Tendon loading in runners with Achilles tendinopathy: Relations to pain, structure, and function during return-to-sport

Abstract

We aimed to (1) compare pain, tendon structure, lower limb function, and Achilles tendon loads while running between limbs in runners with Achilles tendinopathy, and (2) explore the relations of pain, tendon structure, and lower limb function to Achilles tendon loads while running. Twenty runners with Achilles tendinopathy participated in this pilot study. Pain was assessed with questionnaires, quantitative sensory testing, and functional testing. Tendon morphology and mechanical properties were evaluated with ultrasound imaging, continuous shear wave elastography, and ultrasound imaging combined with dynamometry. Lower limb function was assessed with an established test battery. Achilles tendon loads were estimated from biomechanical data acquired during running. Compared to the least symptomatic limb, the most symptomatic limb had lower scores on the Victorian Institute of Sports Assessment – Achilles questionnaire and worse pain during drop countermovement jumping, hopping, and running. Tendon thickness and cross-sectional area were greater, and Young’s modulus, drop countermovement jump height, and plyometric quotient during hopping were lower on the most symptomatic limb. Side-to-side differences in drop countermovement jump height were significantly associated with side-to-side differences in Achilles tendon peak forces and average loading rates during running. Various measures of pain, structure, and function differ between limbs in runners with Achilles tendinopathy during return-to-sport. Tendon forces, however, do not differ between limbs during comfortable running. In addition to measures that differ between limbs, measures of performance during drop countermovement jumping may aid in clinical decision-making during return-to-sport because they are associated with tendon forces while running.

1 |. INTRODUCTION

Achilles tendinopathy is an overuse injury characterized by pain with loading activities and impaired functional performance.¹ Runners frequently develop Achilles tendinopathy,² with deleterious loading behaviors implicated in the multifactorial etiology.³ For treatment, progressive loading programs are recommended and supported by the highest level of evidence.⁴ Despite favorable long-term outcomes with progressive loading, approximately half of those with Achilles tendinopathy will experience either symptom persistence or recurrence.^5–8 Accelerated return-to-sport (RTS) appears to increase the risk of symptom recurrence,⁵ underpinning the importance of a structured RTS phase during rehabilitation. Yet, it is currently unclear what measures should be used by clinicians, patients, and other stakeholders when progressing those with Achilles tendinopathy through the RTS phase.

A RTS program for patients with Achilles tendinopathy was proposed in 2015 by Silbernagel and Crossley.⁹ This RTS program, which is based on research and clinical experience,¹⁰ relies heavily on patient-reported measures of pain to guide the prescription of high-load activities (e.g., running). While measures of pain are likely to help guide clinical decisions during the early stages of RTS, they may be insufficient in late stages of RTS when symptoms are nearly resolved and the athlete is striving toward performance goals at or above their preinjury level. Thus, measures beyond pain may be needed to guide late-stage RTS decisions to increase the likelihood of a successful return-to-performance without symptom recurrence. Measures of tendon structure and lower limb function may be ideal for supplementing measures of pain during late stages of RTS because structural alterations and functional impairments persist in those with Achilles tendinopathy even after symptoms have resolved.^11,12 However, an approach to RTS for those with Achilles tendinopathy that uses symptoms, structure, and function to guide clinical decision-making is currently unsubstantiated.

As first steps toward developing RTS guidelines for those with Achilles tendinopathy, we aimed to gain foundational knowledge in this pilot study. Our first objective was to compare measures of pain, tendon structure, lower limb function, and Achilles tendon loading while running between limbs in runners with Achilles tendinopathy during RTS. Our second objective was to explore the relations of pain, tendon structure, and lower limb function to Achilles tendon loads while running during RTS. These objectives were important and clinically relevant because RTS outcomes for runners with Achilles tendinopathy may be improved by (1) gaining a better understanding of asymmetries that exist during RTS and (2) identifying clinical measures that can provide insights about possibly deleterious tendon loading behaviors. We hypothesized that measures of pain, structure, and lower limb function would be significantly different between limbs and associated with side-to-side differences in Achilles tendon loads while running.

2 |. MATERIALS AND METHODS

2.1 |. Study design

The University of Delaware Institutional Review Board, which is registered with the United States Department of Health and Human Services (IORG#0000279), approved the protocol for this cross-sectional, laboratory-based study. Participants provided written consent after reviewing study procedures and risks. Data collection was performed in two visits. At the first visit, which averaged approximately 1.5 h, we collected participant characteristics followed by measures of pain, tendon structure, and lower limb function. At the second visit, which averaged approximately 3 h, we estimated tendon mechanical properties using various methodologies (see details below) and then evaluated mechanics and pain while running. The visits were at least two, but no more than six, weeks apart. Two visits, rather than one, reduced concerns regarding one experimental procedure influencing another (e.g., calf muscle endurance testing and running mechanics). Furthermore, participants were not actively receiving treatment or changing their physical activity levels between visits, so pain, tendon structure, and lower limb function were unlikely to change beyond minor fluctuations due to recent physical activity. To further ensure pain, structure, and function were similar at each visit, participants were asked to refrain from running and other lower extremity exercise for 2 days before each visit.

