted. For the bench press, a one‐way repeated‐measures ANOVA revealed significant differences between conditions (p < 0.001, η ² = 0.88), with CS20 resulting in a greater SET_tot compared with CON (90.7 ± 26.2 s vs. 17.7 ± 4.3 s, g = 3.8, p < 0.001) and CS10 (40.0 ± 5.9 s, g = 2.6, p < 0.001). A significant difference was also observed between CON and CS10 (g = 4.2, p < 0.001). For the half squat exercise, significant differences were observed between all conditions (g = 5.3–6.8, p < 0.001) with SET_tot being 21.5 ± 2.6, 85.9 ± 15.1 and 344.6 ± 65.3 s, in CON, CS10 and CS20, respectively. Significant differences between bench press and half‐squat were observed under the CS10 and CS20 (g = 3.9, p < 0.01 and g = 4.9, p < 0.001, respectively), whereas no statistically significant difference was detected under the CON condition (p = 1.0), despite a large effect size (g = 1.03).

3.4. Heart Rate Responses

Α two‐way repeated‐measures ANOVA revealed a significant exercise × condition interaction for HR_mean (p = 0.0011, η ² = 0.46). Post hoc analyses indicated that, during the bench press, HR_mean was significantly higher under the CS20 condition compared with CON (70.4% ± 6.5% vs. 66.2% ± 8.2%, g = 0.55, p = 0.008, Figure 4). The comparison between CS20 and CS10 (g = 0.48, p = 0.059) approached but did not reach statistical significance. No difference was observed between CON and CS10 (p = 0.95). For the half‐squat exercise, HR_mean differed between all conditions (CS20: 87.2% ± 4.9% > CS10: 83.4 ± 3.9 > CON: 77.2 ± 6.0, g = 0.84–1.75, p < 0.05, Figure 4). Comparisons between exercises showed that HR_mean was greater during the half‐squat than bench press across all conditions (g = 1.49–2.84, p < 0.001).

FIGURE 4.

Mean heart rate (HR_mean) expressed as a percentage of maximum HR. Values are presented as mean ± SD. *p < 0.05 compared with CON within the same exercise; **p < 0.05 compared with CON and CS10 within the same exercise; ^† p < 0.05 compared with the bench press within the corresponding condition.

No exercise × condition interaction was observed for HR_peak (p = 0.24, η ² = 0.12). However, significant main effects were detected for exercise (p < 0.001, η ² = 0.95) and condition (p < 0.001, η ² = 0.58). Bonferroni‐corrected post hoc tests showed greater HR_peak during the half squat compared with the bench press exercise (91.3% ± 4.6% vs. 77.2% ± 7.1%, g = 2.3, p < 0.001). Additionally, the CS20 condition resulted in greater HR_peak compared with CON (87.5% ± 7.4% vs. 81.9% ± 9.9%, g = 0.62, p < 0.01) and compared with CS10 (83.5% ± 9.7%, g = 0.45, p < 0.05). No other significant differences were observed.

3.5. Blood Lactate Concentration

No exercise × condition × time interaction was observed for blood lactate concentration (p = 0.13, η² = 0.14). However, there were significant condition × exercise (p = 0.011, η ² = 0.33) and exercise × time interactions (p < 0.01, η ² = 0.37). Post hoc analyses indicated that blood lactate concentration peaked immediately after both exercises and remained elevated 3 min later compared with baseline (g = from 1.75 to 4.1, p < 0.05; Figure 5). Blood lactate concentration was significantly higher after exercise in the half squat compared with the bench press across all conditions (~8.5 vs. 5.5 mmol L⁻¹, g = from 0.75 to 1.95, p < 0.01).

It should be noted that pre‐exercise blood lactate values before the half squat (2.1 ± 0.8 mmol L⁻¹) were elevated compared with values before the bench press (1.2 ± 0.4 mmol L⁻¹), indicating that the half squat did not commence from the same physiological baseline (g = 1.4). This difference reflects the fixed exercise order and should be considered when interpreting between‐exercise comparisons.

