Shorter gastrocnemius fascicle lengths in older adults associate with worse capacity to enhance push-off intensity in walking

Abstract

Background:

Reduced push-off intensity during walking is thought to play an important role in age-related mobility impairment. We posit that an age-related shift toward shorter plantarflexor operating lengths during walking functionally limits force generation, and thereby the ability of those muscles to respond to increased propulsive demands during walking.

Research Question:

To determine whether gastrocnemius muscle fascicle lengths during normal walking: (1) are shorter in older than young adults, and (2) correlate with one’s capacity to increase the propulsive demands of walking to their maximum?

Methods:

We used in vivo cine B-mode ultrasound to measure gastrocnemius fascicle lengths in 9 older and 9 young adults walking at their preferred speed, their maximum speed, and with horizontal impeding forces that increased in a ramped design at 1%BW/s to their maximum. A repeated measures ANOVA tested for effects of age and walking condition, and Pearson correlations assessed the relation between fascicle outcomes and condition performance.

Results:

A tendency toward shorter medial gastrocnemius muscle fascicle lengths in older versus young adults was not statistically significant. However, older adults walked with reduced peak fascicle shortening during all conditions compared to young adults – an outcome not explained by reduced muscle-tendon unit shortening and exacerbated during tasks with greater than normal propulsive demand. As hypothesized, we found a strong and significant positive correlation in older subjects between gastrocnemius fascicle lengths during normal walking and performance on the ramped impeding force condition (p=0.005, r²=0.704), even after controlling for isometric strength (p=0.011, r²=0.792) and subject stature (p=0.010, r²=0.700).

Significance:

Our findings provide muscle-level insight to develop more effective rehabilitation techniques to improve push-off intensity in older adults and assistive technologies designed to steer plantarflexor muscle fascicle operating behavior during functional tasks.

Introduction

The plantarflexor muscles provide the majority of mechanical power needed for forward propulsion during the push-off phase of walking [1–3]. Accordingly, reduced push-off intensity during walking, arising from diminished plantarflexor mechanical output, is thought to play an important role in age-related mobility impairment [4–6]. We recently discovered that older adults perform significantly worse than young adults on tasks designed to increase the propulsive demands of walking to their maximum not simply during maximum speed walking, but also at preferred speed using a maximum ramped impeding force condition [7]. We posit that an age-related shift toward shorter plantarflexor operating lengths, perhaps governed in part by increased tendon compliance, functionally limits force generation and thereby the ability of those muscles to respond to increased propulsive demands during walking. This study represents an initial but important step toward understanding the veracity of this overarching premise.

Elderly gait is characterized by hallmark reductions in push-off intensity and thus forward propulsion. Specifically, older adults habitually walk with deficits in peak anterior ground reaction forces and ankle joint kinetics compared to young adults, which are considered relevant to reduced stride lengths and walking speed. [4–6, 8]. The plantarflexor (i.e. gastrocnemius and soleus) muscles are mostly responsible for this forward propulsion provided during walking [2, 4, 9]. Growing evidence is beginning to implicate age-related changes in plantaflexor muscle length-tension behavior, independent of deficits in force-generating capacity, in governing walking performance [10–12].

Recent advances in ultrasound image analysis have enabled in vivo plantarflexor muscle fascicle tracking during functional activities such as walking [13–15]. However, only a small number of previous studies directly compared the plantarflexor muscle operating behavior of both young and older adults during walking. At matched speeds, the soleus muscle fascicles of older adults are shorter across the gait cycle compared to their young counterparts, and undergo less relative shortening [12]. However, the contractile behavior of the uniarticular soleus muscle may differ from the biarticular gastrocnemius muscles, which potentially play a larger role in governing forward propulsion [1, 2]. Older adults gastrocnemius fascicles have also been found to remain shorter during the gait cycle compared to young adults [10]. However, how the medial gastrocnemius muscle accommodates increased propulsive demand at the fascicle level, and how this may change with age, is currently unknown.

Evidence that older adults’ plantarflexor muscle fascicles operate at shorter lengths [10, 12] alludes to the potential for deficits in force production per the force-length properties of muscle [16, 17]. For example, if aging muscle fascicles are habitually operating further down the ascending limb, force generating capacity would be limited compared to young muscle fascicles for the same unit activation. Indeed, in older adults, resting fascicle lengths of the medial gastrocnemius have been correlated with functional walking performance measures such as the 6-minute walk test [11]. Therefore, our overarching hypothesis is that older, shorter muscle fascicles would be less able to respond to increases in force (i.e. task) demands, with potential consequences in the community such as walking uphill, accelerating, and/or stair ascent.