2.2 |. Participants

To be eligible, participants were required to be a runner diagnosed with Achilles tendinopathy currently in the RTS phase. A runner was operationally defined as some-one who (1) self-identifies as a runner; and, (2) runs more than 30 min at least twice per week. Achilles tendinopathy was diagnosed based on self-report of localized Achilles tendon pain and stiffness, and clinical examination with either tenderness to palpation, a positive arc sign, or a positive Royal London Hospital test. Details about these diagnostic tests and their use in clinical practice can be found elsewhere.^13,14 To be considered in the RTS phase, participants were required to self-report (1) no current symptoms with activities of daily living (e.g., standing, walking, and stair negotiation) and (2) symptoms of Achilles tendinopathy within the past year. This operational definition of the RTS phase for Achilles tendinopathy was based on recommendations from Silbernagel and Crossley.⁹ We excluded individuals with a history of surgery to either lower limb or a lower limb injury, besides Achilles tendinopathy, within the past year that limited running participation.

Twenty-four runners fit the study criteria and enrolled. Four participants were excluded from all analyses because of withdrawing prior to the second visit (n = 2) or having unusable biomechanical data (n = 2). Demographics, anthropometrics, as well as injury and running information for the twenty remaining participants are shown in Table 1. Of note, fifteen participants reported unilateral symptoms, while five reported bilateral symptoms. Because we included those with unilateral and bilateral Achilles tendinopathy, we categorized limbs as “most” or “least” symptomatic based on participant-report and clinical examination. This inclusion and categorization process is common in Achilles tendinopathy research because the prevalence of bilateral symptoms is high (~30%), the incidence of developing bilateral symptoms is 41%, and both Achilles tendons are structurally compromised in those with unilateral Achilles tendinopathy.^7,10,15

TABLE 1.

Participant characteristics

Demographics & Anthropometrics
Age (years); mean ± SD	44 ± 13
Height (cm); mean ± SD	174.1 ± 9.1
Body mass (kg); mean ± SD	75.4 ± 12.8
Female; % (n)	30 (6)
Injury information
Unilateral symptoms; % (n)	75 (15)
Midportion Achilles tendinopathy; % (n)	65 (13)a
Duration of symptoms (months); mean ± SD	40.8 ± 52.6
Running information
Current volume (km/week); mean ± SD	36.2 ± 23.5
Current pace (minutes/km); mean ± SD	5.62 ± 0.96
Current frequency (runs/week); median (IQR)	3 (2)
Preinjury volume (miles/week); mean ± SD	39.8 ± 20.1b
Preinjury pace (minutes/km); mean ± SD	5.31 ± 0.63b
Preinjury frequency (runs/week); median (IQR)	4 (2)b
Running Experience (years); mean ± SD	15.9 ± 12.3

^aThe remaining participants were diagnosed with insertional Achilles tendinopathy.

^bn = 19; one participant did not answer questions about preinjury running.

2.3 |. Evaluation of pain

2.3.1 |. Patient-reported outcome measure of symptom severity

Symptom severity was assessed with the Victorian Institute of Sports Assessment – Achilles (VISA–A) questionnaire at the second visit.¹⁶ The VISA–A is an 8-item, disease-specific questionnaire that is scored 0–100, with 100 indicating full recovery. The Victorian Institute of Sports Assessment – Achilles questionnaire (VISA-A) was completed separately for the most and least symptomatic limbs.

2.3.2 |. Quantitative sensory testing

Standardized assessment of pain sensitization may provide useful information about tendon health that informs clinical decisions during RTS if associated with Achilles tendon loading behaviors. We determined pressure-pain thresholds for each Achilles tendon. Participants were prone while pressure was simultaneously applied to the medial and lateral aspects of the tendon with a mechanical algometer (Somedic SenseLab AB, Sösdala, Sweden) (Figure 1). Pressure was applied to the tendon in this manner, instead of through a posterior–anterior pressure, to reduce the likelihood of measuring pain sensitization of anatomic structures deep to the Achilles tendon and to simulate how pressure is applied during clinical examination.¹⁴ The contact area was 1 cm² and pressure was applied at a rate of 30 kPa/s, similar to parameters used by Plinsinga et al.¹⁷ Participants used a response button to indicate the instant when pain was first elicited. Testing was completed at the region of maximal tendon thickness for the most symptomatic limb, as identified with ultrasound imaging. Although seven participants were diagnosed with insertional Achilles tendinopathy, the location of maximal tendon thickness was always distal to the enthesis, allowing for pressure-pain thresholds to be assessed with a medial-lateral approach. For the least symptomatic limb, testing was completed at the region of maximum tendon thickness if pathologic or an anatomically matched location if non-pathologic. The anatomically matched location was determined by measuring the distance from the Achilles tendon osteotendinous junction to the location of maximal thickness on the most symptomatic limb with a tape measure. Three trials were performed for both limbs and average pressure-pain thresholds were determined.

