Background/aim The Achilles tendon is a tissue that responds to mechanical loads at a molecular and cellular level. In vitro and in vivo studies have shown that the expression of anabolic and/or catabolic proteins can change within hours of loading and return to baseline levels within 72 h. These biochemical changes have not been correlated with changes in whole tendon structure on imaging. We examined the nature and temporal sequence of changes in Achilles tendon structure in response to competitive game loads in elite Australian football players.
Methods Elite male Australian football players with no history of Achilles tendinopathy were recruited. Achilles tendon structure was quantified using ultrasound tissue characterisation (UTC) imaging, a valid and reliable measure of intratendinous structure, the day prior to the match (day 0), and then reimaged on days 1, 2 and 4 postgame.
Results Of the 18 participants eligible for this study, 12 had no history of tendinopathy (NORM) and 6 had a history of patellar or hamstring tendinopathy (TEN). Differences in baseline UTC echopattern were observed between the NORM and TEN groups, with the Achilles of the TEN group exhibiting altered UTC echopattern, consistent with a slightly disorganised tendon structure. In the NORM group, a significant reduction in echo-type I (normal tendon structure) was seen on day 2 (p=0.012) that returned to baseline on day 4.
Summary There was a transient change in UTC echopattern in the Achilles tendon as a result of an Australian football game in individuals without a history of lower limb tendinopathy.
- Achilles Tendon
- Assessing Physiological Demands of Physical Activity
- Contact Sports
Statistics from Altmetric.com
Tendinopathy—pain and dysfunction in the tendon—is a prevalent condition in athletes that is often associated with overload.1 ,2 The pathoaetiology of tendon injury is currently unknown, with several theories differing in the proposed primary aetiological event.3–5 Cross-sectional and prospective studies have demonstrated tendon pathology on imaging prior to the development of clinical symptoms.6–8 Despite these observations, the response of the tendon to load and the early stages of tendinopathy are poorly understood.
Tendon is a mechanically responsive connective tissue that reacts to changes in load in the short (24–72 h) and long term (12 weeks–years).9 Changes in tendon anabolic and catabolic cellular processes resulting in protein expression can be present the day after an acute bout of exercise.10 These changes have been shown to return to normal levels as early as 4 days after exercise11 suggesting that transient changes exist in tendons as a response to loading. However, detecting these changes using conventional imaging has been limited due to a number of factors (eg spatial resolution, user-dependency, subjective measurements).
Few studies have reported short-term changes in imaging as a result of load. Previous studies showed conflicting results when using imaging as they had been confined to gross measurements of tendon dimension (cross-sectional area (CSA) or anteroposterior (AP) diameter). A number of studies have found no change or an increase in CSA in the short term in response to increased loading.12–14 Short-term changes in AP diameter have been detected in the Achilles and patellar tendons after exercise; a decrease in AP diameter was reported immediately after exercise that returned to normal at 24 h.15 These changes were proposed to be due to a loss of water within the tendon due to creep, however, changes in intratendinous structure were not investigated.
Recently a novel imaging modality, ultrasound tissue characterisation (UTC), using conventional B-mode ultrasound was introduced for the Achilles tendon.16 UTC collects 600 contiguous transverse US images at 0.2 mm intervals and renders a three-dimensional (3D) image of the tendon allowing quantification of tendon structure by measuring the stability of pixel attributes (brightness) over the length of the scan. This scanning technique standardises operator-dependent parameters such as transducer tilt and angle, gain and depth. van Schie et al16 reported high reproducibility with excellent intraobserver and interobserver reliability (intraclass correlation coefficient >0.92).
As UTC semiquantifies tendon structure, it is an ideal research tool to objectively assess different rehabilitation modalities by monitoring tendon integrity. de Vos et al17 found that the mid-substance of the Achilles in patients with tendinopathy did not improve on UTC after a 24-week eccentric loading programme and injection therapy, despite improvements in clinical and functional outcomes. UTC has detected subtle structural changes in response to load in the superficial digital flexor tendon of the thoroughbred horse, an analogy for the Achilles tendon in humans.18
This research aimed to investigate the presence and the time course of short-term change in the Achilles tendon in response to load in elite Australian football players during an Australian Football League (AFL) game. Australian football is a fast paced sport involving repeated high-intensity bursts of running (cumulative distances of up to 15 km) as well as jumping and cutting manoeuvres. Owing to the physical demands of the sport, elastic storage within the lower limb tendons is high. Based on previous studies,18 ,19 we hypothesised that a maximal bout of exercise (AFL game) will change the Achilles tendon echopattern on UTC imaging on days 1 and 2 with a return to baseline by day 4.