Between‐condition comparisons revealed that there was no difference in blood lactate concentration between conditions in the bench press (p = 0.14 to 1.00, Figure 5). In contrast, in the half squat exercise, blood lactate concentration was significantly higher in the CS20 condition compared with CON and CS10, reaching 10.8 ± 4.5 mmol L⁻¹ (g = from 0.68 to 1.17, p < 0.01; Figure 5).

3.6. Exploratory Repeated‐Measures Correlation Analysis

Repeated‐measures correlation analyses were conducted to examine associations between performance and physiological variables while accounting for within‐subject dependence across experimental conditions. Repetition number was strongly positively associated with mean relative heart rate (rrm = 0.78, 95% CI [0.661–0.864], p < 0.001) and moderately‐to‐strongly associated with post‐exercise blood lactate concentration (rrm = 0.67, 95% CI [0.505–0.790], p < 0.001). Time under tension was also strongly positively associated with mean relative heart rate (rrm = 0.73, 95% CI [0.588–0.830], p < 0.001). Furthermore, mean relative heart rate was strongly associated with post‐exercise blood lactate concentration (rrm = 0.73, 95% CI [0.586–0.830], p < 0.001). These findings suggest that conditions permitting greater work capacity were accompanied by greater cardiovascular and metabolic stress responses.

4. Discussion

The present study examined the acute effects of continuous execution and two cluster‐set configurations (10 and 20 s intra‐set rest) performed to momentary failure under high‐load conditions. The main findings were that in the bench press, CS10 did not significantly increase the number of repetitions compared with continuous execution (approximately one additional repetition on average, p = 0.45), and this was accompanied by no meaningful increase differences in movement velocity, heart rate, or blood lactate concentration. The longer inter‐set interval (CS20) resulted in a large increase in the number of repetitions in both exercises, but this occurred alongside prolonged set duration and elevated cardiovascular and metabolic responses. Interestingly, the CS10 was effective in increasing repetition number in the half‐squat exercise, suggesting that the benefits of cluster configurations are strongly dependent on intra‐set rest duration and exercise selection, and that increased repetition capacity also increases the physiological stress when sets are performed to failure.

It is important to note that the elevated heart rate and blood lactate concentrations observed under CS20, particularly during the half squat, should not be attributed to the cluster‐set configuration per se. Rather, these physiological responses are a direct consequence of the substantially greater number of repetitions and total work performed when longer intra‐set rest intervals enable continued exercise. The cluster structure permits greater volume accumulation, and it is this increased mechanical work that drives the heightened cardiovascular and metabolic demands.

The limited benefit observed with CS10 may be explained by the short duration of intra‐set recovery. Phosphocreatine resynthesis follows a biphasic time course, with a rapid initial component but incomplete restoration within the first 10–15 s [21, 38]. Accordingly, a 10‐s interval is unlikely to substantially restore high‐energy phosphate availability during repeated high‐load contractions. In addition, short recovery periods are unlikely to meaningfully reduce intramuscular accumulation of inorganic phosphate and hydrogen ions, both of which impair cross‐bridge function and excitation–contraction coupling [39]. Within this physiological context, the absence of a significant increase in repetition number with CS10 during the bench press appears consistent with limited metabolic recovery afforded by 10 s of rest. The absence of differences in velocity, heart rate, and lactate further supports the interpretation that short intra‐set rest redistributed effort without materially altering fatigue mechanisms under failure conditions.

In contrast, from a speculative mechanistic perspective, the longer recovery interval employed in CS20 may have allowed greater partial restoration of Phosphocreatine between clusters, thereby supporting force production and repetition sustainability, particularly during the half squat. However, it is also plausible that the prolonged set duration (~6 min) and sustained mechanical loading progressively increased intramuscular pressure, which may have partially constrained local blood flow during later stages of the set. While we did not directly measure intramuscular pressure or muscle oxygenation, previous research suggests that high intramuscular pressures during heavy resistance exercise can transiently restrict microvascular perfusion, particularly in large lower‐body muscles [40, 41]. These potential mechanisms remain hypothetical in the context of the present study and warrant direct investigation in future research.