In order to investigate whether plantarflexor muscle operating lengths are a limiting factor in aging gait, and to gain an improved muscle-level understanding of the mechanisms responsible for generating propulsion during walking, we increased the propulsive demands of walking to their maximum in young and older adults via maximum speed walking and a maximum impeding force protocol at preferred speed. We hypothesized that gastrocnemius muscle fascicle length at peak ankle moment during walking would be shorter in older than young adults. We also hypothesized that, independent of muscle strength, shorter fascicle lengths during normal walking would predict (i.e. correlate with) worse performance on tasks that increase the propulsive demands of walking to their maximum.

Methods

9 healthy young (YA, age: 25.3±4.9 years, height: 1.71±0.1 m, body mass: 71.6±8.4 kg, 5M/4F) and 9 healthy older subjects (OA, age: 75.3±2.7 years, height: 1.72±0.1 m, mass: 71.2±10.9 kg, 5M/4F) participated. The protocol was approved by the UNC Biomedical Sciences Institutional Review Board and all subjects provided written informed consent before participating. Subjects had no neurological disorders or disease, had not suffered neurological or musculoskeletal injury in the previous 6 months, and could walk without an assistive device.

We first recorded subjects’ preferred and maximum safe walking speeds as the average of three times taken to traverse the middle 2 m of a 10 m walkway. Subjects then walked on an instrumented dual belt treadmill (Bertec Corp., Columbus, OH) at their preferred walking speed for 5 min to pre-condition the plantarflexor muscle-tendon units [18] and allow their movement patterns to stabilize. Subjects then performed a series of ramped isometric voluntary contractions at a neutral (i.e., 0°) ankle angle in a dynamometer (Biodex, Shirley, NY). The knee was flexed to ~20° to replicate that near the push-off phase of walking, while straps over the foot and thigh limited extraneous motion. We verbally encouraged subjects to reach their maximum effort for each of two, 4-second ramped contractions separated by at least one minute. We defined subjects’ maximum isometric ankle moment as the peak generated during the isometric contraction, averaged across the two repetitions.

We then collected a normal, baseline trial while subjects walked at their preferred speed on the treadmill for 1 minute (“Pref”). After a 5-minute rest, we used a custom, feedback-controlled, motor-driven horizontal impeding force system (Fig. 1). The details of this system are described in detail elsewhere [7, 19]. Briefly, the system consists of a servo motor (Kollmorgen, Radford, VA) controlled in real-time using a LabVIEW interface (NI PCI 7352, National Instruments, Austin, TX) based upon instantaneous signals recorded from a load cell (Futek, Irvine, CA) in series with a cable connection to a waist belt worn by the subjects. Using this system while subjects walked at their preferred speed, we applied a horizontal impeding force that increased at a rate of 1 %BW/s (“Ramp”), thereby increasing the propulsive demands of walking to their maximum. Subjects were instructed to avoid excessive forward lean and to maintain their position on the treadmill until they could no longer sustain the impeding forces. Subjects received verbal encouragement to continue walking for as long as possible in the presence of the impeding forces. Similar to our earlier study [7], this Ramp trial ended after an inexorable 0.35 m displacement of the subjects’ pelvis, monitored in real-time by the motor’s encoder. Finally, subjects completed a walking trial on the treadmill which started from rest before increasing to their maximum safe overground walking speed (“Fast”) at a constant rate of 0.2 m/s², which was sustained for at least 5 seconds for data collection.

Figure 1.

Schematic showing experimental approach to increasing the propulsive demands of walking. In the maximum impeding force trial, subjects wore a waist belt that connected horizontally via a stainless-steel cable to a feedback controlled, motor-driven, impeding force system capable of prescribing horizontal impeding forces according to instantaneous measurements from a load cell. Specifically, we used a novel ramped impeding force protocol (Ramp) that increased at a rate of 1 %BW/s until the subjects reached the end point criterion, an inexorable 0.35 m posterior displacement of the subject’s pelvis.

