Exploring acute physiological adaptations to circuit strength training: the role of work-to-rest interval variations

Abstract

We investigated the acute physiological responses to circuit strength training performed with different work-to-rest interval durations (10:10, 20:20, and 30:30 s) under a constant 1:1 ratio. Thirty-four trained adults (15 males and 19 females) completed a 14-min circuit protocol consisting of two blocks of 6-min separated by a 2-min rest period. Each session included six alternating upper- and lower-body exercises performed at maximal intended velocity. Heart rate (HR) was continuously monitored during the protocol, while blood lactate concentration, countermovement jump (CMJ) height, and 10-m sprint performance were assessed pre-, mid-, and post-circuit session. Repeated-measures ANOVA was used for CMJ height, blood lactate concentration and 10-m sprint performance, and one-way ANOVA for HR data. The 30:30 protocol elicited the highest post-exercise lactate concentrations (p < 0.05) and the largest decrements in CMJ height. In contrast, HR was consistently higher in the 10:10 protocol across several time points (p < 0.01), indicating greater cardiovascular demand. Sprint performance declined from pre- to post-exercise across all protocols, indicating a time-dependent fatigue response. These findings demonstrate that manipulating absolute work-to-rest durations, even under a fixed 1:1 ratio, produces distinct physiological responses, while performance decrements may occur independently of protocol configuration. Longer intervals (30:30) maximize metabolic stress and neuromuscular fatigue, whereas shorter intervals (10:10) enhance cardiovascular load. The 20:20 protocol represents a balanced alternative between both stimuli.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-34940-1.

Introduction

Resistance training (RT) has been widely recognized for its numerous benefits in terms of health^1–5 including improvements in cardiovascular health and reductions in chronic disease risk, particularly when high-intensity elements are incorporated⁶. However, a commonly reported barrier to practising RT is the lack of time, which, when combined with an increasingly sedentary lifestyle makes adherence to RT programmes challenging⁷. This issue is particularly evident in amateur and team sports settings, where limitations in infrastructure and limited available time complicate the implementation of these conventional RT programmes⁸. In response to these constraints, circuit training (CT) methods have gained popularity as an efficient alternative.

CT has been shown to be an excellent time-efficient strategy, by decreasing training time in 66% compared to traditional RT sessions⁹. CT typically integrates a sequence of strength stations performed consecutively with short rest intervals (15–30 s) and approximately 12–15 repetitions per station¹⁰. In distinction, traditional resistance training frequently involves accomplishing all sets of one exercise at a time employing rest intervals of moderate to extended duration between sets¹¹. In terms of absolute intensity prescription, traditional strength training usually handles heavier loads than circuit training¹²; however, high‑intensity circuit training represents a circuit training approach performed with higher loads¹³.

Building on these distinctions between CT and traditional RT, it is important to consider their differing physiological demands. From a bioenergetic perspective, blood lactate concentration reflects the metabolic stress imposed by resistance exercise and depends on the balance between production and clearance during repeated sets^14,15. Beyond its role as a metabolic by-product, lactate is now recognized as an important fuel and signalling molecule¹⁶. Elevated lactate levels have been associated with neuromuscular fatigue and performance decrements^17,18, such as reductions in CMJ height, highlighting its relevance for circuit training protocols.

Cardiovascular responses during RT are also influenced by work density and rest duration. While HR typically remains at moderate levels during traditional RT⁹, it becomes a more relevant indicator during CT due to the continuous nature of the workload and short inter-exercise rest periods^8,9,19. Shorter rest intervals limit parasympathetic reactivation, resulting in higher sustained HR responses throughout the session^20,21. Consequently, HR may represent a key variable for monitoring cardiovascular load during CT.

All these relations established previously are dominated by the rest periods. This variable is a determinant variable and plays a crucial role in modulating the physiological response to RT programmes²². Resistance training goals are based on the rest duration between sets or rest interval²³. Shorter inter-set rest periods have been associated with greater metabolic demand²¹, and an effect that is particularly accentuated in CT programmes¹⁹.