Similarly, our observation of the pronounced cardiovascular responses during CS20, with heart rate values approaching ~90% of predicted maximal heart rate during the half squat, suggests substantial cardiovascular demand. While we speculate that this may reflect integrated central and peripheral fatigue contributions given the large active muscle mass involved, the present study did not include measures of cardiac output, stroke volume, or central fatigue markers. Therefore, the relative contributions of central versus peripheral mechanisms to the observed fatigue patterns remain to be elucidated. In this context, the cardiovascular strain observed during CS20 approaches values typically reported during moderate‐to‐high intensity aerobic exercise, suggesting a substantial cardiovascular demand. However, no direct measures of aerobic metabolism (e.g., oxygen uptake) were obtained, and therefore these observations reflect elevated cardiovascular and metabolic stress rather than direct evidence of an aerobic‐training stimulus. Accordingly, longer cluster‐set configurations performed with high loads may function as an integrated stimulus, simultaneously challenging neuromuscular performance while eliciting substantial cardiovascular and metabolic demands. Whether these acute responses translate into a meaningful aerobic training stimulus requires direct investigation. This perspective aligns conceptually with concurrent training models, in which resistance and endurance stimuli coexist within the same training session [42, 43]. However, whether such acute physiological responses translate into meaningful aerobic adaptations or influence long‐term strength development warrants further investigation.

Furthermore, the interpretation of the high repetition numbers observed during the half squat requires consideration of several exercise‐specific methodological characteristics. The high repetition capacity observed during the half squat should also be interpreted within the specific methodological context of the present protocol. The half squat involved a reduced range of motion standardized using a mechanical stop, and all repetitions were performed with maximal intended movement velocity. Previous research suggests that reduced range of motion and faster movement execution may influence repetition performance and fatigue responses compared with full‐range conditions [44, 45, 46].

The consistently greater repetition capacity, heart rate responses, and blood lactate concentrations observed during the half squat compared with the bench press likely reflect fundamental differences in muscle mass involvement and systemic physiological demand between upper‐ and lower‐body exercises [35, 42, 43]. The half squat engages the large muscle groups of the lower extremities (quadriceps, gluteals, hamstrings), which collectively represent a greater amount of actively engaged skeletal muscle mass than the primary movers of the bench press (pectoralis major, anterior deltoid, triceps) [47]. Greater active muscle mass is associated with higher oxygen consumption, cardiac output, and glycolytic demand during exercise, explaining the elevated cardiovascular and metabolic responses observed. Additionally, the greater amount of actively engaged musculature during lower‐body exercise may provide a greater capacity to sustain repeated efforts [48]. However, as noted in the limitations, the fixed exercise order (bench press preceding half squat) may have contributed to the observed between‐exercise differences through residual fatigue and elevated baseline lactate concentrations prior to the half squat. Therefore, while muscle mass differences likely contribute to the observed patterns, the magnitude of exercise‐specific effects may be confounded by order effects.

Our findings both align with and extend previous volume‐matched research. Ribeiro‐Torres et al. observed lower cardiovascular stress during resistance exercise with inter‐repetition rest periods, whereas García‐Ramos et al. and González‐Hernández et al. reported lower metabolic stress and reduced mechanical fatigue during cluster configurations under work‐equated conditions [32, 33, 34]. In contrast, the present study did not demonstrate a physiological advantage of cluster configurations when exercise was performed to failure. Instead, the CS20 condition resulted in greater heart rate and lactate responses during the half squat condition. This discrepancy likely reflects fundamental differences in study design, as volume‐matched approaches isolate the effects of rest distribution, whereas failure‐based protocols reveal the physiological consequences of the greater work capacity enabled by cluster configurations. Therefore, cluster strategies may reduce fatigue when work is controlled, but may simultaneously increase physiological stress when they allow greater training volumes to be achieved.

