Reliable and effective novel home-based training set-up for application of an evidence-based high-loading stimulus to improve triceps surae function

ABSTRACT

High-loading interventions aiming for muscle-tendon adaptations were so far implemented in on-site facilities. To make this evidence-based stimulus more accessible, we developed an easy-to-use sling-based training set-up for home-based Achilles tendon and triceps surae muscle strength training and assessed its reliability and effectiveness in healthy men. To assess reliability (n=11), intra-class correlation (ICC) and root mean square (RMS) differences of isometric maximum voluntary contraction (MVC) of the plantar flexors were used. Effectiveness was tested in a controlled intervention trial (n=12), applying one-legged high-loading intervention for 3 months with our mobile set-up, while the contralateral/untrained leg served as control, and assessing plantar flexor MVC, drop (DJ) and countermovement jump (CMJ) height. Reliability was excellent between (ICC^B=0.935) and within session (ICC^Ws=0.940–0.967). The mean RMS difference between and within sessions was 5.3% and 4.7%, respectively. MVCs of the trained/intervention leg increased by 10.2±7% (P=0.004) (dynamometry) and 30.2±22.5% (mobile set-up) (P=0.012). MVC of the untrained/control leg did not change (P>0.05). DJ height increased (P=0.025; D_z=2.13) by 2.37±2.9cm. CMJ height (P>0.05) did not change. We recommend the evidence-based high-loading application with our novel home-based training set-up as reliable and effective improving strength and jump performance of the plantar flexor muscle-tendon unit.

KEYWORDS:

• Injury prevention
• muscle imbalance
• musculoskeletal rehabilitation
• tendon rehabilitation

1 Introduction

To improve musculoskeletal tissue function of muscles and tendons, one approach is to utilise the mechanical stimulus leading to the most pronounced adaptation. To be able to apply this optimal stimulus, it is vital to first know the relationship between mechanical loading and tissue adaptation (Heinemeier & Kjaer, 2011), and second, to accurately control the applied mechanical load to ensure optimal adaptation. Regarding tendon adaptation, the characteristics of exercise stimuli and the associated tendon tissue response in healthy adults have been identified: High tendon strains of 4.5–6.5%, generated by muscle contractions at an intensity of ~90% of the isometric maximum voluntary contraction (MVC) and a stimulus duration of at least 3 s are essential for tendon adaptation (Arampatzis et al., 2007, 2010; Bohm et al., 2014). This high-loading exercise protocol, being applied as five sets of four repetitions four times per week, led to changes in mechanical (i.e. stiffness), morphological (i.e. cross-sectional area), and material properties (i.e. elastic modulus) of the tendon (Arampatzis et al., 2007, 2010; Bohm et al., 2014) and significant strength improvements of the plantar flexors (Albracht & Arampatzis, 2013; Arampatzis et al., 2010). Aforementioned changes may potentially reduce the risk for musculotendinous injuries (Mersmann et al., 2017). In terms of function, increased plantar flexor strength might enhance forward propulsion (Hamner et al., 2010; Liu et al., 2008), improve ankle joint motion control (i.e. during stair descent or jump landings) (Devita & Skelly, 1992; Van Dieën et al., 2008) and improve postural control by stabilising the ankle (Langeard et al., 2020; Ribeiro et al., 2009). Furthermore, higher plantar flexor tendon stiffness has been shown to be related to better jump (Bojsen-Møller et al., 2005) and stability performance (Karamanidis et al., 2008) and may improve running economy (Albracht & Arampatzis, 2013).