FIGURE 1.

Pressure-pain threshold testing of the Achilles tendon with a mechanical algometer (Somedic SenseLab AB, Sösdala, Sweden)

2.3.3 |. Pain with Loading

The Numeric Pain Rating Scale (NPRS; 0 = no pain, 10 = worst pain imaginable) was used to rate pain throughout functional testing (see section 2.5) and treadmill running (see section 2.6). For functional tests with more than one trial, maximum pain was used for analysis.

2.4 |. Evaluation of tendon structure

2.4.1 |. Morphology

We evaluated Achilles tendon morphologic properties with ultrasound imaging at the first visit. Participants were prone with their feet hanging freely. Images were acquired with a LOGIQ e ultrasound system (GE Healthcare) with a wide-band linear array probe (5.0–13.0 MHz). At least three short-axis images and three long-axis extended field of view images were acquired for each Achilles tendon. Images were exported to OsiriX imaging software (Pixmeo SARL) where measurements were made. For each morphologic variable of interest, measurements were taken on three images and averaged. A single, experienced evaluator (PC) collected and processed all ultrasound images.

The morphologic variables of interest were maximal tendon thickness and maximal cross-sectional area, which are reliably assessed with ultrasound imaging.¹⁸ With the long-axis extended field of view images, anterior–posterior tendon thickness was measured at the location of maximal thickness. With the short-axis images, maximal cross-sectional area was determined by manually tracing the outer border of the tendon. For the most symptomatic limb, cross-sectional area measurements were made at the location of maximal thickness. For the least symptomatic limb, measurements were made at the anatomically matched location if healthy or at the location of maximal thickness if pathologic.

2.4.2 |. Mechanical properties

In addition to evaluating tendon morphology, we estimated mechanical properties for each Achilles tendon at the second visit. Mechanical properties were of interest because they reflect how tendon responds to load and are altered in those with Achilles tendinopathy.^19–21 The variables of interest were shear modulus (i.e., resistance to shearing force), viscosity (i.e., rate-dependent resistance to a shearing force), and Young’s modulus (i.e., resistance to a tensile force).

Shear modulus and viscosity were estimated for each Achilles tendon with continuous shear wave elastography (cSWE).^22,23 This technique has been validated,^22,23 and applied in those with Achilles tendinopathy for the purposes of establishing prognosis and evaluating treatment effects.^24,25 The procedures used for cSWE in the current study are identical to those reported in Corrigan et al.²²

Young’s modulus was estimated for each Achilles tendon through a combination of ultrasound imaging and dynamometry. The procedures were modified from techniques reported by Arya & Kulig and Chang & Kulig.^19,20 Participants were prone with their foot secured in a KinCom Dynamometer (Model 500H, Isokinetic International). The ankle was in 0° of plantar flexion, and the dynamometer axis was aligned with the lateral malleolus. Velcro strapping secured the forefoot against the foot pedal, the hindfoot in the heel cup of the dynamometer, and the participant to the table. Once properly positioned, the medial gastrocnemius myotendinous junction was located with ultrasound imaging and marked on the skin. A thin piece of Scotch^® rubber mastic tape (3 M Company) was attached to the skin immediately distal to the mark and perpendicular to the long-axis of the lower leg, creating a shadow within the ultrasound field-of-view to account for movement of the ultrasound transducer. To reduce the chances of the rubber tape moving during testing, athletic tape secured both ends. Participants performed 5-s maximal voluntary isometric contractions (MVICs) of ankle plantar flexion while cine-loop ultrasound recordings of the myotendinous junction were collected (Frame rate = 33 Hz). Before performing MVICs, two submaximal trials at 50% and 75% effort were completed to ensure the participant understood testing procedures and for the evaluator (PC) to identify the optimal position of the ultrasound transducer. For the MVICs, participants were told to push down as hard as possible. Verbal encouragement was provided during each MVIC trial. A 2-min break was given between MVIC trials, and a 5-min break was given when switching limbs. MVICs were repeated until three valid trials were collected for each limb. A trial was considered valid if the displacement of the medial gastrocnemius myotendinous junction was visualized throughout the MVIC, ankle plantar flexion torque plateaued (i.e., indicative of maximal effort), and strapping remained secure. These criteria were used to decrease the likelihood of including a trial with submaximal torque or unwanted movement. If a trial was deemed invalid, an additional trial was performed. No participant performed more than five trials on a single limb.

