Acute mechanical, physiological and perceptual responses in older men to traditional-set or different cluster-set configuration resistance training protocols

Aging is a complex and multidimensional process characterized by a variety of biological changes, degenerative in nature, which contribute to impaired physiological processes. As such, a general decline in function with increased likelihood of adverse health outcomes ensues (Lally and Crome 2007). Even with healthy aging, functional capacity can decline as much as 40% between 60 and 90 years of age (Rikli and Jones 2013), commonly arising from factors such as skeletal muscle atrophy and reduced muscular strength. Low muscular strength causes functional challenges due to the reduced capability to produce forces necessary to accomplish activities of daily living. Importantly, the ability to exert high levels of force at high velocity (i.e. muscle power) decreases with advancing age faster than the muscle mass loss and muscular strength reduction (Metter et al. 1997; Skelton et al. 1994), due to a selective loss of Type II fibers in old age. These age-related impairments in skeletal muscle morphology, strength and more meaningfully, power capabilities, are part of the causal pathway for secondary adverse outcomes such as frailty, reduced mobility (Bischoff-Ferrari et al. 2015; Schaap et al. 2018), longer hospitalisation (Cawthon et al. 2017) and specific comorbidities including poorer bone health, osteoporosis (Bischoff-Ferrari et al. 2015; Schaap et al. 2018), obesity and type 2 diabetes (van Sloten et al. 2011). Moreover, a decline in muscle function is strongly associated with future physical disability and mortality from middle-age into later life (De Buyser et al. 2016).

Given the undesirable consequences of aging on the musculoskeletal system, strategies for preserving muscle function, muscular strength and muscle power are of great importance for the health and wellbeing of older adults. One common strategy to attenuate the effects of aging on neuromuscular function and functional capacity (Borde et al. 2015; Cadore et al. 2014; Peterson et al. 2010) is resistance training. Resistance training programs induce beneficial adaptations in both healthy older adults and those with chronic conditions if performed regularly (e.g., 2–3 days per week) in a periodized manner, with sufficient volume (e.g., 2–3 sets of 8–12 repetitions per exercise) and at an adequate intensity (e.g., 70–85% of 1 repetition maximum [1RM]) (Fragala et al. 2019; Medicine 2009; Pescatello et al. 2014). However, whereas these broad guidelines are commonly recommended to pursue muscle hypertrophy effects and general strength development (Fragala et al. 2019; Medicine 2009; Pescatello et al. 2014), approaches more oriented towards power enhancements could optimize the training responses underpinning greater functional improvements in older people. In support of this rationale, previous studies (Ramírez-Campillo et al. 2014, 2017; Ramirez-Campillo et al. 2016) have consistently reported power training protocols to induce greater improvements in functional performance outcomes and quality of life compared to traditional resistance training protocols in older people. In view of this accumulating evidence, investigating training protocols designed to optimize power output by manipulating resistance training variables seems prudent. One viable strategy to optimize the individual power responses to training is by using optimal power loads (OPL), accurately calculated for a specific resistance exercise and more importantly tailored to individual mechanical profiles (Loturco et al. 2015). Earlier studies conducted in athletic populations have reported greater power responses following resistance training regimens implementing OPL as opposed to protocols designed using absolute percentages of 1RM as reference (Dello Iacono et al. 2018; Dello Iacono and Seitz 2018; Oliver et al. 2016a). Another potential method to maximize power outputs in resistance training is through cluster-set configurations, which include short rest periods between repetitions within a given set. Cluster-set configurations can be designed by redistributing the repetitions within a given set into small clusters (e.g., 2–6 clusters of 2–4 of repetitions) separated by brief rest periods (e.g., 10–60 s). Cluster-set configurations facilitate greater force output, velocity and consequently power at a given load compared to traditional configuration (absent of intra-set or inter-repetition rest) in athletes (Tufano et al. 2016, 2017) and older adults alike (Ramirez-Campillo et al. 2018). Interestingly, cluster-sets also allow greater power output when compared to matched protocols loaded with the same OPL but designed as traditional-set configurations (Dello Iacono et al. 2019). The cumulative beneficial effects of OPL and cluster-set configurations on power output likely stem from psychophysiological (González-Hernández et al. 2020) and metabolic (Gorostiaga et al. 2014, 2010) mechanisms resulting in lower perceived effort and reduced acute muscular fatigue (Tufano et al. 2016, 2017). Furthermore, the rest interval between consecutive clusters significantly affects cardiovascular load (Fleck 1988, 2003; Kraemer et al. 1987), with very short rest periods (i.e. < 20 s) leading to higher heart rate (HR) responses. Therefore, it is worth identifying optimal set configurations (i.e. number of repetitions per cluster and associated inter-cluster rest duration) to induce beneficial cardiovascular responses alongside the known advantageous effects on acute muscular performance. In particular, the combination of cluster-sets and OPL may represent a viable method for optimizing power output and physiological responses in older adults performing resistance training. More importantly, optimization of power output could consequently lead to increased functional capacity and performance of activities of daily living.