Thus, the work-to-rest ratio becomes a crucial factor in CT programmes design¹⁰, knowing that balance or imbalance among lactate production and clearance could be achieved adopting different sequences of work-to-rest ratios and intensities²⁴. Although no universal ratio has been established, the most used formats are 1:1 and 2:1 work-to-rest ratios⁹. However, the fixed ratios alone do not warrant the resolution of differing adaptations; a comprehensive application necessitates pairing them with defined work-to-rest intervals associated with each ratio. For instance, a 1:1 ratio may refer to 10 s of work followed by 10 s of rest or to 1 min of work followed by 1 min of rest, two markedly different stimuli that will likely lead to distinct physiological and neuromuscular adaptations. Thus, the effective application of work-to-rest ratios in CT programming requires not only choosing an appropriate proportion, but also clearly defining the absolute duration of both components.

Therefore, the aim of this study is to analyse the acute effects on HR, lactate concentration, CMJ, and 10 m running sprint of three distinct 1:1 work-to-rest ratio CTs that differ in work and rest interval duration (10:10, 20:20, and 30:30 s) for an equal total duration (14 min).

Based on the existing literature, we hypothesised that longer work intervals (30:30) would elicit greater metabolic stress, reflected by higher blood lactate concentrations, and larger decrements in neuromuscular performance (CMJ), due to greater repetition accumulation and fatigue development. Conversely, we hypothesised that shorter work intervals (10:10) would induce higher cardiovascular strain, manifested by elevated HR responses, particularly during the later stages of the protocol.

Methods

Participants

Thirty-four adult healthy individuals (15 males and 19 females), within 19 and 25 years of age, voluntarily agreed to participate in the study. The sample size was calculated using G*Power software (version 3.1.9.2, Kiel, Germany). In the preliminary analysis, it was determined that with an alpha of 0.05, power of 80% (1-β) and medium effect size (f = 0.5) 30 participants were required. Only individuals with a minimum of 6-months of continuous strength training and no injuries or disorders during this period were included in the study. None of the subjects was taking drugs, medications or dietary supplements known to influence physical performance. Furthermore, participants were instructed to maintain their regular dietary habits as well as to refrain from performing intense physical activity for 48 h approximately before each training session. Preceding enrolment, all participants were thoroughly informed of the experimental procedures and signed a written informed consent. Participants retained the right to terminate their participation at any stage of the study without adverse repercussions. The study was conducted in accordance with the Declaration of Helsinki (2024) and was approved by the Bioethics Committee of Alfonso X El Sabio University (ref. 2024_12/312).

Experimental design

A within-subject repeated-measures (crossover) experimental study was conducted. The implementation of three circuit-based resistance training protocols was involved, each one with a fixed total duration of 14 min. Nevertheless, each protocol differed in terms of work-to-rest intervals: 10:10 (10 s of work, 10 s of rest); 20:20 (20 s of work, 20 s of rest); and 30:30 (30 s of work, 30 s of rest). Each 14-min session was divided into two 6-min blocks separated by a 2-min active rest period (Fig. 1). The order in which the three circuit training protocols were administered was randomized and counterbalanced across participants to minimize potential carryover and learning effects inherent to repeated-measures designs.

Fig. 1.

Experimental design of the three circuit training protocols (10:10, 20:20, and 30:30), including timing of physiological and performance assessments.

Participants were required to execute the maximum number of repetitions possible at maximal intended velocity for the predetermined loads regarding each of the three distinct work-to-rest interval circuit training protocols. Repetitions were systematically recorded during each designated work period. Verbal encouragement was constantly provided to enhance maximal effort. The training circuit consisted of a total of six stations performed in the next sequence detailed: leg press, bench press, unilateral right knee extension, bilateral row, unilateral left knee extension, seated overhead press. The study was conducted over a 4-week data collection period towards the utilization of BH Fitness equipment (BH Fitness, Vitoria, Spain). Participants completed three separate sessions with a minimum intersession recovery period of > 48 h. All sessions and testing procedures were supervised by PhD students and PhD researchers in sport science with prior experience in resistance training and performance assessment. Verbal encouragement was standardized across sessions and delivered consistently by the same researchers, who also ensured correct movement execution and maximal intended velocity during all exercises.