High-loading interventions aiming for muscle-tendon adaptations were so far implemented under laboratory conditions or in a gym with stationary equipment (i.e. weight training machines in on-site facilities) (Beyer et al., 2015; Bohm et al., 2014). To make this evidence-based stimulus accessible to a wider population, it needs to be applicable within a more practical setting, enabling its implementation at home or in physiotherapy facilities without stationary equipment. Providing a home exercise programme as an alternative to a facility-based training programme is imperative to physiotherapy (Kisner & Colby, 2007). It may improve acceptance and commitment due to more convenience. For people without access to gym facilities, e.g., in rural areas or due to pandemic consequences (Covid-19), home-based programmes are highly advantageous (Thiebaud et al., 2014). Indeed, therapeutic effectiveness of home-based programmes has been demonstrated for various musculoskeletal issues (Emery, 2005; Johansson et al., 2009; Keays et al., 2006; Littlewood et al., 2014). However, the lack of immediate control, feedback and correction by the therapist when compared to an on-site setting as well as instantaneous support concerning controlled progression of the therapy may impose limitations to a home-based exercise programme (Stasinopoulos et al., 2010). To meet these limitations and enhance its effectiveness, it may be advantageous, if a home-based exercise programme is applied with the help of initial educational on-site session(s), while the exact mechanical loading stimulus is easy to control for the patient or athlete. Values describing the magnitude, or the progression of the loading stimulus can then be reported to the therapist, allowing comprehensive control of the therapy.

Taking the aforementioned aspects into consideration, we developed a low-cost, simple to replicate and easy-to-use sling-based training set-up for home-based Achilles tendon and triceps surae muscle strength training. This training set-up allows easy administration of the evidence-based exercise stimulus at home, while enabling the users to control and adjust the specific load via a simple biofeedback.

The aim of our study was to test the reliability of the feedback fitted sling-based training set-up, as well as assessing the set-up in terms of its effectiveness in improving strength and vertical jump performance in healthy adults. Jump performance was considered as a specific measure to assess plantar flexor muscle function. We hypothesised that the training set-up allows reliable muscle strength assessment and that its home-based application increases plantar flexor strength and jump performance.

2 Materials and methods

2.1 A priori sample size analysis

To estimate the adequate sample size for the assessment of the intervention effectiveness, we conducted a power analysis (Faul et al., 2007) to test the difference between two dependent group means, using a two-tailed test, a large effect size referring to the training effect for strength (i.e. D_z = 1.15) and an alpha of .05. The effect size was calculated as a mean of the effect sizes of the training effects for plantar flexor strength improvements from two prior trials using the same high-loading training protocol (Albracht & Arampatzis, 2013; Arampatzis et al., 2007). Results showed that a total sample of 11 participants was required to achieve a power of .90.

2.2 Study design and participants

All participants provided informed written consent prior to the experiments, and local institutional ethics approval was obtained (Humboldt-Universität zu Berlin, Faculty of Humanities and Social Sciences). The study was performed in compliance with the Declaration of Helsinki.

2.2.1 Effectiveness

To test the effectiveness of the mobile training set-up (see 2.4), participants were subjected to a controlled 12-weeks long home-based exercise intervention trial in which one leg conducted a high-loading intervention (see 2.5), while the contralateral leg did not receive any intervention and thus served as control. Our intervention was conducted with 12 healthy male adults without any leg injuries in the past 12 months. One participant did not complete the training intervention due to an injury not related to the study and was therefore excluded. The anthropometric data of the remaining 11 participants were as follows: age 27.8 ± 7.1 years (range 22–44 years), body mass 75.1 ± 6.3 kg, height 181.2 ± 6.7 cm, body mass index (BMI) 23 ± 2 kg/m². All participants were physically active, but not involved in high-performance sports. Participants were allowed to maintain their previous individual training habits. However, no additional strength training of the plantar flexors and no implementation of any new sort of lower body strength training was permitted.

MVCs of the plantar flexors of the intervention/trained leg and the control/untrained leg were measured at baseline (PRE) and after completion of the intervention phase (POST) with a stationary lab dynamometer (see 2.3.1) and the mobile training set-up (see 2.3.2). In addition, the MVCs of the intervention/trained leg were monitored weekly with a stationary dynamometer. Countermovement jump (CMJ) and drop jump (DJ) height were assessed PRE and POST (see 2.3.3) (Figure 1).