The aims of this study were twofold: first, to compare mechanical responses to traditional-set or different cluster-set configuration resistance training protocols using OPL in older people. Second, assuming cluster-set configuration would optimize mechanical performance, we aimed to examine the underlying mechanisms by investigating physiological (i.e. HR) and perceptual responses (i.e. effort and fatigue). In view of the above, we hypothesize that resistance training protocols using OPL together with the cluster-set configurations would result in: (1) greater power outputs, (2) greater physiological responses (i.e. HR) and (3) lower perceived effort and fatigue, compared to a protocol configured in a traditional manner. To these ends, healthy old men completed four resistance training protocols, designed as traditional-set or cluster-set configurations, using the back squat exercise loaded with OPL.

ICCs for baseline HR, and first repetition MPP and first repetition impulse across the four protocols were 0.82 (95% CI 0.67, 0.92), 0.96 (95% CI 0.92, 0.98) and 0.76 (95% CI 0.72, 0.78), respectively. These results demonstrate good to excellent between-protocol reliability. No differences were found between experimental protocols for baseline HR responses (all p \(\ge\) 0.91), MPP (all p \(\ge\) 0.21) and impulse (all p \(\ge\) 0.08) of the first repetition (Table 1 for descriptive statistics and 95% CI).

A main effect for protocol (F_{(3, 76)} = 181.5, p < 0.001) and interaction (protocol \(\times\) set differences; F_{(9, 171)} = 86.9, p < 0.001) was observed between experimental protocols for HR response. In particular, CLU1 and CLU2 induced greater HR compared to TRA (CLU1 vs TRA, all p \(<\) 0.001; CLU2 vs TRA, all p \(<\) 0.001) and CLU4 (CLU1 vs CLU4, all p \(<\) 0.001; CLU2 vs CLU4, all p \(<\) 0.001) from the completion of the first sets throughout the entire sessions (Table 1 and Fig. 3).

A main effect for protocol (F_{(3, 57)} = 62.5, p < 0.001) and protocol \(\times\) set interaction (F_{(6, 114)} = 3.7, p = 0.002) was observed for MPP (Table 1; Fig. 4). Across all conditions, three consistent patterns emerged. First, cluster-set configurations consistently resulted in greater MPP compared to traditional-set configuration: CLU1 vs TRA (Set 1: 95% CI 0.42, 0.61; Set 2: 95% CI 0.52, 0.71; Set 3: 95% CI 0.57, 0.77), all p < 0.001; CLU2 vs TRA (Set 1: 95% CI 0.17, 0.36; Set 2: 95% CI 0.23, 0.42; Set 3: 95% CI 0.35, 0.54), all p < 0.001; CLU4 vs TRA (Set 1: 95% CI 0.07, 0.26; Set 2: 95% CI 0.02, 0.21; Set 3: 95% CI 0.15, 0.34), all p \(\le\) 0.007. Second, a step-wise increase in MPP was observed when clusters went from high to low number of repetitions per cluster: CLU1 vs CLU4 (Set 1: 95% CI 0.26, 0.45; Set 2: 95% CI 0.4, 0.59; Set 3: 95% CI 0.33, 0.52), all p < 0.001; CLU1 vs CLU2 (Set 1: 95% CI 0.16, 0.35; Set 2: 95% CI 0.19, 0.38; Set 3: 95% CI 0.13, 0.32), all p < 0.001; CLU2 vs CLU4 (Set 1: 95% CI 0.01, 0.2; Set 2: 95% CI 0.11, 0.31; Set 3: 95% CI 0.1, 0.3), all p \(\le\) 0.03 (Table 1; Fig. 4). Finally, a significant progressive decrease of power output was observed across the three consecutive sets during the TRA protocol (all p \(\le\) 0.037). A similar pattern was also observed during CLU4 protocol, but differences reached significance only between set 1 and the two other sets (p < 0.001).