An entire elite male Australian football team (21 players, age 23.8±3.01 years, mean±SD) who were selected for an in-season competitive match were recruited for this study. Participants with a history of Achilles tendinopathy were excluded. The history of other lower limb tendinopathy (eg. patellar tendinopathy), excluding the Achilles, was noted from club medical records. All participants were determined as fit and healthy by the club medical officer. The protocol was approved by the Monash University human ethics committee and all participants provided written informed consent prior to participating in the study.
Achilles tendon structure was quantified using UTC imaging. UTC has been shown to be reliable in human16 and in equine tendons, and has been validated against pathological specimens histologically.18 ,20 ,21 A 7–10 MHz linear ultrasound transducer (SmartProbe 10L5, Terason 2000; Teratech) was mounted in a tracking device that moves the transducer automatically along the tendon's long axis recording transverse images at intervals of 0.2 mm over a 12 cm distance (600 axial images). The tracking device standardises transducer tilt, angle, gain, focus and depth.
Coupling gel was applied between skin, an integrated stand-off pad and transducer to optimise contact prior to scanning the Achilles tendon. The participant was positioned standing on a raised level surface with the great toe and knee touching the wall in a standardised lunge position (figure 1). The tracking device was placed on the posterior surface of the Achilles region parallel to the long axis of the tendon. The transducer was aligned with the Achilles insertion at the posterior aspect of the calcaneus and the scan collected in a distal to proximal direction. All scans were taken by a single investigator (SID) who has 4 years experience in UTC imaging.
UTC analysis compounds consecutive transverse grey scale images creating a 3D reconstruction.22 Dedicated UTC algorithms (UTC2010, UTC imaging) quantify the dynamics of grey levels of corresponding pixels in contiguous images over a distance of 25 scans (4.8 mm). Fundamental research revealed that the dynamics of grey levels were strongly related to the architecture and integrity of the histomorphology of the tendon.22 Four validated echo-types can be discriminated and related to tendon integrity: echo-type I represents intact, continuous and aligned fibres and fasciculi, echo-type II represents less continuous and/or more wavy fibres and fasciculi, echo-type III represents a mainly fibrillar matrix and echo-type IV represents complete disintegration, with tendon tissue replaced by an amorphous matrix and fluid (refer to van Schie et al16 for further explanation of echo-types). These echo-types are quantified as relative percentages of the tendon in the region of interest (ROI).
Tendon structure was quantified by selecting an ROI, defined by the margin of the Achilles tendon in the transverse plane. An ROI was selected at the mid-substance of the Achilles (defined as 20 mm proximal to the upper border of the calcaneus) and at regular intervals of 5 mm over a distance of 20 mm. The UTC software automatically interpolated contiguous ROIs between the defined ROIs selected by the investigator, creating a tendon volume in which the proportions of echo-types were quantified.
All participants were scanned with UTC imaging on the day prior to the match (day 0), and then reimaged on days 1, 2 and 4 postgame. The scan analysis was completed in a blinded fashion to participants, tendon pathology history, side and day of scan with scans assigned randomly generated numbers before being passed on for scan analysis. Results were decoded for statistical analysis by another investigator using the key.
All players wore a GPS monitor (Catapult minimax, Catapult Sports) to monitor game loads, that is, the total distance covered for each player. Players did not undertake any other training in the 4 days of the study.
Ten Achilles tendons in a sample population of similar demographics were repeat scanned to test repeated measure reliability. Tendon structure was quantified using the same methods as described above. Standard error of the measurement (SEM)=SD of population ×√(1−ICC) was calculated. The ICC was calculated using a two-way mixed single measures (3,1) for absolute agreement between the repeated scans. The minimum detectable change (MDC=1.96×SEM×√2) were calculated with the results displayed in table 1.