Previous investigations have reported 35%–50% lower blood lactate concentrations during cluster‐set protocols when total repetition number is matched between configurations [2, 26]. In the present study, sets were performed to momentary failure, allowing cluster configurations—particularly CS20—to substantially increase total work. Under these conditions, the greater repetition number enabled by intra‐set rest resulted in increased total work, which in turn elevated metabolic and cardiovascular responses. Thus, any transient metabolic recovery afforded by intra‐set rest appears to have been offset by the greater overall exercise volume. We hypothesize that the metabolic advantage often attributed to cluster training may be conditional upon volume equivalence rather than inherent to the structure itself. However, this interpretation requires confirmation through studies directly comparing volume‐matched continuous and cluster‐set protocols performed under similar high‐load conditions. The elevated physiological stress observed is therefore attributable to increased work capacity rather than to the cluster configuration itself.

We acknowledge that the present design, which permitted substantially different total work between conditions, does not allow us to isolate the effects of set configuration independent of volume. The elevated heart rate and blood lactate concentrations observed under CS20 are attributable to the greater number of repetitions performed rather than to the cluster structure itself. However, we contend that this design addresses a practically meaningful question: when the goal is to maximize volume at high loads, that is, how different cluster configurations compare in their capacity to extend performance, and what is the physiological cost. Volume‐matched comparisons, while methodologically important for mechanistic questions, do not capture this applied dimension. Our findings therefore complement the existing literature by quantifying the volume‐expanding potential of cluster sets and characterizing the associated cardiovascular and metabolic demands.

Exploratory repeated‐measures correlation analyses revealed strong associations between performance and physiological variables. Repetition number demonstrated a very strong positive association with time under tension (rrm = 0.98, 95% CI [0.973–0.990], p < 0.001) and was also positively associated with mean relative heart rate (rrm = 0.78, 95% CI [0.661–0.864], p < 0.001) and post‐exercise blood lactate concentration (rrm = 0.67, 95% CI [0.505–0.790], p < 0.001). Time under tension was positively associated with both mean relative heart rate (rrm = 0.73, 95% CI [0.588–0.830], p < 0.001) and post‐exercise blood lactate concentration (rrm = 0.66, 95% CI [0.485–0.780], p < 0.001). These findings suggest that the elevated heart rate and lactate concentrations observed under CS20 were primarily associated with the greater amount of work performed rather than the cluster configuration itself.

The exercise‐specific differences observed in repetition capacity and physiological responses further emphasize the interaction between set configuration and movement characteristics. The half squat consistently elicited a greater repetition number and higher cardiovascular responses than the bench press. This likely reflects differences in muscle mass involvement, mechanical demands, and systemic stress. Importantly, mean movement velocity did not differ substantially between conditions, suggesting that when sets are performed to failure, the inclusion of additional slower repetitions may offset any transient preservation of velocity within earlier clusters.

The observed differences between the bench press and half squat in repetition capacity, heart rate, and blood lactate responses should be interpreted with caution. While these differences are consistent with expected effects of muscle mass involvement, the fixed exercise order means that carryover effects from the bench press may have contributed to the higher physiological responses observed during the half squat. The elevated pre‐exercise lactate concentrations before the half squat confirm that participants did not begin this exercise from the same physiological baseline as the bench press. Consequently, we cannot definitively attribute the between‐exercise differences solely to inherent differences between upper‐ and lower‐body musculature.

Taken together, the present findings indicate that short intra‐set rest intervals (10 s) provide only limited enhancement of repetition capacity under high‐load failure conditions, whereas longer intervals (20 s) meaningfully increase volume but also increase cardiovascular and metabolic demands. Therefore, cluster‐set configurations should not be assumed to reduce physiological stress; rather, they modify the temporal distribution of fatigue and may increase total internal load when repetition number is expanded.