Procedures

One week before the data collection phase began, a familiarization session was conducted to determine the individual loads for each protocol. Participants were instructed to perform as many repetitions as possible at maximum velocity within each protocol to determine the load lifted at muscle failure. This procedure was repeated for each station in the circuit. During the protocol sessions, the load used was one plate less than that employed in the familiarization session, thereby ensuring the ecological validity of the model and enabling participants to complete the protocol with a high level of effort without reaching muscle failure. Notably, shorter work-to-rest ratios resulted in higher absolute training load (e.g., 10:10 > 20:20 > 30:30).

All participants completed the same standardized warm-up prior to each testing session. The warm-up consisted of an 8-min treadmill run at a self-selected brisk walking speed, followed by ten repetitions each of shoulder flexion-extension, abduction-adduction, and circumduction exercises. Participants then performed eight progressive bodyweight squats, emphasizing movement beyond the sticking point. Finally, participants completed one familiarization lap of the circuit, performing six repetitions at each station with 50% of the prescribed training load for the intermediate protocol (20:20), with 30 s of rest between exercises.

Prior to the warm-up, blood lactate concentration was assessed using a fingertip capillary sample using a portable lactate analyser (Lactate Pro 2, Arkray, Kyoto, Japan). During the 2-min rest period between circuit blocks and immediately after completion of the full circuit, participants were escorted to the testing area. Capillary blood lactate samples were obtained first, while the jump measurement system was prepared. Jump height was measured pre and post intervention using an infrared timing system (Optojump®, Microgate, Bolzano, Italy). Once lactate sampling was completed and values were verified, participants performed two CMJ attempts and the highest executed jump was used for analysis, followed by the 10-m sprint test using an infrared timing system (Witty®, Microgate, Bolzano, Italy). This sequence of measurements was standardized and applied consistently across all sessions and protocols. HR was monitored continuously throughout the 14-min circuit (in 30-s intervals) using an optical sensor (Polar Verity Sense, Finland) placed on the arm of the participant.

Statistical analysis

Statistical analyses were performed with JASP statistical software (version 0.19.3.0) for Windows. In this study, the assumption of normality was verified using the Shapiro-Wilk and Levene test (p > 0.05). A 3 × 3 factorial ANOVA with repeated measures was used to evaluate the effects of the different work-to-rest intervals (30:30, 20:20 and 10:10) on blood lactate, CMJ and sprint values. Every single minute one-way ANOVA was analysed to determine the relationship between the HR and the distinct RI. When significant main effects or interactions were detected, post-hoc pairwise comparisons were conducted using the Bonferroni correction. This adjustment was applied to all multiple comparisons, including the minute-by-minute heart rate analyses, to control for inflated Type I error. Data variability is presented graphically using violin plots, which illustrate the full distribution of individual values and provide direct information on data dispersion without relying solely on summary statistics such as standard deviation or standard error. Effect sizes were calculated using Cohen’s D Scale (1988): small (d ≈ 0.20); medium (d ≈ 0.50); large (d ≥ 0.80). In addition, an exploratory Euclidean distance analysis was conducted to descriptively assess the similarity between protocol responses. Mean post-exercise responses for blood lactate concentration, CMJ performance loss, and sprint performance loss were standardized (z-scores) to account for differences in measurement units. Euclidean distances were then calculated between protocol centroids and visualized using a heatmap. All statistical significance was set at p < 0.05.

Cohen’s d was computed as the mean difference between paired conditions divided by the pooled standard deviation of the two measurements.

Results

For blood lactate concentration, a significant main effect of time (“Moment”) was observed (F (2, 66) = 1221.65, p < 0.001, η² = 0.830), together with a significant main effect of (“Group”) protocol (F (2, 66) = 3.26, p = 0.043, η² = 0.006) and a significant time × protocol (“Moment * Group”) interaction (F(4, 132) = 4.76, p = 0.001, η² = 0.006), indicating distinct lactate responses across protocols and measurement moments. For CMJ height, a significant main effect of time was detected (F (2, 66) = 55.86, p < 0.001, η² = 0.043), whereas neither the main effect of protocol (F (2, 66) = 0.04, p = 0.964) nor the interaction effect (F (4, 132) = 0.51, p = 0.729) reached statistical significance. Similarly, 10-m sprint performance showed a significant main effect of time (F (2, 66) = 13.84, p < 0.001, η² = 0.018), with no significant effects of protocol or time × protocol interaction (p > 0.05). These results are displayed in Table 1.