No main effect (F_{(3, 57)} = 0.6, p = 0.62) was found between experimental protocols for impulse; however, a significant protocol \(\times\) set interaction (F_{(6, 114)} = 4.9, p < 0.001) was observed (Table 1; Fig. 5). In particular, the traditional-set configuration and the cluster protocol with higher number of repetitions resulted in greater impulse outputs compared to the cluster protocols with low number of repetitions in set 3: TRA vs CLU1 (95% CI 0.04, 0.15, p < 0.001); TRA vs CLU2 (95% CI 0.04, 0.15, p < 0.001); CLU4 vs CLU1 (95% CI 0.03, 0.14, p < 0.001); CLU4 vs CLU2 (95% CI 0.02, 0.13, p = 0.002).

Significant differences were observed between the experimental protocols for RPE (χ² = 57.4, p < 0.001) and ROF (χ² = 50.5, p < 0.001), with a consistent pattern across protocols: CLU1 and CLU2 induced lower RPE [CLU1 vs TRA (95% CI − 3.2, − 1.8; p = 0.006); CLU2 vs TRA (95% CI − 2.6, − 1.4; p = 0.004); CLU1 vs CLU4 (95% CI − 3.2, − 1.7; p = 0.005); CLU2 vs CLU4 (95% CI − 2.7, − 1.3; p = 0.003)] combined with lower ROF scores [CLU1 vs TRA (95% CI − 4, − 2.2; p = 0.006); CLU2 vs TRA (95% CI − 3.6, − 2.1; p = 0.005); CLU1 vs CLU4 (95% CI − 1.4, − 0.17; p = 0.012); CLU2 vs CLU4 (95% CI − 1, − 0.15; p = 0.006)] compared to the TRA and CLU4 configurations. Lower ROF responses were also observed following CLU4 compared to TRA (95% CI − 1.5, − 3; p = 0.004). (Table 1). Finally, an absolute preference for the cluster-set conditions was observed with 15, 3 and 2 subjects choosing the CLU2, CLU1 and CLU4, respectively.

The relationship between HR response and duration of set 1 is shown in Fig. 6. The polynomial regression model confirmed HR was dependent on duration, which explained the variance HR with 90% confidence (i.e. R² adjusted = 0.90; p < 0.001). The best-fit equation of the model is reported below:

In practical terms, when the duration increases of 1 unit (i.e. 1 s) the HR is expected to change by 0.02%. For example, the model predicted an average difference in HR at completion of set 1 between CLU4 and CLU1 protocols of 9.1% (95% CI 8.7, 9.5) while the actual difference measured from data collected during the study was 8.7% (95% CI 7.7, 9.6).

In the current study, we examined the mechanical, physiological and perceptual responses to resistance training protocols designed with either traditional or cluster-set configurations in older men. Four main findings emerged: (1) cluster-set configurations resulted in greater MPP than traditional-set configuration; (2) clusters of lower repetition number (in this instance one and two) produced greater MPP than clusters of greater numbers (in this instance four); (3) clusters of lower repetition number induced greater HR and lower RPE and fatigue responses than traditional-set and clusters of greater numbers; and (4) cluster-set protocols were preferred over the traditional-set configuration, with clusters of two repetitions reported as the favorite.

