Effect of resistance exercise dose components for tendinopathy management: a systematic review with meta-analysis

Objective

To investigate potential moderating effects of resistance exercise dose components including intensity, volume and frequency, for the management of common tendinopathies.

Design

Systematic review with meta-analysis and meta-regressions.

Data sources

Including but not limited to: MEDLINE, CINAHL, SPORTDiscus, ClinicalTrials.gov and ISRCTN Registry.

Eligibility criteria for selecting studies

Randomised and non-randomised controlled trials investigating resistance exercise as the dominant treatment class, reporting sufficient information regarding ≥2 components of exercise dose.

Results

A total of 110 studies were included in meta-analyses (148 treatment arms (TAs), 3953 participants), reporting on five tendinopathy locations (rotator cuff: 48 TAs; Achilles: 43 TAs; lateral elbow: 29 TAs; patellar: 24 TAs; gluteal: 4 TAs). Meta-regressions provided consistent evidence of greater pooled mean effect sizes for higher intensity therapies comprising additional external resistance compared with body mass only (large effect size domains: β _{BodyMass: External} = 0.50 (95% credible interval (CrI): 0.15 to 0.84; p=0.998); small effect size domains (β _{BodyMass: External} = 0.04 (95% CrI: −0.21 to 0.31; p=0.619)) when combined across tendinopathy locations or analysed separately. Greater pooled mean effect sizes were also identified for the lowest frequency (less than daily) compared with mid (daily) and high frequencies (more than once per day) for both effect size domains when combined or analysed separately (p≥0.976). Evidence for associations between training volume and pooled mean effect sizes was minimal and inconsistent.

Summary/conclusion

Resistance exercise dose is poorly reported within tendinopathy management literature. However, this large meta-analysis identified some consistent patterns indicating greater efficacy on average with therapies prescribing higher intensities (through inclusion of additional loads) and lower frequencies, potentially creating stronger stimuli and facilitating adequate recovery.

Keywords: Tendinopathy, Exercise Therapy, Systematic review

What is already known on this topic?

• Tendinopathy is a prevalent condition in both athletic and non-athletic populations commonly affecting the Achilles, rotator cuff, lateral elbow, patellar and hip tendons.

• Exercise therapy is the main mode of conservative treatment for tendinopathies with a focus on resistance exercise, which is shown to be effective in improving patient outcomes.

• Little is known about the effect of different resistance exercise dose components, including intensity, volume and frequency, on patient improvement.

• Previous systematic reviews and meta-analyses that have attempted to investigate exercise dose in tendinopathy have been limited to small numbers of studies.

What this study adds?

• This extensive systematic review with meta-analysis included 91 studies (126 treatment arms), and identified common patterns despite large variations across interventions.

• Interventions involving higher intensity resistance exercise, with the addition of external loads, showed greater efficacy compared with body mass only exercise.

• Greater efficacy was seen with interventions performed less frequently, potentially allowing for adequate recovery, compared with higher frequencies of once or more per day.

• There were no consistent results or patterns identified from analyses of resistance exercise volume.

How this study might affect research, practice or policy

• Clinicians prescribing resistance exercise therapy for tendinopathy should consider including higher intensities of resistance exercise (through addition of external loads) and allowing adequate recovery between sessions.

• We urge future research on exercise interventions to make use of reporting guidelines and to include full details of all components of exercise dose (intensity, volume, frequency).

Inclusion criteria

Inclusion criteria and methods were influenced by the project aims, the results of an initial scoping review¹⁵ and two subsequent stakeholder workshops (n=13). The first included nine individuals who delivered exercise therapy for tendinopathy and had an academic interest. The second included four women with lived experience. Finally, an online survey (n=26) was conducted to gather the views of a more diverse international sample of purposefully selected clinicians and academics. A completed Preferred Reporting Items for Systematic review and Meta-Analysis checklist for reporting of systematic reviews can be found in online supplemental file 1. The inclusion criteria were framed according to a modified PICOS (participant, intervention, comparator, outcomes, study type) approach which also included context.