Table 1.

ANOVA results.

		df	F	p	η2
Lactate	Moment	2	1.221.650	< 0.001	0.830
	Group	2	3.257	0.043	0.006
	Moment * group	4	4.755	0.001	0.006
% decrease jump height	Group	2	0.729	0.485	0.015
CMJ	Moment	2	55.860	< 0.001	0.043
	Group	2	0.037	0.964	6.734 × 10− 4
	Moment * group	4	0.509	0.729	7.832 × 10− 4
10-m sprint	Moment	2	13.838	< 0.001	0.018
	Group	2	0.362	0.697	0.006
	Moment * group	4	0.885	0.474	0.002
% decrease sprint speed	Group	2	0.682	0.508	0.014

df = degrees of freedom; F = Fisher’s F ratio; p = probability value; η² = partial Eta squared; CMJ = countermovement jump.

The exploratory Euclidean distance analysis (Supplementary Fig. 1) revealed clear differences in overall response profiles between protocols. The 20:20 protocol exhibited comparable Euclidean distances to both the 10:10 and 30:30 protocols, indicating an intermediate physiological and neuromuscular response pattern. In contrast, the largest Euclidean distance was observed between the 10:10 and 30:30 protocols, reflecting the greatest dissimilarity in their overall response profiles.

Results from the blood lactate concentration measurements are presented in Fig. 2. Post-exercise blood lactate concentration was significantly higher in the 30:30 protocol compared with the 10:10 protocol (p = 0.037; Cohen’s d = 1.109). No other between-protocol differences were observed at this time point. Regardless of protocol, blood lactate levels at the pre-moment differed significantly from both the mid- and post-moments (p < 0.001). Interaction effects across protocols showed significant differences between mid and post moments in the 20:20 group (p = 0.003; Cohen’s d = 0.698) and in the 30:30 group (p < 0.001; Cohen’s d = 1.327) resulting in higher values for both groups on the post moment.

Fig. 2.

Blood lactate concentration measured pre-, mid-, and post-session across the three protocols. Violin plots represent data distribution and individual variability. Significance levels: p < 0.05; p < 0.01; p < 0.001.

HR responses during CT protocols are shown in Fig. 3. Significant differences between protocols were observed only at specific time points. During minutes 2–3, HR was significantly higher in the 30:30 protocol compared with the 10:10 protocol (p = 0.032). Conversely, during minutes 7–8 (p = 0.002), 8–9 (p = 0.004), 10–11 (p = 0.006), and 11–12 (p = 0.005), HR values were significantly higher in the 10:10 protocol compared with the 30:30 protocol. No significant differences were observed between protocols at the remaining time points.

Fig. 3.

Mean heart rate response across the 14-min circuit for each protocol. Symbols indicate significant between-protocol differences at specific time points (p < 0.05).

For CMJ height, the repeated-measures ANOVA revealed a significant main effect of time (p < 0.001), whereas neither the main effect of protocol nor the time × protocol interaction reached statistical significance (p > 0.05). Results of the CMJ jump height (left panel) and percentage of jump height loss (right panel) are presented in Fig. 4. Significant differences in CMJ jump height were found between pre and post moments in all conditions: 10:10 (p < 0.001; Cohen’s d = 0.432), 20:20 (p < 0.001; Cohen’s d = 0.514), and 30:30 (p < 0.001; Cohen’s d = 0.586) with the lowest values observed at post-moment. For the 20:20 group, significant differences were also found between mid and post moments (p = 0.037; Cohen’s d = 0.248) leading to higher outcomes regarding the mid moment. Furthermore, relevant differences were observed in the 30:30 group between pre and mid moments (p = 0.001; Cohen’s d = 0.355). Regarding CMJ height loss, no significant differences were observed between protocols. Descriptively, CMJ loss tended to be smaller in the 10:10 protocol and larger in the 30:30 protocol.