Regardless of the set configuration, the three clusters protocols investigated in this study were associated with greater power outputs than the traditional-set protocol throughout the entire session. These findings are in agreement with previous investigations endorsing cluster-set configurations to optimize power outputs during the back squat exercise (Oliver et al. 2015, 2016b; Tufano et al. 2016; Wetmore et al. 2019). For example, in the study of Oliver et al. (2015) subjects performed back squats loaded with 70% of 1RM either as 4 sets of 10 repetitions with 120 s rest between sets, or as 4 sets of 2 clusters of 5 repetitions each with 30 s between clusters and 90 s between sets. Although the two training configurations were matched for exercise intensity, total volume and work: rest ratio, greater power outputs were observed during the cluster-set protocol. The authors concluded that the superior effects of the cluster-set configuration likely arose from the different rest structure and the inclusion of a short rest between consecutive clusters. Building on the findings of Oliver et al., we further examined the effects of cluster protocols designed with different configurations (one, two and four repetitions) and rest intervals duration (10, 20 and 40 s). The novel finding emerging from this study is that smaller clusters (one and two repetitions) including shorter (10–20 s) but more frequent rest intervals are an effective strategy to attenuate the velocity loss across repetitions and sets and to maximize power outputs. We assume that the rest interval embedded between clusters can reduce the rate of adenosine triphosphate (ATP) depletion and favor partial or complete regeneration of phosphocreatine (PCr) within the working muscles, thereby allowing maintenance of greater power outputs (Bogdanis et al. 1998; Gorostiaga et al. 2010, 2014). Moreover, we presume this mechanism may be exploited by implementing clusters with lower number of repetitions and frequent shorter rest intervals. Our assumption is supported by two main observations. The first, subjects produced greater power output during protocols with smaller clusters as confirmed by the stepwise trend of power output observed in this study (Fig. 3). Second, while power output decrement was present in all protocols, the decline was significant only during TRA and CLU4 set configurations (Table 1). In particular, power output decrements observed across consecutive sets during the TRA protocol, may have been caused by an acute increase in metabolic stress. This hypothesis is in agreement with previous studies, in which traditional-set configurations similar to the one used in this study led to greater blood lactate concentration and the consequent inability to maintain optimal power outputs (Girman et al. 2014; González-Hernández et al. 2020; Gorostiaga et al. 2014, 2010). As a result, prescribing resistance training using cluster-set configurations with low repetition numbers may help to avoid these detrimental metabolic effects and provide a greater stimulus for power enhancement adaptations than resistance training using traditional-set configuration and clusters of many repetitions. This is particularly pertinent in an aging population, as generating force at a high velocity is imperative to maintain mobility, independence and quality of life (Clark and Manini 2012; Runge et al. 2004). This hypothesis conforms to the findings of the study of Ramirez-Campillo et al. (2018) who reported greater improvements in functional performance and quality of life in older women (> 65 years) following a 12-week cluster-set configured resistance training programme compared to a matched programme with a traditional-set configuration and a control group.

The superior effects of cluster-set configurations with a low number of repetitions on power outputs are further supported by the HR data, whereby we observed greater HR responses during CLU1 and CLU2, than CLU4 and TRA across sets and throughout the entire session. These findings are in agreement with previous evidence suggesting the recovery within and between sets in resistance training sessions has an impact on cardiovascular responses (Fleck 1988, 2003; Kraemer et al. 1987). Differences in HR between protocols were already found at the completion of set 1 (Table 1; Fig. 4), most likely due to the significant longer duration of the same set in CLU1 (\(\sim 118\) s) and CLU2 (\(\sim 109\) s) compared to TRA (\(\sim 45\) s) CLU4 (\(\sim 82\) s), as confirmed by the results of the regression analysis between HR responses and set duration (Fig. 6). This suggests a greater metabolic demand during CLU1 and CLU2, and a greater aerobic training intensity and load, which could induce superior aerobic adaptation than the other protocols. In particular, the average HR responses achieved in CLU1 and CLU2 sessions were approximately 65% of HR_max, thus within the 60–90% range that is recommended for the development of cardiorespiratory fitness and promotion of body composition changes (Medicine 2009; Pescatello et al. 2014). These findings have practical importance as resistance training with multiple exercises in cluster-set configurations could lead to superior cumulative peripheral and central stimuli and increase effective time within optimum training zones with consequent greater adaptations on both muscular and cardiovascular function.