Figure 1. Experimental design: Reliability assessment of the mobile training set-up and effectiveness assessment of the 12-week long intervention of an isometric high-loading home-based plantar flexor muscle strength protocol. BMI = body mass index, MVC = maximum isometric voluntary contraction, PRE = baseline, POST = after the intervention phase, DJ = drop jump, CMJ = countermovement jump

To compare the novel mobile set-up with the dynamometer (i.e. gold standard), we estimated the strength of the relationship between the MVC values measured with the dynamometer and the MVCs measured with the mobile training set-up applying a regression analysis (Bland & Altman, 2003).

2.2.2 Reliability

The reliability of the MVC measurements with the training set-up was tested in a separate experiment with five repeated measurement sessions within 2 weeks (Figure 1). Healthy male amateur athletes without any leg injuries conducted plantar flexor MVCs with their dominant leg on five different days with an interval of 48 hours between MVC sessions (n = 11: age 38.7 ± 11.2 years, body mass 82.0 ± 9.5 kg, height 181.1 ± 8.3 cm, BMI 25.0 ± 2.3 kg/m²). After the standardised warm-up (see 2.3.2), the participants conducted five MVCs with a 1-min rest between repetitions in each session.

2.3 Strength and jump performance assessment

2.3.1 MVC with dynamometer

Participants were seated on the dynamometer (Biodex-System 3, Biodex Medical Systems Inc., Shirley, NY, USA) in a standardised position with a hip (i.e. femur-to-spine) angle of 110°, extended knee and an ankle angle of 90°. The pelvis was fixed with a rigid belt at the dynamometer seat. Rotational axis of the ankle joint was individually aligned with the axis of the dynamometer during the contraction condition. Individual settings of each participant were recorded and used for the weekly and post-intervention measurements. After a warm-up with three sets of five isometric sub-maximal plantar flexor contractions, participants performed five MVCs with 1-min rest between repetitions.

2.3.2 MVC with mobile training set-up

The MVC measurements with the mobile set-up were similar to the MVC measurements on the dynamometer referring to warm-up and resting time. Standardised sitting position and use of the set-up were applied according to its general settings (see 2.4) (Figure 2). Each MVC was recorded manually by the examiner from the strain gauge display.

Figure 2. Novel training set-up: Mobile training belt-sling-system for evidence-based plantar flexor muscle-tendon unit training with an integrated strain gauge. Seating position with knee extended and ankle angle of 90°. The strap with the ratchet was individually adjusted downwards near the hip bone to reduce spinal load (↧)

2.3.3 Vertical jump performance

After a warm-up with up to 12 jumps of low to moderate intensity, five maximum effort CMJs and five DJs were performed with bare feet and 1-min rest between repetitions. Ground reaction forces were measured with two separate force plates at a rate of 1000 Hz (Kistler, Type 9260AA, 600 × 500 × 50 mm, Switzerland) linked to an analogue digital converter (DAQ-System, USB 2.0, Type 5691A1). Data were recorded (BioWare Software, Type 2812A) and jump height was calculated using a custom written Matlab interface (version R2012a; MathWorks, Natick, MA, United States). For the CMJ, the hands were akimbo, while no specifications were given regarding depth. For CMJ height calculation, we used the impulse momentum method (Linthorne, 2001). DJs were performed from a 15 cm box. Participants were asked to jump with maximum effort, keeping the contact time as short as possible while jumping as high as possible. DJ height calculation was based on the flight-time method (Moir, 2008).