Several limitations should be acknowledged. A key design characteristic of this study was the absence of volume matching between conditions. While this approach was intentional—designed to quantify the volume‐expanding capacity of cluster sets under failure conditions—it limits our ability to isolate the effects of set configuration per se. The elevated cardiovascular and metabolic responses observed under CS20 reflect the substantially greater total work performed rather than an inherent effect of the cluster structure. Researchers and practitioners should interpret these findings within this context, recognizing that volume‐matched comparisons address different questions regarding the acute effects of rest redistribution.

A notable limitation of the present study is the fixed exercise order, with the bench press always preceding the half squat. Although a 20‐min rest interval was provided between exercises, residual cardiovascular and metabolic effects from the bench press may have influenced baseline values and subsequent responses during the half squat. Indeed, pre‐exercise blood lactate values were higher before the half squat than before the bench press, indicating that the two exercises did not commence from identical physiological baselines. Consequently, the observed differences between exercises in heart rate, blood lactate, and repetition capacity cannot be attributed solely to inherent differences between upper‐ and lower‐body musculature but may also reflect cumulative session effects. Future studies employing counterbalanced exercise orders or separate testing sessions for each exercise would help isolate exercise‐specific responses.

Additionally, the relatively small sample size (n = 12) may have limited statistical power to detect smaller effects, although sensitivity analysis confirmed adequate power for the large effects observed on the primary outcome variables.

4.1. Perspective

From a practical standpoint, the present findings indicate that the effectiveness of cluster‐set configurations depends on rest duration and exercise selection. Short intra‐set rest intervals (10 s) provide limited (half squat) or no enhancement (bench press) of repetition capacity when sets are performed to momentary failure at high loads. In contrast, longer intervals (20 s) substantially increase repetition number, particularly during lower‐body multi‐joint exercises such as the half squat, albeit at the cost of prolonged set duration and elevated cardiovascular and metabolic stress.

The practical implications of the present findings should be interpreted within the context of the study design, which employed sets performed to momentary failure. While this approach may not represent the most common application of cluster training, it provides valuable information for specific training contexts.

For practitioners implementing cluster training to failure, short intra‐set rest (10 s) offers minimal benefit for upper‐body exercises and only modest volume gains in lower‐body work, whereas longer rest (20 s) substantially increases volume capacity—particularly for lower‐body exercises—at the cost of extended set duration and elevated cardiovascular and metabolic demands, with implications for concurrent training and recovery. For those using cluster sets under submaximal, non‐failure conditions, the present findings delineate the upper limit of volume expansion these configurations can achieve, which may inform decisions around set termination criteria—such as velocity loss thresholds or target repetition numbers—that balance fatigue management with the performance benefits of intra‐set rest.

Practitioners should recognize that longer cluster intervals may shift the physiological profile of the exercise from predominantly local muscular fatigue toward a combined neuromuscular and cardiovascular demand. When the objective is to maximize volume at high loads, longer intra‐set rest intervals may be appropriate, especially for lower‐body exercises. Conversely, when tighter control of session duration or cardiovascular strain is required, shorter cluster intervals or continuous execution may be preferable. Notably, the pronounced cardiovascular responses observed during CS20 raise the possibility that certain cluster‐set configurations may impose a meaningful cardiovascular and metabolic stress when large muscle mass is engaged, although aerobic metabolism was not directly assessed in the present study. Thus, under specific conditions, cluster‐set training may provide an integrated mechanical and systemic challenge within a single session. However, whether such responses contribute to long‐term aerobic adaptations or influence concurrent training outcomes requires further investigation.

Funding

The authors have nothing to report.

Ethics Statement

Ethical approval for this study was obtained from the local Institutional Review Board (Approval No. 1779/20‐03‐2025). All procedures complied with the ethical standards of the Declaration of Helsinki.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data are available upon request.