For treadmill trials, 3D trajectories of 31 retroreflective markers recorded at 100 Hz on the pelvis and lower limbs were recorded using a 14-camera motion capture system (Motion Analysis Corp. Santa Rosa, CA) while 3D ground reaction force (GRF) data were simultaneously recorded at 1000 Hz. Data were then filtered using 4^th order low-pass Butterworth filters with cutoff frequencies of 6 Hz (marker data) and 100 Hz (GRF data). We used a standing trial and functional hip joint centers from right and left leg circumduction tasks [20] to scale a seven segment, 18 degree-of-freedom model of the pelvis and lower limbs [21]. Finally, we used these scaled models and an inverse dynamics routine described in detail previously to estimate leg joint kinematics, muscle tendon unit (MTU) lengths, moments, and powers [22]. For Pref, we averaged time-normalized right leg outcome measures across the 1-minute trial. For Ramp and Fast, we found the time-normalized right leg stride associated with the peak anterior ground reaction force and averaged this with the previous 3 strides.

During all trials, a 60 mm linear array ultrasound transducer (Echoblaster 128, 10 MHz, Telemed, Vilnius, Lithuania) was placed over the mid-belly of the right medial gastrocnemius which recorded B-mode images at 61 frames/s through an image depth of 65 mm. We used open source MATLAB routines [14] based on an optic flow algorithm to analyze the cine B-mode images. Following previously outlined techniques [13], the same investigator tracked all muscle data by first defining a region of interest surrounding each muscle and their aponeuroses before defining a gastrocnemius muscle fascicle in the mid-belly of the image from the superficial to deep aponeuroses. Manual corrections were made following visual inspection. These MATLAB routines then quantified time series of muscle fascicle lengths averaged across 3 strides, the derivative of which provided fascicle shortening velocity across the gait cycle or contraction. For the Ramp and Fast conditions, we analyzed strides during the last 5 seconds of each trial.

Our walking outcome measures were peak ankle moment, medial gastrocnemius fascicle length at peak ankle moment, peak fascicle length change, and muscle-tendon unit length change. Our secondary outcome measures were preferred and maximum walking speeds, performance on the impeding force Ramp condition, and isometric strength. For primary outcome measures, a repeated-measures analysis of variance (rmANOVA) tested for significant main effects of condition (Pref, Ramp, Fast) and age (young, older). For significant main effects, post-hoc pairwise comparisons identified the effects of age at each condition using an alpha level of 0.05. For secondary outcome measures, an independent samples t-test statistically compared older vs. young adults using the same alpha level. Finally, we calculated Pearson correlation coefficients between gastrocnemius fascicle measurements and performance on the Ramp (i.e., %BW) and Fast conditions (i.e. maximum speed).

Results

Older and young adults walked overground at similar preferred speeds (YA: 1.32±0.16 m/s, OA: 1.28±0.13 m/s, p=0.557, d=0.283). However, older subjects reached a 25% slower maximum walking speed (YA: 2.51±0.29 m/s, OA: 1.89±0.17 m/s, p<0.001, d=2.661, Fig. 3A) than young subjects, and performed 35% worse on the Ramp condition (p=0.036, d=l.081, Fig. 3C). We also found no significant difference in average isometric strength (p=0.125, d=0.762) nor peak ankle moment during preferred speed walking between young and older adults (p=0.744, d=0.121, Table 1). However, a significant age×condition interaction (p=0.003, η_p²=0.514) revealed that peak ankle moment was significantly lower in older adults for Ramp (−23%, p=0.017) and Fast (−20%, p=0.010).

Figure 3.

(A) Individual and group average (standard deviation) performance on the Ramp condition as percent bodyweight (% BW) for young and older adults. (B) Bivariate correlations between performance in the Ramp condition and gastrocnemius fascicle length at peak ankle moment from preferred walking for older (grey) and young (black) adults. (C) Individual and group average (standard deviation) maximum (i.e. Fast) walking speed for young and older adults. (D) Bivariate correlations between maximum (i.e. Fast) walking speed and gastrocnemius fascicle length at peak ankle moment from preferred walking for older (grey) and young (black) adults. Lines in B and D represent best-fit linear regressions. Single asterisks (*) represent a significant age effect (p<0.05). Double asterisks (**) represent a significant correlation (p<0.05).

Table 1.

Group mean ± standard deviation peak medial gastrocnemius fascicle shortening velocity and shortening velocity at the instant of peak ankle moment.

	Peak Shortening Velocity (cm/s)		Shortening Velocity at Peak Ankle Moment (cm/s)
	Young	Older	Young	Older
Pref	3.22 ± 0.53	1.95 ± 0.67	2.22 ± 0.76	1.33 ± 1.21
Fast	3.47 ± 1.02	2.74 ± 0.83	2.61 ± 0.73	1.51 ± 0.63
Ramp	3.67 ± 1.31	2.40 ± 0.86	3.22 ± 1.40	1.55 ± 0.98
Age	0.021*		0.005*
Condition	0.089		0.261
Age×Condition	0.175		0.401

Age, Condition, and Age×Condition effects reported as p-values for each outcome measure.