Search strategy

The search strategy used for this study was part of a larger search conducted to scope the entire exercise for tendinopathy research base. We employed a three-step search strategy. First, a limited search of MEDLINE and CINAHL using initial keywords (MH tendinopathy OR TX tendin* OR TX tendon*) AND (MH exercise OR TX exercis*) was conducted to develop a full search strategy. Second, the full search strategy was adapted to each database and applied systematically to: MEDLINE, CINAHL, AMED, Embase, SPORTDiscus, Cochrane library (Controlled trials, Systematic reviews), JBI Evidence Synthesis, PEDRo and Epistemonikos. The following trial registries were also searched: ClinicalTrials.gov, ISRCTN The Research Registry, EU-CTR (European Union Clinical Trials Registry), ANZCTR (Australia and New Zealand Clinical Trials Registry) (all search strategies are presented in online supplemental file 4). Finally, the third step involved a search of cited and citing articles using Scopus and hand-searching 130 systematic reviews that were identified to include information relevant to exercise therapy and tendinopathy. As a final check, the list of identified studies was sent to experts external to the research team to identify any potentially missing studies. Research studies published in languages other than English were translated via Google Translate or international collaborations of the review team. Searches were initiated from 1998 as (1) the heavy load eccentric calf-training protocol for Achilles tendinosis by Alfredson et al ²⁰ was published in 1998 and may be considered seminal work in the field of tendinopathy, and (2) there has been a proliferation of research on exercise interventions for tendinopathies post 1998. The final search was conducted on 25 March 2022.

Study selection

Proquest Refworks was used to manage references and remove duplicates, before importing to Covidence (Melbourne, Australia) for screening and further de-duplication. Each title and abstract was independently reviewed by any two members of the review team (PAS/KC/LA/RM/LG/EP/JSCS/AVP). Full-texts of included studies were similarly screened independently by any two team members. Conflicts were resolved by discussion or by a third reviewer.

Data extraction

Following extraction training, data were extracted independently by eight members of the review team (PAS/KC/LA/RM/LG/EP/JSCS/AVP) into prepiloted excel spreadsheets and independently coded as described in the accompanying extraction codebook (online supplemental file 5). Each entry on the spreadsheet was double-checked by a different member of the team. Where pre-post intervention data were not presented in text but in figures, data were extracted using PlotDigitizer V.2.6.8 Windows (WebPlotDigitizer - Copyright 2010–2021 Ankit Rohatgi (automeris.io)).

Risk of bias assessment

Risk of bias was assessed using the earlier version of the Cochrane risk of bias (RoB) tool²¹ since a recent review of RoB tools in tendinopathy management studies did not identify one tool as being superior to the others.²² Furthermore, it allowed us to streamline the process by combining with RobotReviewer,²³ a machine learning software that semi-automates the Cochrane tool. A risk of bias judgement was made for each outcome and time point within studies for each of the seven domains²¹ and reported as either ‘low risk’, ‘high risk’ or ‘uncertain’ when there was insufficient detail or the outcome was not addressed. RobotReviewer was used to make initial assessments on domains 1, 2 and 3 and validated manually using the extracted free text to agree on a final selection of risk of bias. This semi-automated process provided greater efficiency and consistency during the review process. Results are presented using an overall summary risk of bias assessment, obtained for each domain by selecting the mode risk category across all outcomes and time points. An assessment was made by any two members of the team (AVP, JSCS, RM, EP, LG] with comments made to justify scoring and regular consultation between team members where uncertainties arose.

Coding of resistance exercise therapies

Attempts were made to code exercise dose components (intensity, volume and frequency) for each study; however, sufficient information was not always available to code all three components. Coding of exercise intensity was initially achieved by identifying whether exercise load was prescribed in absolute (eg, kilogrammes when using dumbbells or isoinertial loads) or relative terms (as a percentage of the maximum load that can be lifted) and the magnitude of the load recorded. Additionally, a binary coding was used to identify whether exercise was performed with body mass only (eg, whole body mass or mass of a limb), or with the addition of external loads (such as a loaded backpack, dumbbell or elastic resistance). Exercise volume was coded by quantifying the number of sets and repetitions. Exercise frequency was recorded as the total number of resistance exercise sessions performed per week (including where there were multiple sessions a day). In cases where several resistance exercises were prescribed, intensity and volume were extracted for the primary resistance exercise only, which we defined as the exercise that was the focus of the paper or, if unclear, whichever had the greatest volume. In cases where exercise dose progressed, we took the average value for the primary exercise. Where progressions led from an initial mobility component to a resistance exercise focus, the latter was extracted.

Statistical analysis

The purpose of the meta-analysis was to investigate responses to exercise therapies where resistance exercise was the dominant treatment class. A broad modelling perspective was selected where outcomes across a range of tendinopathies and outcome domains were combined to investigate whether central estimates (eg, pooled mean) were associated with different levels of moderator variables representing exercise dose (frequency, intensity or volume). Due to the use of different outcome domains and different tests within the same outcome domain, pooling of data required standardisation. This was achieved using the standardised mean difference (SMD_pre) effect size, dividing the mean group change by the pre-intervention SD. Where baseline SD values were not presented these were estimated using statistical information presented²⁴ (eg, CIs, SEs, t values, p values, F values) or imputed based on the simple linear regression quantifying the relationship between the log-transformed means (explanatory) and log-transformed SDs (response) from studies with complete data.²⁵ Where required, SMD_pre values were reflected by multiplying by –1 to ensure that positive values represented an improved clinical effect. Where multiple outcomes were reported from the same study (different outcomes and/or the same outcome at multiple time points), all possible SMD_pre values were calculated and included in the meta-analysis models. To account for covariances created, all meta-analyses were conducted using a nested four-level model²⁶ comprising the individual study (level 4), the outcome (level 3), the measurement occasion (level 2) and the sampling variance (level 1) levels. A comprehensive description of the model and further details of statistical analysis can be found in online supplemental file 6.