Young’s modulus was estimated for each Achilles tendon as follows: YM = (F/A)/(ΔL/L₀). YM is Young’s modulus, F is the peak Achilles tendon force during MVIC, A is the average cross-sectional area of the tendon, ΔL is the linear displacement of the medial gastrocnemius myotendinous junction during MVIC, and L₀ is the resting length of the Achilles tendon. For reporting YM, pascals were converted to megapascals (MPa). F was determined by dividing the peak plantar flexion moment during MVIC by a 5 cm Achilles tendon moment arm. A was determined by averaging tendon cross-sectional area measurements that were obtained 2, 4, and 6 cm proximal to the osteotendinous junction. ΔL was determined by measuring displacement of the myotendinous junction during the MVICs. With OsiriX imaging software (Pixmeo SARL), positions of the myotendinous junction and rubber tape shadow were recorded at rest and during each MVIC (Figure 2). Displacement was calculated after adjusting for movement of the ultrasound transducer (i.e., change in shadow position). L₀ was determined using a reliable method as described by Silbernagel and colleagues that uses long-axis extended field of view ultrasound images.²⁶

FIGURE 2.

Ultrasound images revealing position of medial gastrocnemius (MG) myotendinous junction at rest (1) and during ankle plantar flexion maximal voluntary isometric contraction (2). Line A was drawn parallel to the right edge of the rubber tape shadow. The change in Line B was used to adjust for movement of the ultrasound transducer. The change in Line C was used to measure linear displacement of the medial gastrocnemius myotendinous junction along the long-axis of the lower leg

2.5 |. Evaluation of lower limb function

Participants completed three countermovement jumps (CMJ), three drop countermovement jumps (drop CMJ), two bouts of hopping, and one heel-rise endurance test for each limb. These tasks are part of a functional test battery for evaluating individuals with Achilles tendinopathy.²⁷ For each CMJ, participants jumped with a single leg as high as possible from flat ground with their arms behind their back. Using a light mat (MuscleLab^™, Ergotest Innovation AS), jump height was derived from flight time (max height = $\frac{1}{8}$ gt²; g = gravity, t = flight time). Drop CMJs were like CMJs, except the participant dropped from a 20 cm box prior to jumping as high as possible. Hopping performance was assessed by having the participant repetitively hop 25 times on a single leg. After removing the first 3 and last 2 hops, plyometric quotient (i.e., flight time/contact time) was calculated.²⁷ The heel-rise endurance test was performed on a single leg while standing on a 10° incline box and following a metronome that led to 30 repetitions per minute.²⁷ A linear encoder (MuscleLab) was attached to the posterior heel to track the height of each repetition. Participants stopped when they were unable to perform any additional repetitions or when they were unable to follow proper testing procedures (e.g., straight knee). Total work was calculated (Σ displacement × body mass x gravity). Participants wore standardized shoes (Glycerin 15, Brooks Sports Inc.) during each functional task, including the laboratory-based treadmill run, but were barefoot during the heel-rise test. For each variable obtained during the jumping tasks, an average was calculated.

2.6 |. Evaluation of Achilles tendon loads while running

2.6.1 |. Data collection

Participants were equipped with 47 retroreflective markers on their pelvis and lower limbs. The marker set was modified from Willy et al²⁸ and placed by the same evaluator (PC) for all participants. A 10-s static trial was collected while standing in anatomical position. Twenty markers were then removed, leaving 27 markers for segment tracking. The pelvis, thighs, and shanks were tracked with rigid marker clusters. The feet were tracked with markers over the 5th metatarsal head as well as the superior, inferior, and lateral aspects of the heel.

Participants ran at their self-selected endurance speed (mean ± SD of 2.93 ± 0.29 m/s) for 7 min on a treadmill instrumented with force plates (Bertec Corp.). Endurance speed was defined as the participant’s speed during the middle of a long run. The first 6 min were used for familiarization and fine-tuning the speed of the treadmill based on participant feedback. During the 7th minute, ground reaction forces were sampled at 2000 Hz and marker trajectories were sampled at 100 Hz with an 8-camera motion capture system (Motion Analysis Corp).