2.4 Novel mobile sling-based training set-up

The mobile set-up consisted of the following components: a belt (Neoprene padded powerlifting belt 11.5 cm width, RDX, US) providing counter resistance on the hip, a ratchet strap (automatic ratchet strap, EAN 4260246942700, capacity of 600 kg, 1.85 m length, Iapyx) for adequately tightening the sling-belt-system with one hand, and a foot plate (aluminium stirrup, 350 g, AMKA) for rigid force transmission. The neoprene material of the underlying powerlifting belt provided some cushioning for the overlying ratchet strap. In order to combine the foot plate with the strain gauge, we used carabiners (Carabiner Micro 3, Edelrid) with a polyamide quickdraw in between (Quickdraw Pad 19 mm, Ocun). To allow MVC measurements and real-time feedback and control of the applied load, the mobile set-up was fitted with a strain gauge (HS-70, Voltcraft, Germany) displaying the force (kg) applied to the system (Figure 2).

Participants were advised to sit with extended knees. The forefoot (with shoes) was placed in the foot plate with the widest part of the shoe (i.e. ball of the foot) positioned in the sagittal centre of the foot plate pad to ensure a standardised contact point. The participants were advised to always use the same shoes with a rigid sole. The ratchet was individually set and fixed as tightly as possible, to allow for maximal isometric plantar flexor contractions at a standardised ankle angle position (90°) (Figure 2). The straps on both sides of the pelvis were placed as close as possible to the hip bone (i.e. os ilium) to reduce spinal load. It was recommended to place one hand on the ground behind the body providing a hip (i.e. femur-to-spine) angle of >90°. For standardised alignment during the contractions, both legs had to be parallel in the frontal plane and in neutral rotational position in the hip joints in the transverse plane.

2.5 Intervention

The leg with the lower baseline MVC was chosen as the training leg, while the contralateral leg served as control. The first training session was supervised by one examiner. The evidence-based home-based training consisted of four training sessions per week for 12 weeks. Each training session consisted of five sets with four repetitions each and a 1-min rest between sets. Each repetition was a ~ 90% MVC isometric plantar flexor contraction held for 3 s followed by 3 s of rest according to previous studies (Arampatzis et al., 2007, 2010; Bohm et al., 2014). Based on the PRE MVC values measured with the mobile training set-up, the initial training value (i.e. 90% of the PRE MVC values mean) was calculated. During intervention training, participants were instructed to observe and track the contraction-induced force magnitude on the display of their mobile set-up (Figure 2) in order to achieve their calculated training load. To control for delayed onset muscle soreness, any increase in training intensity was prohibited for the first 2 weeks. From the third week onwards, the training value was re-calculated based on a weekly MVC test. The intervention training was monitored during the weekly MVC tests and via telephone calls to ensure that the exercise programme was conducted as planned throughout the intervention phase. The participants used a timer app with audio signal function on their mobile phone to help implement the contraction frequency and contraction duration of 3 s.

2.6 Data and statistical analysis

The a priori power analysis was conducted using G*Power 3 (Faul et al., 2007). All further statistical analyses were performed using IBM SPSS Statistics software for Windows, Version 21.0 (Armonk, NY, IBM Corp). For CMJ, DJ and MVC analysis, the mean of the middle three out of five attempts was used. To assess normal distribution of the data, a Kolmogorov-Smirnoff test was used. In case of normal distribution, paired T-tests were applied to establish effectiveness in terms of Pre to POST comparisons for CMJ, DJ and MVC measures. The effect size concerning the effect of training on MVC, DJ and CMJ height was calculated by Cohen's D_z and defined as D ≤ 0.2 small, D ≤ 0.5 medium, and D ≤ 0.8 large effect (Lakens, 2013).