Confidence in cumulative evidence

Assessments were made using the Grading of Recommendations Assessment Development and Evaluation guidelines²⁷ in addition to recommendations on transparent reporting of evidence for tendinopathy management.²⁸ Confidence in evidence was assessed at the outcome level with: (1) overall risk of bias ranked as high, low or unclear risk, as identified by the mode rating across all data in the specific analysis; (2) inconsistency assessed based on meta-analysis results and comparisons of central and variance parameter estimates (downgraded where σr >0.9 γ^0 ); (3) imprecision judged by the number of available data points (studies, treatment arms, outcome measures) and the width of credible intervals for central estimates; (4) indirectness identified as low risk for all outcomes based on inclusion criteria from our previous scoping review and stakeholder recommendation; and (5) small study effects assessed by visual inspection of effect size distribution and sampling variance (downgraded when substantive number of points outside bounds). Overall confidence in evidence for each analysis was recorded as either high, moderate, low or very low. Categorisations began with high confidence in cumulative evidence and were downgraded to a level for each domain not judged as low risk.

Protocol deviations

A deviation from our PROSPERO registered protocol in relation to the included tendinopathies was made. We intended to include all tendinopathies but were guided by identification of studies reporting on resistance exercise following our scoping review. Our final inclusion criteria incorporated RCRSP, lateral elbow, patellar, Achilles and gluteal tendinopathies.

Equality, diversity and inclusion statement

The authors on this project were chosen on merit and came from a diverse range of backgrounds, occupations and levels of seniority. Although we excluded studies from countries not on the ‘very high’ Human Development Index (HDI) list, this was done to make findings more generalisable to the UK.

Study selection

The search strategy identified a total of 12 379 potential studies, with 6944 remaining following de-duplication (figure 1). After title and abstract screening 440 studies were retained for full-text screening. Of these studies, a further 330 were excluded (online supplemental file 7A: Excluded studies with reasons reference list) based primarily on insufficient description of the exercise stimulus (141 studies) and not including exercise-only treatment arms (79 studies). In total, data from 110 studies comprising 148 treatment arms and 3953 participants were included in the meta-analyses (online supplemental file 7B, C: Table of included studies and reference list). Exercise therapies for the treatment of five different tendinopathies (RCRSP: 48 (32%) treatment arms; Achilles: 43 (29.0%) treatment arms; lateral elbow: 29 (20%) treatment arms; patellar: 24 (16%) treatment arms; and gluteal: 4 (3%) treatment arms) were identified (table 1). Over half of the treatment arms (82/55%) comprised resistance only therapies, with the remaining predominantly including additional flexibility exercises (table 1). The dominant resistance exercise treatments are presented in table 1, with eccentric only exercise the most common for Achilles (79% of relevant treatment arms), lateral elbow (62% of relevant treatment arms) and patellar (38% of relevant treatment arms) tendinopathies; and both concentric and eccentric resistance exercise most common for gluteal (75% of relevant treatment arms) and RCRSP (73% of relevant treatment arms). Overall, eccentric-only (68 treatment arms) was the most common dominant treatment, followed by concentric and eccentric (55 treatment arms) then isometric (16 treatment arms).

Risk of bias and confidence in cumulative evidence

RoB and confidence in evidence assessments are presented for the primary meta-analyses and more broadly in online supplemental file 8. Summarised according to studies, the most frequent risk of bias for randomised controlled trials was blinding of participants (40% studies high risk of bias) and ‘other bias’ (54% studies high risk of bias). Similarly, ‘other bias’ was also the most frequent risk of bias for non-randomised studies as assessed by the RoB tool. In general, confidence in evidence was frequently low based on imprecision due to wide credible intervals and inconsistency due to large between study variance estimates. Overall confidence in cumulative evidence varied from very low to moderate with low confidence most commonly identified.