2.6.2 |. Data processing

The static and running trials were trimmed to 1 s and 10 gait cycles, respectively. Marker trajectories were gap-filled and labeled in Cortex software (Version 8, Motion Analysis Corp.) then exported to Visual3D software (Version 8, C-motion Inc.) for all remaining processing. Lower extremity models were constructed using trajectory data from the static trial. We modeled the pelvis as a cylinder with iliac crests defining the proximal end and greater trochanters defining the distal end. The medial-lateral diameter of the pelvis was defined by the greater trochanter markers, while the anterior–posterior diameter was defined by taking a physical measurement with breadth calipers at the base of the sacrum during data collection. We modeled the thighs, shanks, and feet as truncated cones. The length of each thigh was defined as the distance from the hip to the knee joint center. Hip joint centers were positioned 23.4% and 76.6% of the inter-trochanteric distance, then offset superiorly by 4.7% of the inter-trochanteric distance.²⁹ Knee joint centers were determined as the centroid of the femoral condyle and tibial plateau markers.³⁰ The proximal end diameter of each thigh was calculated as 50% of the inter-trochanteric distance, while the distal end diameter was defined by the femoral condyles. The proximal end of each shank was defined by the knee joint center, with a diameter determined from medial and lateral tibial plateau markers. Distal end location (i.e., midpoint between markers) and diameter for each shank were determined with the medial and lateral malleoli markers. The feet were defined proximally by the malleoli and distally by the first and fifth metatarsal heads. Each segment had six degrees of freedom with coordinate systems and segmental inertial properties scaled to height and body mass.^31,32

Lower extremity models were applied to the running trials for inverse dynamic analyses and subsequent estimation of Achilles tendon forces. We identified stance (i.e., 20 N threshold for initial contact and toe-off events) with vertical ground reaction force data that were low-pass filtered at 30 Hz with a fourth-order Butterworth filter. With raw marker and force data, we performed Newton-Euler inverse dynamic analyses to calculate net ankle joint moments that were resolved in the shanks coordinate system. Achilles tendon forces were then estimated by dividing the ankle plantar flexion moment by the Achilles tendon moment arm, which was held constant at 5 cm. Achilles tendon force waveforms were then low-pass filtered at 15 Hz with a fourth-order Butterworth filter and normalized to each participant’s bodyweight. We chose to filter our data after inverse dynamic calculations because it is a valid technique that reduces the likelihood of (1) impact-like artifacts when filtering marker and force data at different cutoff frequencies and (2) over-smoothing force plate data.³³

The variables of interest were peak forces, average loading rates, and impulses for each Achilles tendon. Peak force was the maximum load experienced by the Achilles tendon during stance. Average loading rate was the rate at which force was applied to the Achilles tendon in the middle 60% of the time from initial contact to peak force. To determine average loading rate, we created events at 20% and 80% of the duration from initial contact to peak force. With the first derivative of the Achilles tendon force waveform, we determined the average loading rate within the 60% interval. Impulses were calculated as the time integral (i.e., the area under the Achilles tendon force-time curve) during stance.

2.7 |. Analytic approach

We summarized variables of interest from each limb with medians and interquartile range (IQR). Wilcoxon signed-rank tests were performed to determine whether differences existed between most and least symptomatic limbs. Spearman’s correlations were performed to evaluate the associations of side-to-side differences in pain, structure, and lower limb function to side-to-side differences in Achilles tendon loading. Associations with Achilles tendon loading were not evaluated for side-to-side differences in pain during CMJ, pain during drop CMJ, or pain during heel-rise testing because most participants did not experience pain in either limb during these functional tasks. All side-to-side differences were calculated by subtracting the most symptomatic limb from the least symptomatic limb for each variable of interest. Statistical analyses were performed in SPSS (Version 26, IBM Corp.) with a significance level set to p < 0.05.

3 |. RESULTS

3.1 |. Pain

3.1.1 |. Between-limb comparisons

Victorian Institute of Sports Assessment – Achilles questionnaire scores were significantly lower for the most symptomatic limb compared to the least symptomatic limb (Table 2). Furthermore, pain with loading was significantly worse in the most symptomatic limb compared to the least symptomatic limb during drop CMJ, hopping, and running (Table 2). Pressure-pain thresholds, pain during CMJ, and pain during heel-rise testing were not significantly different between limbs (Table 2).

TABLE 2.