For assessing reliability, four different methods for mean calculation were applied and compared, to identify the most reliable approach: Mean of all five measurements per session (Mean), the maximum measurement per session (Best), mean of the three maximum measurements per session (Ø3Best) and mean of the middle three measurements per session (Mid 3). A two-way mixed, single measure, absolute agreement intra-class correlation (ICC 3.1) was performed (Shrout & Fleiss, 1979). ICC results were interpreted based on the classification scale: excellent (0.90–1.00), good (0.75–0.90), moderate (0.50–0.75), and poor (<0.50) (Rosner, 1982). Between session reliability (ICC^B) and corresponding RMS differences were calculated based on the mean MVC values of each participant per session over all sessions and based on four different MVC mean calculation approaches. Within session reliability (ICC^W) and corresponding RMS differences were calculated based on every single MVC value per session and participant.

To assess the strength of the relationship between the PRE and POST MVC values of the trained and the untrained leg measured with the dynamometer (i.e. independent variable) and the PRE and POST MVC values of the trained and the untrained leg measured with the mobile training set-up (i.e. dependent variable), a linear regression model was used for analysis (Bland & Altman, 2003). The interpretation of the regression coefficient (R) was: very weak (0.00–0.19), weak (0.20–0.39), moderate (0.40–0.59), strong (0.60–0.79), very strong (0.80–1.0) (Evans, 1996). For all statistical tests, significance was established at an alpha level of .05. For all statistical procedures concerning PRE to POST plantar flexor MVC measures due to multiple comparisons of possible intervention effects, Bonferroni adjustments were made by the number of tests (n = 16) performed.

3 Results

3.1 Reliability of the mobile sling-based training set-up

The training set-up demonstrated excellent reliability both between sessions (ICC^B) and within session (ICC^W) (TABLE 1). Independent of the mean calculation method, the calculated ICC^B mean values showed high similarity with a range of 0.927–0.938. Mean RMS difference was 5.3% between sessions and 4.7% within session (TABLE 1).

Table 1. Reliability of the mobile training set-up: ICC^B and RMS differences of mean values of the MVCs between measurement sessions (i.e. over all 5 time points T1 – T5); and ICC^W and RMS differences of the MVCs within one measurement session (i.e. over all MVCs per time point based on 5 MVCs per time point and participant) using the mobile training set-up. Four different calculation methods of the mean MVC (i.e. Mean, Best, Ø3Best and Mid 3) (n = 11) were applied to identify the best approach

n = 11	Mean±SD (kg)	Best±SD (kg)	Ø3Best±SD (kg)	Mid 3±SD (kg)	ICC^W	95% CI	RMS±SD (kg)
T1	59.0 ± 14.9	63.0 ± 14.9	61.0 ± 14.9	58.9 ± 14.9	0.949*	0.887–0.984	2.9 ± 1.1
T2	61.6 ± 17.6	65.3 ± 17.2	63.5 ± 17.4	61.7 ± 17.5	0.967*	0.924–0.990	3.4 ± 1.0
T3	61.9 ± 16.1	66.1 ± 18.5	63.9 ± 17	61.7 ± 15.8	0.944*	0.877–0.982	3.0 ± 1.9
T4	64.4 ± 16.1	69.5 ± 18.6	66.3 ± 16.7	63.8 ± 15.6	0.940*	0.867–0.981	3.1 ± 2.0
T5	65.6 ± 15.5	68.9 ± 15.8	67.3 ± 15.8	65.5 ± 15.6	0.967*	0.925–0.990	2.4 ± 0.9
Mean	62.5 ± 2.3	66.6 ± 2.4	64.4 ± 2.2	62.3 ± 2.2	0.953
ICC^B	0.935*	0.927*	0.938*	0.933*
95% CI	0.840–0.980	0.829–0.977	0.848–0.981	0.839–0.979
RMS±SD (kg)	3.3 ± 1.7	3.8 ± 1.7	3.3 ± 1.7	3.4 ± 1.7

Abbreviations: ICC^B = Intra-class correlation of mean values of the isometric maximum voluntary contractions (MVCs) over all 5 sessions (T1-T5); ICC^W = Intra-class correlation of mean values of the MVCs within one session; kg = kilogram; Mean = average of all 5 measured MVC values; Best = highest MVC value; Ø3Best = average of highest 3 MVC values; Mid 3 = mean of the middle 3 MVC values, excluding the highest and the lowest value; RMS = root mean square differences; SD = standard deviation; T = time point; 95% CI = 95% confidence interval. * indicates significant difference (P < .001).

