Isokinetic eccentric exercise substantially improves mobility, muscle strength and size, but not postural sway metrics in older adults, with limited regression observed following a detraining period

Neuromuscular and musculoskeletal functional decline in ageing presents with comorbidities associated with falls including, but not limited to, a loss of muscle mass and strength (Hairi et al. 2010) and impaired balance and mobility (Delbaere et al. 2010). Whilst a range of exercise interventions have been demonstrated to reverse age-related reductions in physical function (Chou et al. 2012), limited cardiorespiratory capacity, locomotor ability, and fear of falling can reduce exercise capacity and tolerance (LaStayo et al. 2003). These issues can subsequently limit the ability to exercise at a sufficient intensity to prevent further losses in neuromuscular function, accelerating decline and contributing to the development of frailty (Cesari et al. 2014), increased levels of disability (Freedman et al. 2002), and poor quality of life (Chou et al. 2012). Consequently, improvements in muscle strength evoked by traditional progressive resistance training can be somewhat limited and, importantly, are often not transferred to improved balance or mobility performances (Orr et al. 2008). Therefore, developing resistance exercise strategies with sufficient stimulus to effectively and simultaneously counteract multiple impairments, whilst minimising the challenges older adults face with traditional forms of exercise, remains a high priority in ageing research.

Resistance training is widely used to improve muscle mass and strength in older adults within multi-component fall prevention programmes (Sherrington et al. 2019), with high-intensity resistance training also providing the necessary stimulus for neuromuscular adaptations leading to enhanced mobility (Hess and Woollacott 2005). Although reduced cardiorespiratory function can limit exercise tolerance and adherence to traditional forms of high-intensity exercise in older adults (LaStayo et al. 2003), the metabolic demand of eccentric exercise is ~ 25% that of concentric exercise at a comparable workload (LaStayo et al. 1999). Consequently, perceived effort during eccentric-only exercise is reduced (LaStayo et al. 2003) and exercise tolerance has been shown to be improved in older adults and individuals with cardiorespiratory impairments when compared with traditional resistance-based exercises that include cyclical concentric and eccentric contractions (Rooyackers et al. 2003). Furthermore, seated resistance exercise (e.g. performed on isokinetic machines) can be used in populations with poor mobility or fear of falling, whilst allowing for an adequate stimulus to be provided for lower-limb muscle strength improvement, as previously demonstrated in both young adult (Kay et al. 2018) and older (LaStayo et al. 2003) populations. Collectively, these findings suggest that seated eccentric-only exercise may be ideally suited to the specific physical challenges older adults face with traditional aerobic- and resistance-based exercises.

The superiority of adaptations to eccentric resistance training over isometric or concentric exercise has been demonstrated repeatedly in young adult populations with greater increases in muscle mass (Roig et al. 2009), strength (Hortobagyi et al. 1996), and changes in muscle architectural characteristics (i.e. fascicle length and pennation angle [Ema et al. 2016]). Interestingly, strength and mass also remained significantly improved after 6 weeks of detraining in well-trained young adults after eccentric, but not concentric, exercise (Coratella and Schena 2016). These findings are particularly relevant for older populations who often undergo periods of incapacitation; however no data exist describing the detraining effects after the cessation of eccentric exercise in older adults. Nonetheless, comparable improvements in functional mobility (i.e. timed-up-and-go test; TUG) have been reported after eccentric resistance training compared to traditional resistance training (Gault et al. 2012), with TUG times remaining significantly improved after 4 months of detraining following high-intensity (~ 80% maximal voluntary contraction [MVC]) but not low-intensity (~ 55% MVC) traditional strength training (Fatouros et al. 2005). Whilst chair-based exercise is often characterised as low intensity and lacking sufficient stimulus to elicit substantial adaptation, seated isokinetic and isotonic eccentric exercises have been reported to rapidly improve muscle strength and mass in young (Kay et al. 2018) and older (Reeves et al. 2009) adult populations. As muscle weakness is strongly related to fall risk (Benichou and Lord 2016), these findings indicate that eccentric exercise may be used to reverse several important comorbidities associated with fall risk, preventing the development of frailty, disability and institutionalisation, possibly with limited regression during periods of detraining, outcomes likely having important clinical implications for exercise prescription in older populations.