Side-to-side differences in pain, tendon structure, lower limb function, and Achilles tendon loads while running for individuals with Achilles tendinopathy during return-to-sport

	Most symptomatic limb; median (IQR)	Least symptomatic limb; median (IQR)	p-value
Pain
VISA-A (scores)	80 (22)	95 (12)	0.01 a
PPT (kPa)	260 (134)	257 (172)	0.28
CMJ Pain (0–10)	0 (1)	0 (0)	0.12
Drop CMJ Pain (0–10)	0 (1)	0 (0)	0.03 a
Hopping Pain (0–10)	1 (3)	0 (0)	<0.01 a
Heel-Rise Pain (0–10)	0 (0)	0 (0)	0.07
Running Pain (0–10)	2 (3)	0 (0)	<0.01 a
Tendon structure
Thickness (mm)	5.98 (2.60)	4.95 (1.79)	0.01 a
Cross-sectional area (cm2)	0.671 (0.296)	0.609 (0.283)	0.04 a
Shear modulus (kPa)b	88.9 (34.3)	85.1 (31.5)	0.81
Viscosity (Pa*s)b	60.1 (12.4)	52.0 (30.4)	0.47
Young’s Modulus (MPa)	1556 (582)	1622 (1036)	0.02 a
Lower limb function
CMJ Height (cm)	8.3 (3.6)	9.9 (3.8)	0.09
Drop CMJ Height (cm)	7.5 (4.6)	9.0 (3.7)	0.01 a
Hopping PQ	0.45 (0.14)	0.49 (0.11)	0.01 a
Heel-Rise Work (J)	2017 (1180)	1908 (1142)	0.65
Achilles tendon loads during running
Peak Force (BW)	4.90 (0.43)	4.92 (0.82)	0.74
Average Loading Rate (BW/s)	45.6 (10.8)	47.7 (12.4)	0.37
Impulse (BW*s)	0.673 (0.126)	0.633 (0.137)	0.65

Abbreviations: BW, bodyweight; CMJ, countermovement jump; IQR, interquartile range; J, Joules; kPa, kilopascals; MPa, megapascals; Pa*s, pascal-seconds; PPT, pressure-pain threshold; PQ, plyometric quotient; VISA-A, Victorian Institute of Sports Assessment – Achilles questionnaire.

^aIndicates p < 0.05 when comparing limbs with Wilcoxon signed-rank tests.

^bn = 19 because of equipment malfunction.

3.1.2 |. Correlations with Achilles tendon loading

Side-to-side differences in VISA–A scores, pressure-pain thresholds, pain during hopping, and pain during running were not significantly associated with side-to-side differences in Achilles tendon peak forces, average loading rates, and impulses (Table 3).

TABLE 3.

Associations of side-to-side differences in pain, tendon structure, and lower limb function with side-to-side differences in Achilles tendon loads during running in individuals with Achilles tendinopathy who are in the return to sport phase

	Peak force	Average loading rate	Impulse
Pain
VISA-A	0.256 (0.276)	0.427 (0.061)	0.333 (0.151)
PPT	0.391 (0.088)	0.322 (0.166)	0.346 (0.135)
Hopping Pain	−0.302 (0.195)	−0.367 (0.111)	−0.348 (0.132)
Running Pain	−0.054 (0.822)	−0.070 (0.769)	−0.169 (0.476)
Tendon structure
Thickness	−0.247 (0.295)	−0.343 (0.139)	−0.308 (0.186)
Cross-sectional area	−0.214 (0.366)	−0.189 (0.424)	−0.295 (0.207)
Shear Modulusa	−0.263 (0.276)	−0.307 (0.201)	−0.205 (0.399)
Viscositya	0.054 (0.825)	0.125 (0.611)	0.058 (0.814)
Young’s Modulus	0.241 (0.307)	0.154 (0.523)	0.229 (0.332)
Lower limb function
CMJ Height	−0.126 (0.596)	−0.017 (0.945)	−0.182 (0.443)
Drop CMJ Height	0.462 (0.040) b	0.481 (0.032) b	0.438 (0.054)
Hopping PQ	0.254 (0.280)	0.170 (0.474)	0.236 (0.316)
Heel-Rise Work	−0.015 (0.950)	−0.015 (0.950)	−0.054 (0.821)

Note: Results are reported as Spearman’s rank correlation coefficients (ρ) and p-values.

Abbreviations: CMJ, countermovement jump; PPT, pressure-pain threshold; PQ, plyometric quotient; VISA-A, Victorian Institute of Sports Assessment – Achilles questionnaire.

^an = 19 because of equipment malfunction.

^bIndicates significant association.

3.2 |. Tendon structure

3.2.1 |. Between-limb comparisons

Compared to the Achilles tendon of the least symptomatic limb, the Achilles tendon of the most symptomatic limb was significantly thicker, had a larger cross-sectional area, and a lower Young’s modulus (Table 2). Estimates of shear modulus and viscosity of the Achilles tendons were not significantly different between limbs (Table 2).

3.2.2 |. Correlations with Achilles tendon loading

There were no significant associations between side-to-side differences in Achilles tendon structural variables and side-to-side differences in Achilles tendon peak forces, average loading rates, and impulses (Table 3).