3.2 Effectiveness of training

One participant (age 23 years, body mass 71.1 kg, height 174.0 cm, BMI 23.5 kg/m²) was excluded from the trial and any data analysis due to an injury not related to the trial. Plantar flexor MVCs of the trained leg measured with the dynamometer increased during the intervention period, being significantly different (P < .05) to the start of the intervention from the 5^th week onwards (Figure 3).

Figure 3. Time course of plantar flexor strength increase: Isometric maximum voluntary contraction (MVC) plantar flexor strength of the intervention/trained leg during the 12-week intervention period presented as percent increase compared to baseline (i.e. PRE). POST = week 12. Data are presented as mean ± standard deviation (except PRE showing 100% for all participants with no standard deviation). * indicates significant difference (P < .05) compared to baseline

From PRE to POST, MVCs of the intervention/trained leg increased significantly (P = .004) by 10.2 ± 7.0% when measured with the dynamometer (Cohen D_z = 1.74) and by 30.2 ± 22.5% (P = .012) when measured with the sling-based training set-up (Cohen D_z = 1.43) (Figure 4). MVCs of the control/untrained leg measured with either the dynamometer (P > .05) or the training set-up (P > .05) did not significantly change from PRE to POST.

Figure 4. PRE-POST comparison of plantar flexor strength of trained (intervention) and untrained (control) leg: Mean torque values (Nm) measured with the dynamometer (dynam.) and mean force values (kg) measured with the strain gauge of the mobile training set-up of the plantar flexors of the trained (intervention) leg and the untrained (control) leg before (PRE) and after (POST) the intervention phase. Data are presented as mean ± standard deviation. * indicates significant difference (P < .05) compared to PRE value

Measures by the dynamometer and measures by the sling-based training set-up demonstrated a strong linear positive relationship between (R = 0.704) (P < .05) (Figure 5).

Figure 5. Relationship of MVC values measured by dynamometry and mobile set-up: Linear regression analysis for all baseline (PRE) and three months after (POST) isometric maximum voluntary contraction (MVC) values of the trained (intervention) and untrained (control) leg measured with the dynamometer in Newton metres (Nm) to all PRE and POST MVC values of the trained (intervention) and untrained (control) leg measured with the mobile training set-up in kilogram (kg). Each data point represents the mean for each leg, being calculated using the middle three of five measures; R² = coefficient of determination; * indicates significant difference (P < .05)

DJ height did significantly change from PRE to POST (P = .0250) indicating a large training effect (Cohen D_z = 2.13) with a jump height difference of 2.37 ± 2.9 cm. CMJ height did not significantly change from PRE to POST (P > .05) (Figure 6).

Figure 6. PRE-POST comparison of vertical jump height: Mean jump height values of two different jump manoeuvres (CMJ = countermovement jump; DJ = drop jump) measured at baseline (PRE) and after the intervention (POST). Data are presented as mean ± standard deviation. * indicates significant difference (P < .05)

4 Discussion

As hypothesised, our study showed excellent reliability of the feedback fitted sling-based training set-up. Further, the home-based plantar flexor training was effective, leading to strength gains of 10.2 ± 7.0% (measured by dynamometry) in healthy adults. Regarding improvements in function (i.e. vertical jump performance), our hypothesis was confirmed as DJ height increased from PRE to POST by 9.6 ± 19%.