The association between lower-limb strength and balance is well established (Lord et al. 1991); nonetheless, reviews have reported that strength training often fails to improve static balance (i.e. postural sway [Orr et al. 2008]), possibly because the hip and knee extensors are targeted due to their relative importance for stair climbing and rising from a chair, but not the ankle musculature that influences postural sway assessments (Winter 1995). Therefore, as eccentric resistance training provides the greatest improvements in strength (Hortobagyi et al. 1996), and the strength of the ankle musculature is also important for balance (Winter 1995), eccentric resistance training programmes targeting the ankle musculature may provide simultaneous improvements in both strength and balance, two primary risk factors associated with falls in older adults (Hairi et al. 2010; Johansson et al. 2017). However, currently no data exist describing the effects of lower-limb isokinetic eccentric exercise on static balance (i.e. using quantitative posturography), an important predictor of future falls (Johansson et al. 2017), nor regression following the cessation of eccentric training in older adults, limiting our understanding of mid- to long-term potential benefits through periods of reduced activity more common in older adults. To overcome some of these issues, the aims of the present study were to examine the effects of two 6-week low-intensity, isokinetic eccentric lower-limb resistance training programmes targeting the hip and knee extensor muscle groups (ECC) or the hip, knee and ankle extensor muscle groups (ECC_PF), and 8 weeks of subsequent detraining, on functional mobility, static balance, lower-limb eccentric muscle strength, and vastus lateralis muscle thickness and architecture in moderately active older adults. We hypothesised that both isokinetic eccentric resistance training programmes would significantly reduce TUG time and increase muscle strength, thickness and architecture, and that the positive effects of training would be largely retained at 8 weeks following the cessation of training. We also hypothesised that postural sway would be significantly reduced only in the ECC_PF group (i.e. with the inclusion of plantarflexor training).

The two-way mixed model ANOVA revealed a significant main effect of time (F = 40.67, P < 0.001, η_p² = 0.63), but not group (F = 3.19, P = 0.09, η_p² = 0.12) for mobility (i.e. the time to complete the TUG test). Post hoc within-subject analyses revealed significant (P < 0.01) reductions in the time to complete the TUG test after training in ECC (− 12.0 ± 8.9% [− 0.94 ± 0.82 s], d = 1.15) and ECC_PF (− 7.7 ± 6.4% [0.64 ± 0.56 s], d = 1.14) (Table 1). No significant regression was detected after 8 weeks of detraining (~ 100% of the mean improvement being retained), with TUG time remaining significantly (P < 0.001) faster than at pre-training in ECC (− 11.3 ± 7.6% [− 0.84 ± 0.63 s], d = 1.34) and ECC_PF (− 8.2 ± 6.1% [0.66 ± 0.53 s], d = 1.24).

No significant main effects of time (F = 1.11–3.54, P > 0.05, η_p² = 0.06–0.13) or group (F = 0.005–0.46, P > 0.05, η_p² = 0.001–0.020) were detected in any postural sway metric (COP_ML, COP_AP, 95% ellipse, COP velocity) in either EO or EC conditions (Table 1).

Significant main effects of time (F = 60.44, P < 0.001, η_p² = 0.72) and group (F = 7.20, P = 0.04, η_p² = 0.23) were detected in eccentric leg press strength. Post hoc within-subject analyses revealed a significant (P < 0.01) increase in strength after training in ECC (58.8 ± 39.9% [573 ± 319 N], d = 1.80) and ECC_PF (39.4 ± 25.0% [323 ± 158 N], d = 2.04) (Fig. 3a). No significant regression occurred after detraining (~ 71–85% of the mean increase being retained), with strength remaining significantly (P < 0.01) greater than at pre-training in ECC (50.4 ± 38.0% [513 ± 326 N], d = 1.57) and ECC_PF (30.5 ± 28.7% [227 ± 166 N], d = 1.36). Post hoc between-subject analyses revealed no significant difference between groups at baseline, but significant (P < 0.05) differences post-training between ECC (1667 ± 631 N) and ECC_PF (1223 ± 379 N) and after detraining between ECC (1666 ± 540 N) and ECC_PF (1127 ± 311 N) were observed.

A significant main effect of time (F = 51.20, P < 0.001, η_p² = 0.68), but not group (F = 0.33, P = 0.57, η_p² = 0.01) was detected in VL muscle thickness. Post hoc within-subject analyses revealed significant (P < 0.01) increases in VL thickness after training in ECC (9.8 ± 1.3% [1.89 ± 0.74 mm]; d = 2.54) and ECC_PF (9.9 ± 1.8% [1.80 ± 1.05 mm], d = 1.71) (Fig. 3b). Although a significant (P < 0.05) reduction in VL muscle thickness was detected after detraining in both groups, VL thickness remained significantly (P < 0.01) greater than at pre-training in ECC (6.1 ± 1.6%; d = 1.14) and ECC_PF (7.1 ± 1.7%; d = 1.30), with ~ 62–71% of the mean increase in muscle thickness being retained.