3.3 |. Lower limb function

3.1.1 |. Between-limb comparisons

Drop CMJ height and plyometric quotient during hopping were significantly lower on the most symptomatic limb compared to the least symptomatic limb (Table 2). CMJ height and heel-rise work were not significantly different between limbs (Table 2).

3.3.2 |. Correlations with Achilles tendon loading

Side-to-side differences in drop CMJ height were significantly associated with side-to-side differences in Achilles tendon peak forces and average loading rates during running (Table 3; Figure S1). Side-to-side differences in CMJ height, plyometric quotient during hopping, and heel-rise work were not significantly associated with side-to-side differences in Achilles tendon peak forces, average loading rates, or impulses (Table 3).

3.4 |. Achilles tendon loads while running

Ensemble curves for Achilles tendon forces are shown in Figure 3. There were no significant differences between the most and least symptomatic limbs for Achilles tendon peak forces, average loading rates, or impulses (Table 2).

FIGURE 3.

Average Achilles tendon forces during comfortable treadmill running for the most (black solid line) and least (gray solid line) symptomatic limbs of individuals with Achilles tendinopathy during the return-to-sport phase. Dashed lines represent ±1 standard deviation

4 |. DISCUSSION

In this pilot study, we compared pain, tendon structure, and function between limbs in individuals with Achilles tendinopathy who are in the RTS phase, and explored relations to Achilles tendon loads during running. One of our key findings was that, on average, runners with Achilles tendinopathy have significant differences between limbs for various measures of pain, structure, and jump performance during RTS. Additionally, we found that measures of drop CMJ height are associated with side-to-side differences in Achilles tendon loading behaviors during running. Collectively, these findings provide preliminary evidence that several factors may need to be targeted during RTS to decrease the likelihood of recurrent, persistent, or contralateral Achilles tendinopathy.

Since this is one of the first studies to focus on individuals with Achilles tendinopathy during RTS, it is critical to describe characteristics of our sample. All participants were runners with Achilles tendinopathy on the RTS continuum. Meaning, at the time of data collection, participants were pain-free with activities of daily living (e.g., walking, stairs), but had not yet run for a year without symptoms. Notably, in combination with our exclusion criteria (i.e., history of surgery or injury elsewhere), this definition of RTS yielded a cohort of individuals who had nearly returned to their preinjury running levels (Table 1). However, based on variability in measures of pain, structure, and function, it appears that our sample was heterogeneous. Including individuals at any point along the RTS continuum, from return-to-participation to return-to-performance, may partially explain the large variability seen in our results. Future work may want to consider additional inclusion or exclusion criteria when studying RTS in those with Achilles tendinopathy to ensure a homogenous sample of individuals at a similar sub-phase of RTS, as described by Ardern et al.³⁴

An interesting finding from the current study is that, on average, individuals with Achilles tendinopathy symmetrically load their Achilles tendons while running (Figure S1 and Table 2), even though asymmetries are present for certain measures of pain, structure, and function. It was hypothesized that Achilles tendon forces to be lower on the most symptomatic limb compared to the least symptomatic limb because of pain, altered structure, and impaired function. The intensity of the laboratory-based treadmill run may partially explain the lack of asymmetry in tendon loading. In the current study, participants ran for 7 min on a treadmill at their self-selected endurance speed (i.e., comfortable speed for every participant that mimicked their speed in the middle of a long duration run). Previous research on persons post-Achilles tendon rupture suggests that loading asymmetry increases with task difficultly.^28,35 Thus, we may have found greater asymmetry in Achilles tendon loading if we increased the intensity by either increasing the speed, extending the duration, or changing the environment of the run. Future research should aim to understand how individuals with Achilles tendinopathy modify their loading behaviors in response to different exposures, such as speed, fatigue, inclination, and surface.

Twenty-five percent of our sample had bilateral Achilles tendinopathy. We recognize that inclusion of these individuals may have affected our results, including the lack of side-to-side difference in tendon loads while running. However, when repeating the analyses without those with bilateral Achilles tendinopathy, the results were unchanged. This may be because each participant with bilateral symptoms in the current study was able to easily identify a more problematic limb, making asymmetry in pain, structure, function, and running mechanics more probable. Additionally, since many individuals with Achilles tendinopathy progress from unilateral to bilateral disease,⁷ the lack of asymmetry in tendon loads while running may be because those with unilateral symptoms have bilateral alterations in structure and function.¹⁵ Although our results were unaffected by including those with bilateral symptoms, future research should carefully consider how Achilles tendinopathy is defined and the appropriateness of including individuals regardless of unilateral or bilateral injury.