Fig. 4.

CMJ height at pre-, mid-, and post-session (left) and percentage CMJ height loss (right). Violin plots illustrate data distribution across protocols. Significance levels: p < 0.05; p < 0.01; p < 0.001.

Figure 5 presents 10-m sprint performance across pre-, mid-, and post moments for the three protocols. A significant main effect of time was observed, with sprint time increasing from pre- to post-moments (p = 0.027; Cohen’s d = 0.454). No significant main effect of protocol or time × protocol interaction was detected, indicating that this decline occurred similarly across all three protocols.

Fig. 5.

Time to complete the 10-m sprint measured pre-, mid-, and post-session for each protocol. Violin plots represent individual responses. p < 0.05.

To provide an integrated overview of the relationships among the analysed variables, a correlation heatmap was generated illustrating the associations between blood lactate concentration, heart rate, CMJ performance, and 10-m sprint performance, including their corresponding significance levels (Supplementary Fig. 2).

Discussion

The present study analysed the acute physiological effects of three circuit training protocols with different work-to-rest interval durations while maintaining a constant 1:1 ratio. Our finding demonstrate that the manipulation of work-rest intervals affects significantly metabolic, cardiovascular and neuromuscular outcomes showing that longer intervals (30:30 protocol) eliciting greater physiological stress and performance decrements compared to shorter intervals (10:10 protocol).

In our study, the 30:30 protocol produced the highest blood lactate concentrations, primarily due to the longer work duration and the greater accumulation of repetitions across sets. This outcome may be explained by the fact that longer work duration, led to a greater number of repetitions and, consequently, higher neuromuscular and metabolic demand. Importantly, the longer work bouts in the 30:30 protocol likely promoted a greater accumulation of repetitions per set and a more pronounced velocity loss, both of which are closely associated with increased glycolytic contribution and lactate production¹⁷. As intermittent exercise bouts become longer, ATP resynthesis relies progressively on anaerobic glycolysis, particularly when recovery periods are insufficient to fully restore phosphocreatine stores^25–27. This interaction between work duration, repetition density, and energetic pathway contribution provides a more comprehensive explanation for the elevated lactate responses observed in the 30:30 condition²⁸. However, when focussing solely on rest intervals, previous studies^29,30 reported higher lactate concentrations with shorter rest intervals. These findings collectively underscore that the length of the rest interval alone does not determine the physiological response and must be interpreted in the context of the total training structure, particularly work duration.

Cardiovascular responses are typically higher in CT than traditional RT⁹, a distinction attributed to the shorter rest intervals performed in circuit training¹⁹. In the present study, significant differences in HR between protocols were observed only at specific time points, particularly between the 10:10 and 30:30 protocols during the later stages of the circuit (minutes 7–8, 8–9, 10–11, and 11–12), with higher HR values in the shorter work-to-rest condition (10:10). These findings indicate a time-dependent effect rather than a consistent protocol-related difference across the entire session. Longer inter-set rest intervals are consistently associated with lower HR responses during resistance training, a pattern reported in controlled experimental protocols³¹. Circuit training settings where HR may remain elevated even during recovery periods⁹, and team-sport contexts such as football training⁸. The higher HR values observed in shorter rest conditions may be explained by limited parasympathetic reactivation during brief recovery periods alongside sustained sympathetic activity^32,33. When rest intervals are very short, parasympathetic dominance may not be fully re-established, which could partially account for the elevated HR observed during the 10:10 circuit training protocol particularly as fatigue accumulates over the session³³. It should be noted that autonomic nervous system activity was not directly measured in the present study. Therefore, interpretations related to parasympathetic reactivation and sympathetic dominance are based on previously established physiological models and should be considered inferential.

Finally, while a significant “Moment × Group” interaction was observed for blood lactate concentration, no such interaction was detected for CMJ or 10-m sprint performance. All CT conditions induced a clear reduction in CMJ height from pre- to post-exercise, reflecting the development of acute neuromuscular fatigue ³⁷ regardless of the work-to-rest interval duration. Although descriptively greater CMJ losses were observed in the 20:20 and 30:30 protocols, these differences were not statistically significant. These results align with the well-documented relationship between metabolic stress in terms of lactate and neuromuscular fatigue reported in the literature¹⁷, and with evidence showing that fatigue can alter neuromuscular control and lower-limb movement characteristics under fatigued conditions³⁴.