Median and IQRs were calculated for all four echo-types in both limbs on each of the four test days. Tests of normality using Kolmogorov-Smirnov test demonstrated that the data were not normally distributed. Hence, data were analysed using non-parametric statistics. No Bonferroni adjustment was made and was set at 0.05 as this is the first study looking at tendon response to load using this technology. It is reasonable to be comprehensive in the data analysis with the view of not missing potential important findings.23
At baseline, there was no difference between the left and right Achilles for all participants in all four echo-types (data not shown) hence, the left Achilles was randomly selected and subsequently used for all statistical tests to minimise the risk of type I errors.
Difference in the overall echopattern was analysed in participants with and without a history of lower limb tendinopathy (excluding the Achilles) on day 0 using a Mann-Whitney U test. The median and IQR for all echo-types were plotted over the 4 days. In an attempt to reduce the number of statistical tests performed and minimise the risk of type I errors, a related samples Wilcoxon signed-rank test was performed only if the changes in the echopattern on days 1, 2 and 4 were greater than the MDC in comparison to day 0. Linear regression modelling was used to identify potential interaction between distance covered in the game compared with changes in echopattern. Significance was set at p<0.05; all analyses were conducted using statistical package, SPSS V.20.0 (SPSS for Windows, SPSS Inc, Chicago, Illinois, USA).
Three participants were excluded based on a current/prior history of Achilles tendinopathy. Of the remaining 18 participants, 6 were noted to have had current or a history of lower limb tendinopathy excluding the Achilles (patellar or hamstring). All players were currently asymptomatic and were not taking any medication or interventions that may have had a systemic effect.
Baseline tendon structure was investigated in the TEN and NORM group as previous studies have suggested that tendinopathy at one site can alter lower limb biomechanics, which may overload other structures.24 ,25 Significant differences in baseline tendon structure on UTC imaging were observed between the TEN and NORM group. A significant increase in echo-type II was observed in the TEN group (p=0.032), with no significant changes in echo-types I, III and IV (p=0.053, 0.616, 0.053, respectively, table 1). As differences were observed between the two groups, changes in echopattern over the 4 days were analysed within the group.
The NORM group demonstrated a change in echo-types I and II greater than the MDC on day 2 (table 1). Post hoc analysis showed a significant reduction in echo-type I on day 2 in comparison to day 0 (p=0.012), which returned to baseline on day 4 (p=0.594, figure 2). This coincided with a significant increase in echo-type II (p=0.013) on day 2 that returned to baseline on day 4 (p=0.789, table 1). These changes in the UTC echopattern suggest that the Achilles mid-portion exhibited a loss of normal tendon structure 2 days postmaximal load that returned to baseline at day 4.
Changes in the echopattern on UTC were observed over the 4 days in the TEN group that were greater than the MDC (table 1). Post hoc analysis was performed for all changes greater than the MDC with no significant differences observed when compared with baseline (table 1).
The NORM (13.4±1.1 km, median±IQR) and TEN group (12.9±2.0 km, median±IQR) covered similar distances during the game, with no correlation observed between distance covered during the game and change of echopattern on day 2 in the NORM group (R=0.297, 0.237, 0.293 & 0.921, for all echo-types, respectively).
This study demonstrated a loss of normal tendon structure on UTC imaging 2 days after maximal exercise in players with normal tendons and no history of tendinopathy. UTC imaging echo-types have been validated histologically against equine tendons, which have similar structural and compositional properties. The extrapolation of the UTC echo-types to structural features in human tendon is unknown.22 As echo-type I indicates high stability in the grey scale pixels, it reflects homogeneity that corresponds with aligned tendon fibrils within the matrix. Echo-types II, III and IV represent increasing degrees of variability in grey scale pixel brightness. Slight separation and increased waviness of tendon fibrils are represented by echo-type II, disorganised fibrillar matrix correlated with echo-type III, with echo-type IV indicative of a more amorphous collagen structure.
A decrease in echo-type I coinciding with an increase in echo-type II observed on day 2 in comparison to baseline in the NORM group suggests that the normal tendon integrity has been negatively affected. The observation that echo-types III and IV were consistent across the 4 days indicates that there was no increase of disorganised fibrillar or amorphous matrix structure. As echo-types I and II returned to baseline by day 4, this observation supports the concept of a short-duration and fully reversible tendon response without loss of integrity of the collagen matrix. In response to high loads, this transient loss of normal tendon structure on UTC imaging may be considered normal. As it is not ethically possible to collect matching tendon biopsies from elite athletes for histology, the specific structural and extracellular matrix changes that lead to the alteration in echopattern observed in this study are, as yet, unknown.