A significant main effect of time (F = 9.49, P < 0.01, η_p² = 0.28), but not group (F = 1.61, P = 0.22, η_p² = 0.06) was detected in VL fascicle length. Post hoc within-subject analyses revealed significant (P < 0.01) increases in VL fascicle length after training in ECC (5.4 ± 4.3% [5.3 ± 4.5 mm]; d = 1.17) and ECC_PF (4.8 ± 4.4% [5.1 ± 4.7 mm], d = 1.07) (Fig. 4a). No significant regression occurred after detraining (~ 94–99% of the mean increase retained), with fascicle length remaining significantly (P < 0.01) longer than at pre-training in both ECC (5.8 ± 7.5% [5.2 ± 7.1 mm], d = 0.74) and ECC_PF (4.9 ± 6.8% [4.8 ± 7.3], d = 0.65).

A significant main effect of time (F = 8.25, P < 0.01, η_p² = 0.26) and group (F = 7.87, P < 0.05, η_p² = 0.25) was detected in VL pennation angle. Post hoc within-subject analyses revealed a significant (P < 0.01) increase in VL pennation angle after training in ECC (4.4 ± 5.7% [0.52 ± 0.69°], d = 0.75) and ECC_PF (5.0 ± 5.0% [0.50 ± 0.48°], d = 1.05) (Fig. 4b). Significant reductions in VL pennation angle were detected after detraining, with no significant difference detected compared to pre-training in ECC (0.5 ± 5.1% [0.01 ± 0.64°], d = 0.02) and ECC_PF (2.5 ± 8.3% [0.23 ± 0.83°], d = 0.28), indicating that pennation angle had largely returned to baseline. Post hoc between-subject analyses revealed significant (P < 0.05) differences in VL pennation angle (1.1–1.3°) between groups at pre-training (ECC = 11.6 ± 1.6°; ECC_PF = 10.3 ± 0.7°), post-training (ECC = 12.1 ± 1.8°; ECC_PF = 10.8 ± 0.7°), and after detraining (ECC = 11.7 ± 1.4°; ECC_PF = 10.6 ± 0.8°).

When the relationships between absolute change data were examined, significant correlations were observed between changes in VL pennation angle and both muscle thickness (r = 0.54; P < 0.01) and fascicle length (r = − 0.47; P < 0.05), but not between changes in muscle thickness and fascicle length (r = 0.35; P = 0.07). No significant correlations were observed between the changes in strength or TUG and any other measure (r = − 0.19 to 0.22; P > 0.05).

A significant main effect of time (F = 4.78, P < 0.01, η_p² = 0.16), but not group (F = 0.76, P = 0.39, η_p² = 0.03) was detected in weekly RPE. Post hoc within-subject analyses revealed significantly greater RPE (P < 0.05) in weeks 3 and 5 (compared to other weeks) in the ECC_PF group (i.e. weeks where absolute intensity was increased). No further difference in RPE was observed between any other week in either programme, with RPE remaining consistently low throughout the ECC (3.3–4.1 out of 10; i.e. ‘moderate’ to ‘somewhat hard’) and ECC_PF (3.6–4.5 out of 10) programmes (Fig. 5a).

A significant main effect of time (F = 12.43, P < 0.001, η_p² = 0.33), but not group (F = 1.79, P = 0.19, η_p² = 0.07) was detected in weekly leg press step accuracy. Post hoc within-subject analyses revealed significantly (P < 0.05) greater leg press step accuracy (i.e. the ability to match the required 50% MVC target force) in weeks 2–6 (ECC = 91.3–96.6%, ECC_PF = 95.0–98.7%) than week 1 (ECC = 82.8%, ECC_PF = 90.7%). No further difference in leg press accuracy was observed from weeks 2 to 6 in either programme (Fig. 5b). Similarly, a significant (F = 14.93, P < 0.001, η_p² = 0.52) increase in straight-legged plantarflexor step accuracy was detected during the ECC_PF training programme with accuracy greater in weeks 2–6 (91.7–96.8%) than week 1 (88.1%). No further difference in straight-legged plantarflexor accuracy was observed from weeks 2 to 6.