Single asterisks (*) represent a significant effect (p<0.05).

During Pref, gastrocnemius fascicle lengths at peak ankle moment in older adults averaged 7% shorter than those in young, however this difference was not statistically significant (p=0.189, Fig. 2A). A significant main effect (p=0.009, η_p²=0.600) revealed that gastrocnemius fascicle shortening relative to heel-strike was significantly diminished by age, and that this was exacerbated during the Fast (vs. YA, −43%, p=0.017) and Ramp (vs. YA, −65%, p=0.001) conditions relative to Pref (vs. YA, −29%, p=0.092) (Fig. 2B). This age-related reduction in peak shortening at the fascicle level was not explained by differences in MTU length (Fig. 2C–D). Significant main effects (p=0.021, η_p²=0.506) revealed that peak gastrocnemius shortening velocity was significantly slower in older than young adults, especially during the Pref (vs. YA, - 65%, p<0.001), and Ramp conditions (vs. YA, −53%, p=0.028, Table 1, Supplementary Figure 1). Similarly, a main effect (p=0.005, η_p²=0.644) revealed slower peak shortening velocity at the instant of peak ankle moment for older adults, particularly during Fast (vs. YA, −72%, p=0.004) and Ramp conditions (vs. YA, −107%, p=0.010, Table 1, Supplementary Figure 1).

Figure 2.

Muscle fascicle and muscle-tendon unit outcomes for preferred (Pref), maximum ramp (Ramp), and maximum speed walking (Fast) for both young (solid lines) and older adults (dashed lines), plotted against an average gait cycle, from heel-strike to heel-strike. Group average (A) absolute and (B) change (Δ) in medial gastrocnemius fascicle lengths. Group average (C) absolute and (D) change (Δ) in gastrocnemius muscle-tendon unit lengths. Shaded regions indicate the timing of peak ankle moment across our cohort. Single asterisks (*) represent local minima with significant age effect (p<0.05).

Across our study cohort, we found that gastrocnemius fascicle length at peak ankle moment during Pref exhibited a significant moderate correlation with performance on the Ramp condition (p=0.020, r²=0.293, Fig. 3B). This relation was driven by that in older adults (p=0.005, r²=0.704) and remained significant after controlling for isometric strength (OA: p=0.011, r²=0.792) and subject height (OA: p=0.010, r²=0.700). Conversely, we found no correlation between fascicle length at peak ankle moment during Pref and maximum speed (All: p=0.134, r²=0.135. OA: p=0.759, r²=0.014) or change to maximum speed (i.e. Fast-Pref, All: p=0.096, r²=0.164. OA: p=0.172, r²=0.249).

Discussion

We investigated muscle-level mechanisms responsible for performance on walking tasks designed to increase the propulsive demands of walking to their maximum in older compared to young adults. Our findings did not support our first hypothesis; the tendency toward shorter medial gastrocnemius muscle fascicle lengths at peak ankle moment in older versus young adults was not statistically significant. However, in partial support of our second hypothesis, shorter medial gastrocnemius fascicle lengths during normal walking did correlate with worse performance on our maximum ramped impeding force protocol performed at subjects’ preferred walking speed. This relationship was particularly dominant in older adults, even after controlling for plantarflexor muscle strength or subject stature. In contrast, independent of age, gastrocnemius fascicle lengths during habitual walking appeared unrelated to maximum walking speed.

Previous work has shown that older adults’ fascicle lengths are generally shorter than their younger counterparts [10, 12]. Few studies have shown time-series medial gastrocnemius fascicle data during the gait cycle for older adults. However, our fascicle data are well-aligned with previously reported in vivo ultrasound imaging of the plantarflexors during walking [10, 12, 13, 15, 23–26]. Our older adult cohort walked with reduced peak fascicle shortening during all conditions compared to young adults – an outcome not explained by reduced MTU shortening or ankle angle alone. Although the specific mechanisms driving this age effect are unclear, it would facilitate relatively longer fascicle lengths at slower shortening velocities during push-off and was disproportionately prevalent when the propulsive demands of walking were increased. Longer fascicle lengths at slower shortening velocities may convey a physiological advantage to muscle force generation when task demands are high, and deficits in muscle strength may precipitate the need for such changes. Consistent with this premise, older and young adults exhibited more pronounced fascicle lengthening during early to mid-stance in response to Ramp than during Pref. However, similar behavior was also observed for muscle-tendon unit length, suggesting that at least part of this change can be explained by changes in ankle and knee joint kinematics. Nevertheless, these findings suggest that gastrocnemius fascicle operating lengths are fundamentally altered by age and contribute to adaptation within the muscle-tendon unit complex in response to increased propulsive demands during walking.