The changes observed in the Achilles tendon over the course of the 4 days may be a result of a cell driven mechanism4 as the tenocyte is primarily responsible for remodelling of the tendon extracellular matrix in response to mechanical stimuli.26 The continuum of pathogenesis, proposed by Cook and Purdam4 suggests that the tendon responds to overload by increasing the expression of large proteoglycans (aggrecan/versican). This results in an increase in bound water within the ground substance,27 which may lead to matrix disorganisation. Aggrecan has been shown to be upregulated rapidly (<24 h), with 60% of this and other larger proteoglycans being degraded within 3 days.27–29 As the echopattern returns to baseline by day 4, deposition and enzymatic breakdown and clearance of proteoglycans may be responsible for the findings of this study.
Recent findings in a tendon explant model showed increases in bound water content and an increase in large-diameter tendon fibrils when the explant was incubated in phosphate-buffered saline, despite the dry weight of collagen remaining similar to the control.30 This swelling of tendon fibres as well as the separation of tendon fibres was proposed to be mediated by proteoglycan interactions with bound water. Echo-type II is indicative for (reversible) tendon matrix remodelling and/or swelling of tendon fibres, suggesting that the findings of the current study may be explained by alterations in fibre diameter and separation due to increases in bound water and proteoglycan content.
The concept that the tendon is responsive to mechanical stimuli in a matter of days is not new and has been described by a number of authors. A number of studies have shown increases in markers of collagen synthesis and degradation collected from the tendon and peritendinous space; these were elevated 24 h postexercise and remained elevated after 72 h.9 ,19 ,31 However, alterations in the collagenous matrix are unlikely to explain the findings in this study as their restoration would take up to 11 weeks. As echo-types III and IV did not change over the 4 days postexercise, it further supports that collagen fibre remodelling and integrity was not affected.
The reduction in the normal Achilles tendon structure on UTC imaging at baseline between participants with (TEN) and without (NORM) a history of lower limb tendinopathy (other than Achilles tendons) may have substantial implications, yet needs further investigation. No participants in the TEN group had a history of Achilles symptoms or abnormality on imaging, yet exhibited a compromised Achilles tendon structure at baseline. Andersson et al24 described changes within the unloaded tendon (increase in cell number and neovascularisation) in a unilateral Achilles tendon overload model in rabbits. Unfortunately the findings in other tendons were not reported. These changes coincided with increased expression of the neuropeptide, substance P, suggesting a systemic or central nervous system role in the development of pathology despite the absence of load. Systemic upregulation of substance P may lead to disorganisation of tendon structure and account for changes in the echopattern between the two groups. These changes may also be due to genetic susceptibility32 ,33 changes in kinetic chain biomechanics adversely loading the Achilles25 or be related to systemic conditions.34 Regardless of the mechanism, these findings warrant further investigation as confirmation and explanation of the underlying mechanism, that structural changes in one tendon affect other tendons of the body, is of considerable clinical importance if confirmed in further research.
This study demonstrated a transient change in the Achilles tendon's response on UTC imaging to a single game, however, the effect of repeated tendon load over a season was not assessed. Without sufficient recovery time, repeat tendon loading may result in cumulative tendon matrix adaptation or degradation. Malliaras et al35 demonstrated that tendons can transition forward and back along the tendon pathology continuum4 over the course of a volleyball season. Monitoring tendons over a greater period of time with UTC may provide insight into variations in tendon morphology across the season. Furthermore it may lead to a better understanding of the ‘point of no-return’ where homeostatic capacities of the tendon are exceeded.36 Identifying critical changes in the tendon pathology continuum that lead to non-reversible degeneration may help in developing a prognostic criteria for decision making in the prevention and treatment of tendinopathy.