Sprint performance over 10 m deteriorated from pre- to post-exercise across all protocols, reflecting a general effect of fatigue rather than a protocol-specific response. This fatigue likely reflects a combination of peripheral mechanisms, such as impaired excitation-contraction coupling and metabolite accumulation, as well as central factors affecting motor unit recruitment^17,18. Metabolic stress, reflected by elevated blood lactate concentration, has been consistently associated with neuromuscular performance decrements, particularly reductions in jump height following high-intensity resistance and sprint-based exercise^16,17. Although higher lactate concentrations were descriptively associated with greater CMJ loss in the longer work-duration protocols, the absence of statistically significant between-group differences suggests that neuromuscular fatigue development may reach a similar threshold across protocols when total session duration is matched, a phenomenon previously reported in resistance-based protocols where different loading or density strategies induce comparable acute fatigue when overall volume is equated^22,35.

Several limitations of the present study should be acknowledged. Although the work-to-rest ratio was held constant, the absolute external loads differed between protocols, with shorter work intervals requiring higher loads, which may have influenced both metabolic and neuromuscular responses, as external load and effort density are known to modulate lactate accumulation and fatigue development during resistance exercise. CMJ and sprint performance are known to present considerable inter-individual variability, particularly in trained populations, which may reduce statistical sensitivity to detect between-protocol differences in acute fatigue responses. Another limitation of this study is that menstrual cycle phase was not formally controlled in female participants. However, all experimental sessions were completed within a 10-day period, and the within-subject repeated-measures design helped reduce inter-individual variability.

Future studies should incorporate repetition-based analyses, standardized external loading strategies, complementary neuromuscular assessments to further elucidate these mechanisms and, with larger sample sizes, should explore inter-individual variability in physiological responses to different circuit training configurations using exploratory clustering or responder classification approaches.

Conclusions

Protocol outcomes were affected by the prescribed work‑rest time structure, leading to distinct acute physiological responses. Longer work bouts (30:30) produced greatest metabolic stress, as reflected by higher blood lactate concentrations, and were associated with larger decrements in neuromuscular performance, as indicated by CMJ loss. However, all protocols induced a reduction in CMJ height, with descriptively larger losses observed under longer work durations (20:20 and 30:30), although these differences were not statistically significant between protocols. These findings are consistent with the established association between metabolic stress and neuromuscular fatigue reported in the literature. In contrast, 10-m sprint performance declined over time across all protocols, indicating a time-dependent fatigue effect rather than a protocol-specific response. Heart rate responses differed between protocols only at specific time points, with higher values observed during the later stages of the session in the 10:10 protocol, indicating that HR responses during circuit training are influenced by rest duration in a time-dependent manner rather than by a uniform protocol effect. These findings reflect acute physiological responses, and caution is therefore warranted when extrapolating the present results to long-term training adaptations or chronic programming outcomes.

From an applied perspective, coaches and practitioners may select the 30:30 protocol when the primary goal is to maximize metabolic stress and neuromuscular fatigue, whereas the 10:10 protocol may be preferred to elicit higher cardiovascular demands with lower lactate accumulation. The 20:20 protocol may represent a practical intermediate option when a balanced metabolic and cardiovascular stimulus is desired.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ANOVA

Analysis of variance

CMJ

Countermovement jump

RT

Resistance training

CT

Circuit training

HR

Heart rate

Author contributions

Aaron Agudo-Ortega and Marcos Gomez-Castro contributed to data collection and statistical analysis. Rafael Olivares-Llorente and Adrián Martín-Castellanos participated in the experimental design and supervised the exercise testing sessions. Manuel Barba-Ruíz assisted in data interpretation and manuscript editing. Francisco Hermosilla-Perona conceived the study, coordinated the research project, and drafted the manuscript. All authors reviewed and approved the final version of the paper.

Funding

The authors declare that they have not received any funding or financial support for the preparation of this manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.