Impaired mobility is an important risk factor for falls, with individuals recording TUG times of > 13.5 s having a 90% probability of being a faller (Shumway-Cook et al. 2000). However, subjects in the present study were moderately active and completed the TUG in 7.3–7.8 s (mean range for both groups) during pre-training assessments, times at the faster end of the normative spectrum (7.1–9.0 s) for community dwelling-older adults aged 60–69 years (Bohannon 2006). Nonetheless, in agreement with our hypothesis, a 9–12% (− 0.64 s to − 0.94 s [d = 1.14–1.15]) improvement in mobility performance (faster TUG time) was observed after both eccentric training programmes, placing the subjects well above the expected range within this demographic. While one meta-analysis revealed that traditional forms of resistance training did not improve mobility in older adults (Chou et al. 2012), the improvement in mobility in the present study is consistent with previous studies examining the effects of eccentric resistance training in older adults (LaStayo et al. 2003; Gault et al. 2012]. Novel to the present study, however, is the assessment of detraining effects following the cessation of the interventions. Importantly, no regression in mobility occurred after the 8-week detraining period, indicating that a relevantly short period of eccentric training was sufficient to improve mobility in an already-mobile cohort with the retention of improvements maintained through a detraining period. As older adults can be more prone to training interruptions (due to injury and/or illness, caring for a spouse or child, etc.), the retention of mobility is valuable to practitioners, care providers and exercise professionals involved with exercise programming and implementation. From a fall-risk perspective, it may be a worthwhile strategy to prescribe eccentric resistance training to older adults who are expecting a future period of inactivity (i.e. prehabilitation to assist in the preservation of mobility performance).

Previous studies examining the effects of eccentric resistance training in older adults (Gault et al. 2012; LaStayo et al. 2003) have not examined quiet standing postural sway, which is an important risk factor associated with falls (Johansson et al. 2017). Despite improvements in dynamic balance (i.e. mobility) being evoked in the present study, and the inclusion of plantarflexor training in the ECC_PF group, no significant change in any COP measure of postural sway (6.9–14.0%; η_p² = 0.06–0.13) was observed in the present study. This result runs counter to our hypothesis. Nonetheless, these data are consistent with previous systematic reviews (Orr et al. 2008) and meta-analyses (Low et al. 2017) examining the impact of traditional resistance exercises on static balance performance. The underlying mechanisms governing the significant improvements in dynamic balance (i.e. TUG time), but not static balance (i.e. postural sway), in the present study are not clear, although distinct neuromuscular qualities may need to be targeted separately for their respective improvements (Muehlbauer et al. 2015). Importantly, eccentric training is characterised by the preferential recruitment of fast twitch motor units (Douglas et al. 2017), which would logically improve performance during fast velocity dynamic tasks (i.e. TUG [Miller et al. 2015]), but not during quiet standing as the antigravity muscles that minimise body sway show a greater predominance of type I muscle fibres (Paillard 2017). Alternatively, relatively low-intensity and short duration training programmes may not be sufficient to generate adaptations in standing balance, and the seated exercise did not involve body movements that stress, and therefore improve, standing balance (Lesinski et al. 2015). Furthermore, the present subjects had good balance scores at the commencement of training, which may have provided less scope for further change, i.e. a ceiling effect (Orr et al. 2008), or the quiet bipedal standing task may not have been adequate to challenge the postural control system in these minimally impaired individuals (Clifford and Holder-Powell 2010). It will be of interest to observe the adaptive response in individuals with poorer balance and mobility, possibly after using greater exercise intensities, over a longer training period, and with more challenging assessments of static (i.e. standing on foam) or reactive (i.e. surface perturbations) balance.