Clinically, measures of pain are used to monitor recovery, gauge response to treatment, and make RTS decisions. Although patient outcomes are favorable when pain is used to guide rehabilitation and RTS decisions,¹⁰ it remains unclear which measures of pain are most valuable. In the current study, we found significant differences between limbs for pain during drop CMJ, hopping, and running. Interestingly, these functional tasks involve more ballistic movement compared to the functional tasks that showed no differences in pain (i.e., CMJ and heel-rise). Even though we found several between-limb differences in pain, the differences were not associated with Achilles tendon loading behaviors while running. Taken together, it appears that measures of pain during functional tasks may be useful for determining status of recovery, but provide little insight into tendon loading. Further research in this area is needed to understand how measures of pain should be used to guide RTS decisions for those with Achilles tendinopathy.

Tendon morphology and mechanical properties are altered in those with Achilles tendinopathy.^1,19–21,36 Namely, the size of the tendon increases and tissue stiffness reduces. Importantly, evidence suggests that tendon structure can be modified, or even normalized, with load-based treatments.^37–39 Normalization of tendon properties may be critical to return fully to preinjury levels of sport since Achilles tendon mechanical properties are associated with running efficiency.^40–42 Surprisingly, we found that side-to-side differences in tendon morphology and mechanical properties were not associated with side-to-side differences in Achilles tendon loading during running. Prior to performing the study, we expected that Young’s modulus would have the strongest association with Achilles tendon loading behaviors. We speculated a significant relation because Young’s modulus is a material property that reflects the tendon’s response to a tensile load, which is the principal type of load the tendon experiences during running. Interestingly, although we found no relation to tendon loading, there were significant differences between the most and least symptomatic limbs for tendon thickness, cross-sectional area, and Young’s modulus. When considered with the lack of difference between limbs for shear modulus and viscosity, it seems that tendon morphology and tensile properties remain affected into the RTS phase while shear properties are more similar between limbs. Further research is needed in this area to determine whether measures of tendon morphology and mechanical properties can be used for prescribing runs of optimal load and monitoring structural response to RTS activities.

Functional performance measures are used in clinical research and practice to identify impairments and monitor recovery. Functional measures have also been incorporated in RTS criteria for various musculoskeletal injuries (e.g., anterior cruciate ligament injury). For Achilles tendinopathy, specific RTS criteria do not exist, although a RTS program that uses pain and rate of perceived exertion has been proposed.⁹ Functional testing may be a critical part of a RTS program, since deficits in muscle-tendon function exist after symptoms have resolved.¹¹ In the current study, we found that side-to-side differences in drop CMJ height were associated with side-to-side differences in peak Achilles tendon force and average loading rates while running in individuals with Achilles tendinopathy during the RTS phase (Figure S1). These relationships suggest that the limb with worse drop CMJ performance has lower Achilles tendon loads while running. This finding may be partially confounded by the presence of pain, but we did not adjust for pain because of the exploratory nature of the study. Regardless, drop CMJ height, which may be a good surrogate of the Achilles tendon’s stretch-shortening function, may provide critical information that could be used for progressing patients through RTS because of its relation to tendon loading. Furthermore, evaluating differences between limbs for hopping plyometric quotient may be helpful in clinical practice when assessing for dysfunction that needs addressed during RTS.

There are several limitations of this study. The inclusion of individuals with bilateral injury may have affected the side-to-side differences detected in our variables of interest. However, the sample is probably more representative of what is seen in clinical practice since bilateral symptoms and asymptomatic structural changes are common in this population.^7,15 Another limitation of this study is that participants could be at any point in the RTS phase. Meaning, some participants were close to a year of running without symptoms, while others still had minor symptoms with running (max pain was 4 out of 10 while running). Despite this limitation, we found significant side-to-side differences in several domains of tendon health, which helps identify factors that could be critical for a successful transition from RTS to return-to-performance. The small sample size with multiple correlations and comparisons is also a limitation of this study, resulting in the possibility of type I and type II error. However, given the pilot nature of this study, no a priori power analysis was completed and non-parametric statistical analyses were used. An additional limitation of this study was the inclusion of persons with midportion or insertional Achilles tendinopathy. We lacked adequate statistical power to evaluate these subgroups separately. Further research is needed to determine whether RTS should be guided by different measures based on location of injury.

5 |. PERSPECTIVE

Runners with Achilles tendinopathy in the RTS phase have side-to-side differences in several measures of pain, structure, and function. Specifically, tasks that are more difficult appear to elicit greater side-to-side differences in pain and function. Furthermore, performance during drop countermovement jumping seems to be associated with Achilles tendon loading behaviors while running. Taken together, our results suggest that measures of pain and lower leg function may aid in clinical decision-making during RTS. Future research is needed to develop RTS criteria that optimizes loading and improves outcomes for those with Achilles tendinopathy.

DATA AVAILABILITY STATEMENT

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