As RPE and ROF were also lower when participants completed CLU1 and CLU2, it appears that these protocols, although exhibiting greater internal and external intensity, and subsequent load, were perceived as easier and less fatiguing. While the speculation mentioned previously concerning the likely differences between protocols in lactate concentration may hold true for greater perceived effort and fatigue following TRA and CLU4 protocols, we consider another reason to explain why such differences occurred. Greater impulses were observed in set 3 during TRA and CLU4 compared to CLU1 and CLU2 coupled with a dissimilar pattern for power outputs (i.e. CLU1 and CLU2 greater than TRA and CLU4 in set 3). The combination of greater impulse (i.e. force \(\times\) time) and lower power (i.e. force \(\times\) velocity [distance/time]) predicates longer duration for the eight repetitions in the last set, given that distance covered was constant. The longer duration of these repetitions in TRA and CLU4 may have caused greater perceptions of effort, fatigue and discomfort (Fisher and Steele 2017). Individuals tend to remember the peak and the end of an event (i.e. peak-end rule) (Hargreaves and Stych 2013; Kahneman et al. 1993), which in this study was the last set completed, and report perceptions accordingly. This also explains the divergent pattern which emerged between HR and RPE, although they are generally well correlated in response to exercise (Kraemer et al. 1987). In light of the above and in consideration that perception of effort and fatigue are key determinants of exercise adherence (Hartman et al. 2019; Prasad and Cerny 2002; Zenko et al. 2016), it seems prudent that CLU1 and CLU2 should be advocated when prescribing muscle strengthening exercises in older men. Finally, participants reported a preference for CLU2 over the other protocols (15, 3 and 2 for CLU2, CLU1 and CLU4, respectively). Taken together, results presented here suggest CLU2 may be the optimal resistance training design in older men, due to preference, low perception of effort and fatigue, greater HR responses and kinetic responses. CLU1 was superior in terms of MPP, and matched CLU2 in terms of HR and perception of effort and fatigue, yet it was not preferred by the majority of participants. Therefore, we feel it pragmatic to suggest CLU2 would be better received in this cohort.

This study is not without limitations. First, our participants were not resistance exercise naïve, being accustomed to this form of training. As exercise is a primary countermeasure to human aging, and the age-associated decline in muscle strength and mass, these individuals are not an at-risk cohort for frailty, so whether these findings translate to untrained individuals required further research. In this context, as our participants were trained, they presumably had a predilection for exercise, so perceptual responses may be different in this cohort compared to an exercise-adverse cohort. Second, the effects of the four protocols were investigated only on the back squat exercise loaded with OPL. This fact narrows the ability to generalize the results from this study to other conventional lower and upper body resistance training exercises across a broader range of loads and intensities. Finally, another limitation was the absence of additional physiological measurements (e.g., hormonal and lactate concentrations), which may have helped in better understanding the underlying mechanisms of resistance training under traditional and cluster-set configurations.

The findings of the current study have few important practical implications. First, they suggest cluster-set configurations effectively maintain power outputs during the back squat, one of the most commonly prescribed lower limbs resistance training exercises (Tufano et al. 2017). Second, they indicate cluster-set configurations combining lower repetition numbers with shorter and more frequent rests lead to greater internal training load and low perception of effort and fatigue. Finally, cluster-set configurations are consistently preferred over a traditional-set configuration by trained older men. Consequently, it would be of great interest to investigate the long-term effects of different cluster-set modes on muscular and cardiovascular adaptations and functional performance of aging cohorts (Ramirez-Campillo et al. 2018). Future studies are warranted to elucidate the optimal combination of investigated protocols into a mixed approach, including aerobic exercise and balance training for older adults (Fragala et al. 2019).