Reductions in muscle strength and mass are commonly associated with both fall risk (Benichou and Lord 2016) and mortality (Chuang et al. 2014) in older adults, therefore identifying exercise modalities that limit or reverse functional decline is a priority in ageing research. In agreement with our hypothesis, a very large increase in lower-limb eccentric muscle strength (~ 40–60%; d = 1.80–2.04) was detected after both 6-week eccentric training programmes, data consistent with a previous study employing a longer-duration (11 weeks) eccentric training programme in frail older adults (LaStayo et al. 2003). In addition to the rapid and substantial increase in strength, significant increases in VL thickness (9.8–9.9%; d = 1.71–2.54) and important functional architectural characteristics (fascicle length, pennation angle) were observed. These findings are consistent with the functional and architectural adaptations observed in young adults after this type and duration of training (Blazevich et al. 2007; Kay et al. 2018), indicating similar adaptive profiles of muscle tissue in young and older adult populations. A significant correlation was observed between changes in muscle thickness and pennation angle (r = 0.54; P < 0.01), but not between changes in muscle thickness and fascicle length (r = 0.35; P = 0.07). However, no significant correlations were observed between the improvements in muscle strength or size and mobility. Possibly of greater interest was that although a small but significant reduction in muscle thickness occurred during the detraining period, no significant reduction in strength occurred, indicative that neuromuscular adaptations (Cadore et al. 2014) were important for both the improvement, and then retention, of strength. Regardless, strength (~ 71–85% of the mean increase maintained) and muscle thickness (~ 62–71% of the mean increase preserved) remained significantly greater than baseline after the 8-week detraining period. These are the first data reporting the detraining regression profiles in older adults after a period of isokinetic eccentric training and are indicative of prolonged functional benefit. As limited regression in strength and muscle mass has been reported in younger adults following isotonic eccentric, but not concentric, training (Coratella and Schena 2016), the collective findings of the present study, and those of previous studies, are indicative that the substantial improvements in muscle mass and strength triggered by eccentric exercise are well maintained, with minimal regression after training cessation; this is important information relating to exercise prescription in older adults.

Older adults often present with a fear of falling, poor mobility, and limited aerobic capacity, reducing exercise tolerance and adherence to traditional forms of exercise (LaStayo et al. 2003) that can contribute to the development of frailty (Cesari et al. 2014). However, substantial improvements in muscle strength, size, and mobility were achieved in the present study despite the subjects reporting a low-to-moderate perceived exertion during the training. Furthermore, subjects in the present study were able to accurately achieve the required contraction intensity (i.e. 50% MVC) in > 90% of contractions and had an 89% programme completion rate, with those completing the programme recording 100% adherence to the training sessions. These data are indicative of a well-accepted and tolerated exercise modality, and are consistent with previous studies observing that eccentric exercise was well tolerated in older adults and other individuals with cardiorespiratory impairments (LaStayo et al. 2003; Rooyackers et al. 2003). The improved exercise tolerance is likely a consequence of eccentric contractions having a low metabolic demand, ~ 25% that of concentric exercise of a comparable workload (LaStayo et al. 1999), with the low demand reducing perceived effort and improving tolerance (LaStayo et al. 2003). Similarly, the seated exercise modality used in the present study also eliminated issues with poor mobility and fear of falling, which are common in older populations (LaStayo et al. 2003). Collectively, these findings are indicative that seated isokinetic eccentric exercise mitigates several limitations common to other exercise modes (i.e. aerobic exercise, traditional concentric–eccentric resistance exercise) that compromise older adults’ ability to exercise at a sufficient intensity to provide meaningful adaptations.

In summary, this is the first empirical study to quantitatively evaluate the effects of seated isokinetic eccentric resistance training and, importantly, a period of detraining, on neuromuscular function and musculoskeletal characteristics in older adults. While there were no changes in postural sway metrics following training, a significant reduction in TUG time and substantial increases in strength and muscle thickness occurred after 6 weeks of eccentric resistance training. Notably, these improvements remained significantly elevated after 8 weeks of detraining, indicating a prolonged beneficial effect with limited structural and no functional regression. Importantly, seated exercise removes fear of falling during the exercise and the eccentric modality results in lower cardiovascular demand, enabling subjects to exercise at a low RPE, important for improving exercise tolerance in both healthy and clinical populations. Collectively, these findings have important clinical and practical implications for exercise prescription in older adults that may be prone to neuromuscular and musculoskeletal decline and periods of inactivity, as the functional outcomes as well the exercise modality itself are well suited to the specific exercise needs and physical challenges of older adults. A limitation of the present study is that physical activity was not measured during the detraining period. The minimal regression observed during the detraining period may be a consequence of increased physical activity, maintaining the improvements in functional outcomes rather than a prolonged effect of eccentric training. Future research should compare activity profiles before training and during detraining periods to clarify the potential prolonged impact of eccentric training. A practical limitation of the study is that isokinetic dynamometry equipment is expensive and usually restricted to research or clinical facilities, limiting access for older adults. Future studies should identify alternative strategies to impose eccentric loading to determine the feasibility, acceptability, tolerance, adherence and efficacy in older adults to improve practicality, clinical application and outcomes for older adults.