4.1 Reliability

While reliability and validity of the strain gauge on its own is warranted per certificate, we investigated reliability of the whole mobile sling-based training set-up. The excellent reliability in terms of ICC within and between MVC measurement sessions provided by the feedback mechanism (strain gauge) (TABLE 1) enables the user to precisely control the mechanical stimulus, and to adjust the load in a progressive manner depending on strength gain. Excellent reliability was confirmed with the small within and between session RMS difference of 4.7% and 5.3%, respectively. In this regard, coaches, therapists, and users can rely on the values given by the strain gauge in terms of exact application of the stimulus and adequate planning of exercise progression. Excellent reliability was also supported by the fact that we did not find any difference between the four different calculation methods determining the mean of the MVCs. This indicates a lack of outliers and a continuous MVC performance throughout the five MVC repetitions. Further, this suggests that one MVC measurement is sufficient in determining the evidence-based training load of 90% MVC. We conclude that once the set-up is individually set according to the initial and standardised instructions, it allows reliable application over time.

4.2 Effectiveness

The effectiveness of our home-based exercise protocol applied with our novel mobile set-up was indicated by significant increases in plantar flexor strength (Figure 4) from week 5 onwards (Figure 3). These positive effects were also reported in studies, which applied the same plantar flexor training protocol for 14 weeks of training at 90% MVC under controlled conditions in a lab on a stationary dynamometer (Albracht & Arampatzis, 2013; Arampatzis et al., 2007). As the home-based application was not inferior to exercising under laboratory conditions, it provides a sufficiently effective alternative training set-up.

Statistically significant PRE to POST differences for the DJ height by 9.6 ± 19% indicated an improvement in jump performance as a specific measure of plantar flexor function (Figure 6). Considering that the training was unilateral, larger effects may be expected with bilateral training. CMJ height did not change with plantar flexor training. However, as the plantar flexors contribute little to CMJ height with weak correlations between plantar flexor MVC and CMJ height being reported (Tsiokanos et al., 2002), the detected increase in plantar flexor strength may not necessarily lead to measurable changes in CMJ height. In sum, the application of high-loading exercise might have the potential to improve jumping capacity in movements performed with a larger contribution of the plantar flexors.

The novel training set-up was highly feasible, if feasibility is defined as the extent to which a planned intervention protocol is fully realised by the participants as planned without impeding incidents reported (Wang et al., 2006). Participants reported no adverse effects or impeding incidents throughout the course of intervention. Thus, the exercise protocol was implemented as planned and safe.

Intervention-related effects on muscle strength and jump performance varied considerably between individuals in our study with strength improvements ranging from 2.9% (i.e. moderate response) to 29% (i.e. strong response). High variability in response to exercise programmes is frequently reported in the literature (Bohm et al., 2019; Hecksteden et al., 2018) and may be explained by inter-individual variations (Hubal et al., 2005) in training experience, anthropometric or genetic characteristics (Atkinson & Batterham, 2015). Moreover, the characteristics of the home-based protocol allow variation in individual compliance rates and differences in exercise execution, which may be considered as another factor that might affect outcome variability. Thus, when applying this home-based protocol, practitioners need to be aware that the training effect may vary considerably between individuals.

4.3 Limitations

For the regression analysis, we included dependent variables (i.e. PRE and POST). Since all participants contributed with the same number of measurements, the effect of dependency should not be severe. Despite a strong linear positive relationship between MVC measures established with the dynamometer and with the mobile set-up (Figure 5), the extent of the PRE to POST plantar flexor strength increase varies between both methods. Thus, we do not consider the novel set-up to represent a laboratory dynamometer substitute in terms of a precise strength and performance assessment tool. However, measures established by the mobile set-up confirmed the positive adaptations measured by the dynamometer. Hence, our mobile set-up is capable of displaying and monitoring individual strength gains.

As jumping performance was considered as a specific assessment of function, transferability to any further functional aspects of the plantar flexor muscle-tendon unit is limited. Moreover, as our participants were advised to execute the exercises with extended knees, our results might not be equally transferable to the application of this protocol with bended knees. We decided for males only which has to be considered when transferring our results to females. However, as strength training interventions may lead to larger relative strength improvements in females (Hubal et al., 2005), we would expect similar or larger improvements in plantar flexor strength and function in females following our intervention protocol.