Medial gastrocnemius fascicle lengths strongly correlated with performance in the ramped impeding force condition in older adults. This association was not prevalent in younger subjects and persisted even after controlling for isometric plantarflexor strength and subject stature. We interpret this finding to suggest that gastrocnemius operating behavior, independent of interindividual differences in strength or anthropometries, plays a mechanistic role in governing one’s capacity to enhance push-off intensity during walking at a constant speed. We are certainly not the first to implicate age-related changes in plantarflexor muscle fascicle length-tension behavior in reduced performance in functional mobility tasks. Indeed, Stenroth et al. (2015) reported that medial gastrocnemius fascicle lengths in older adults, in addition to Achilles tendon stiffness, correlated with improved distance on the 6 minute walk test [11]. Combined, these associations warrant further research into rehabilitation techniques that focus on ways to steer muscle contractile dynamics in those with diminished push-off intensity. For example, attempts to stiffening the Achilles tendon may have favorable implications for the corresponding plantarflexor muscle fascicle behavior. In addition to enhanced series elastic energy storage and return, this could permit longer fascicle operating behavior which could in turn allow older adults to better accomplish tasks in the community that require enhanced push-off intensity (i.e. walking faster and walking uphill).

Maximum walking speed was significantly slower in older than in young adults. We posit that maximum walking speed has higher efficacy to assess functional walking ability than what can be inferred from preferred walking. For example, although we found no difference in peak ankle moment between older and young adults during normal walking, a significant interaction revealed that only young adults increased peak ankle moment to walk at their maximum speed. However, in neither group did fascicle length at peak ankle moment during normal walking correlate with maximum walking speed or the change in walking speed from preferred to maximum. This is surprising, as longer fascicles allude to a higher number of serial sarcomeres and thus higher possible contractile velocities. Despite this, prior literature in young adults has shown that fascicle length change behavior is relatively insensitive to changes in walking speed [15]. While these muscle-level insights provide an entirely novel contribution, this is not the first evidence to suggest that maximum walking performance is disassociated from habitual locomotor patterns. For example, previous work has shown that strengthening the plantarflexor muscles in older adults elicits improvements in maximum walking speed but has little to no effect on habitual ankle joint kinetics or preferred walking speeds [27, 28]. We advocate for the more widespread use of dynamic in vivo ultrasound during functional tasks in aging research to better understand the mechanisms underlying cross-sectional differences and interventional outcomes.

There are several limitations to this study. First, our trial order prevented the impeding force condition from influencing preferred locomotor patterns, and ensured maximum exertion during Ramp. As such, we cannot exclude the possibility of ordering effects. Despite similarities in biomechanics during treadmill and overground walking [29, 30], we also cannot be certain that use of a treadmill did not influence the muscle-level response. We also acknowledge that although our sample size is consistent with that used in previously published studies [10, 12], our independent-samples comparisons may be underpowered. Finally, fascicle lengths measured in vivo during walking cannot provide direct evidence for where on the force-length curve older and young adults are actually operating. We tend to interpret the limited available literature [10, 12] to suggest that older adults may operate further down the ascending limb of this relation during walking, which could in turn have implications for force generating capacity and ability to accomplish the more rigorous walking tasks as well as young adults. However, there is a critical need for empirical data to describe aging effects on fascicle (and sarcomere) length-tension behavior in the same human subjects as functional walking tasks such as those included in this study. Such data would allow us to gauge the extent to which reaching a fascicle shortening “limit”, i.e. fascicles being too short to generate sufficient force, precipitates failure during demanding walking tasks.

In conclusion, we report that shorter gastrocnemius fascicle lengths in older adults associate with worse capacity to enhance push-off intensity in walking, even when controlling for isometric strength and subject anthropometries. We also found that older adults undergo less relative fascicle shortening, especially in tasks which increase the propulsive demands of walking. These findings provide muscle-level insight for rehabilitation techniques that improve push-off intensity in older adults and assistive technologies designed to steer plantarflexor muscle fascicle operating behavior during functional tasks.

Highlights.

• Measured gastrocnemius muscle response to increased propulsive demands in walking

• Shorter gastrocnemius fascicle lengths associate with worse push-off capacity

• Older adults undergo less fascicle shortening, especially in higher demand tasks

• Muscle-level insight for techniques to improve push-off intensity in older adults