This pilot study had a number of limitations. A major limitation of this study was the small sample size in both groups. Also, this pragmatic study within the elite athletic environment did not allow standardised load across participants. In future, a standardised tendon specific loading protocol to induce a tendon response should be attempted. The clinical relevance of this tendon response needs to be determined, as the current study is unable to provide any insight whether this response is a pathological or adaptive response.
The results of this study need to be considered in the context of previous studies published using UTC.16 ,37 ,38 The technique described in this study differs in five critical points; (1) the transducer is moved automatically by a mechanically-driven arm compared with manually; (2) the transducer captures images over 12 compared with 9.6 cm in previous studies; (3) the stability of brightness over contiguous transverse images is quantified across 4.8 vs 4.2 mm; (4) UTC scans were performed in a standing lunge position in this study and (5) the region of tendon selected for quantification. Caution is advised in using the UTC data from this study, or others, as reference data unless the scanning parameters are identical.
The sensitivity and reproducibility of UTC to detect and monitor these changes offers new insight into the pathophysiology of tendon responses and opens a pathway to investigate tendons and their response to loading. As understanding of the response of normal tendons to load improves, so will the opportunity to explore various pathological states, the processes involved in their pathogenesis and to design appropriate exercise protocols for prevention and rehabilitation.
This study showed the normal Achilles tendon may respond in the short-term (reduction in UTC echopattern corresponding with a loss of normal tendon structure) to exercise loads in individuals with no history of lower limb tendinopathy. These findings are preliminary in a small cohort, yet were similar to the results published in Docking et al18 who reported a similar tendon response on UTC in racehorses. A normal acute and transient response in tendons as a result of load application has been suggested from basic science research. This study suggests that UTC may be able to detect these changes in the tendon structure in response to load, with the potential to elucidate the clinical relevance of this response. Future studies with greater numbers are indicated to further investigate the factors as well as studies to investigate long-term effects of repeat application of bouts of exercise loads.
What are the new findings?
Normal tendon structure, as quantified by ultrasound tissue characterisation (UTC) imaging (echo-type I), was slightly reduced within the mid-substance of the Achilles tendon 2 days postmaximal exercise, with tendon structure on UTC imaging returning to baseline after 4 days.
The Achilles tendon in participants with a history of lower limb tendinopathy excluding the Achilles exhibited subtle disorganisation in the tendon structure on UTC imaging in comparison to participants without a history of tendinopathy.
How might it impact on clinical practice in the near future?
Tendon response, where normal tendon structure is reduced on ultrasound tissue characterisation (UTC), 2 days postmaximal exercise should be considered in the development of loading programmes. Future research is required to investigate the effect of repeat maximal load on tendon structure.
UTC may be useful as a monitoring tool in the elite athletic population to guide loading, with the prospect of informing load reduction prior to the development of symptoms.
Achilles tendon structure may be compromised on UTC imaging in individuals with a history of other lower limb tendinopathies. Further research is required to understand the mechanism that underpins poorer UTC imaging echopattern.
The authors would like to thank the players and staff at the Carlton Football Club, Melbourne, Australia for participation and assistance during this study.
Correction notice This paper has been amended since it was published Online First. It has been brought to the attention of the authors by Dr Rod Whiteley that they had inadvertently calculated the minimum detectable change (MDC) incorrectly. In calculating the MDC (MDC=1.96×SEM x√2), the standard error of the mean was used rather than the standard error of the measure. Based on previous studies, the intra-class correlation (ICC) was calculated using a two-way mixed single measures (3,1) for absolute agreement. From this, the standard error of the measurement (SEM=SD×√(1-ICC)) and the MDC was calculated. The recalculated MDC's for each of the echo-types (0.91%, 0.94%, 0.25% & 0.59% for each echo-type respectively) has not altered any of the findings for this study.
Contributors All the authors were involved in the conception and design of the study. SDR, JLC and SID were responsible for the collection, analysis and interpretation of the data and creation of the manuscript. SDR, JLC and SID are responsible for the overall content as guarantors.
Funding This paper was supported by the Australian Centre for Research into Sports Injury and its Prevention, which is one of the International Research Centres for Prevention of Injury and Protection of Athlete Health supported by the International Olympic Committee.
Competing interests SR, JC and SD attended a conference in April 2013 supported by UTC imaging company. For the remaining authors none were declared.
Patient consent Obtained.
Ethics approval Monash University Human Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.