As we used the contralateral leg of each participant as control, the control leg may have been affected by cross-educational or systemic effects.

4.4 Implications for practice

This easily accessible mobile training set-up provided reliable values to control the evidence-based stimulus and demonstrated effectiveness for home-based strength training of the plantar flexors. As the joint angle–moment relationship has an impact on strength development (Gordon et al., 1966; Moo et al., 2020), users have to observe carefully that an adequate tightness or tension of the belt and the standardised 90° of ankle angle during the contraction are strictly adhered. Any positional change of the shoe onto the foot plate alters the length of the lever arm and thus has an impact on the peak torque magnitude. Therefore, strict adherence must also be applied regarding the standardised position of the shoe within the foot plate. Advising the participants to perform the exercises with an ankle angle of 90° and with extended knees seemed advantageous for several reasons. Firstly, it might be simple and easy to replicate especially when using the mobile training set-up at home or in a clinical setting. Second, maximum plantar flexor moments have been shown to be achieved in close-by positions (i.e. 85° ankle angle) (Arampatzis et al., 2006).

In healthy adults or athletes, this evidence-based protocol may contribute to improve plantar flexor strength (Albracht & Arampatzis, 2013; Arampatzis et al., 2007). Its application may reduce muscle strength imbalances between legs, which are observed after injury even long after the rehabilitation process is completed (Alfredson et al., 1996). Furthermore, the intervention protocol may improve athletic performance as enhanced running economy has been observed (Albracht & Arampatzis, 2013). Accompanying increases in plantar flexor strength, a training programme that utilises muscle contractions at an intensity of ~90% of MVC to induce high tendon strains has been shown to improve Achilles tendon stiffness (Kubo et al., 2012; Arampatzis et al., 2007, 2010; Bohm et al., 2014). Regarding the patellar tendons, similar effects have been described (Kubo et al., 2001; Kongsgaard et al., 2007). As our home-based protocol similarly utilised muscle contractions at ~90% MVC, which led to significant strength improvements of the plantar flexor muscles, positive alterations in tendon stiffness may be expected (Muraoka et al., 2005). Increased stiffness means less strain at a given load. As strain is a key predictor for overuse tendon injury (Obst et al., 2018; Wren et al., 2003), this approach of high-loading exercise may have the potential of reducing muscle-tendon imbalances (Arampatzis et al., 2020), preventing tendon overload injuries (Bohm et al., 2019; Muraoka et al., 2005) and rehabilitating tendinopathic tendons (Radovanović et al., 2019).

As minimal equipment (i.e. mobile set-up), minimal facilities and staff (i.e. initial on-site supervision and subsequently on demand remote monitoring) was required, our home-based protocol was a low-cost programme and thus might meet economical needs (Ribeiro et al., 2009).

4.5 Conclusions

Based on our results, we recommend this home-based set-up and the evidence-based high-loading exercise protocol to be implemented for training and treatment of the triceps surae muscle-tendon unit. This set-up meets economical needs and everyday use challenges. The mobile character provides additional value, particularly if stationary gym facilities cannot be accessed or time flexibility is needed. The highly reliable character of the set-up allows to control the adequate training stimulus. It may support athletes aiming to enhance plantar flexor performance of the lower leg, or patients suffering from muscular imbalances due to an injury or post-surgery. Due to previously reported adaptive effects in healthy tendons associated with this specific training protocol, our approach may have the potential to contribute to Achilles tendon injury prevention and rehabilitation.

Acknowledgments

We would like to thank Vanyo Tanchev for contributing to data collection, Arno Schroll for statistical and Lars Janshen for methodological advice.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

We acknowledge support by the Open Access Publication Fund of Humboldt-Universität zu Berlin.