Resistance exercise intensity

Of the 148 treatment arms included, 123 provided sufficient information to categorise the intensity as lower intensity in the form of body mass only (31 treatment arms; 25%), or higher intensity with the addition of external resistance (92 treatment arms; 75%) prescribed based on absolute loads (eg, addition of weights to a backpack, isoinertial loads, resistance band and dumbbells) or percentage of a maximum. Meta-regressions provided consistent evidence of greater pooled mean effect sizes for increased training intensity with the addition of external loads. Primary meta-analyses pooling data across all tendinopathy locations identified median increases of β _{BodyMass: External} = 0.50 (95% credible interval (CrI): 0.15 to 0.84; p=0.998) for outcomes generating large effect sizes, and an increase of β _{BodyMass: External} = 0.04 (95% CrI: −0.21 to 0.31; p=0.619) for outcomes generating small effect sizes (individual levels presented in table 2). Similarly, point estimates indicated greater pooled mean values for the addition of external resistance for all analyses separated by tendinopathy location for which there was sufficient data (online supplemental file 9A).

Frequency of resistance exercise

Of the 148 treatment arms included, 135 provided sufficient information to categorise the frequency as low frequency (less than daily: 48 treatment arms; 36%), moderate frequency (daily: 34 treatment arms; 25%) or high frequency (more than once per day: 53 treatment arms; 39%). Consistent evidence of a moderating effect was also identified for resistance exercise frequency with greater pooled mean effect sizes identified for the lowest frequency of less than once per day. Primary meta-analyses pooling data across all tendinopathy locations identified median increases of β _{<Daily: Daily} = −0.50 (95% CrI: −0.88 to −0.11; p=0.992) between less than once per day and once per day, and β _{<Daily: >Daily} = −0.44 (95% CrI: −0.829 to −0.05; p=0.951) between less than once per day and more than once per day for outcomes generating large effect sizes. Similarly, median increases of β _{<Daily: Daily} = −0.32 (95% CrI: −0.55 to −0.09; p=0.999) and β _{<Daily: Daily} = −0.21 (95% CrI: −0.42 to −0.00; p=0.976) were identified across the comparisons for outcomes generating small effect sizes (individual levels presented in table 3). Consistent evidence of increased pooled mean effect sizes for resistance exercise performed less than once per day was also obtained when analyses were separated by tendinopathy location (online supplemental file 9B). In contrast, effect size estimates tended to be similar for exercising once per day or more than once per day (table 3) with wide overlap of potential values also identified when analyses were separated by tendinopathy location (online supplemental file 9B).

Resistance exercise volume

Resistance exercise volume was categorised for 128 treatment arms as the product of the number of sets and repetitions for the primary resistance exercise. The most common number of total repetitions was 45 (eg, 3 sets of 15 repetitions) and this accounted for almost half of the training interventions (51 treatment arms; 40%). As a result, training volume was coded as a binary variable characterised as lower volume (<45 total repetitions: 67 treatment arms; 52%) and higher volume (≥45 total repetitions: 61 treatment arms; 48%). In general, considerable overlap was identified between pooled mean effect size estimates of lower and higher volume exercise including primary meta-analyses of outcomes generating large effect sizes (β _{Lower: Higher} = −0.02 (95% CrI: −0.40 to 0.37; p=0.553)) and outcomes generating small effect sizes (β _{Lower: Higher} = −0.14 (95% CrI: −0.35 to 0.09; p=0.782; individual levels presented in table 4)). While the median point estimates from the primary meta-analyses favoured lower volume exercise, this ordering was not consistently maintained when analyses were separated by tendinopathy location (online supplemental file 9C).

Combined analysis

As a final analysis, a meta-regression including the above intensity, frequency and volume variables were included to assess for differences in the pooled mean effect size while controlling for each other across all tendinopathy locations. A total of 76 studies (101 treatment arms) provided sufficient information for simultaneous coding of all three dose variables for outcomes generating large effect sizes, and 40 studies (53 treatment arms) provided sufficient information for outcomes generating small effect sizes. Results were consistent with analyses conducted individually on dosing variables, with evidence of increased pooled means with greater intensity for both outcomes generating large ( βBodyMass:External= 0.38 (95% CrI: 0.00 to 0.77; p=0.975)) and small (β _{BodyMass: External} = 0.17 (95% CrI: −0.11 to 0.46; p=0.888)) effect sizes. Similarly, evidence of increased pooled means was obtained for the lowest frequency therapies for both outcomes generating large (β _{<Daily: Daily} = −0.60 (95% CrI: −1.1 to −0.13; p=0.993); β _{<Daily: >Daily} = −0.32 [95% CrI:−0.76 to −0.02; p=0.977)) and small (β _{<Daily: Daily} = −0.37 (95% CrI:−0.63 to −0.08; p=0.995); β _{<Daily: >Daily} = −0.26 (95% CrI:−0.51 to −0.01; p=0.976)) effect sizes. Finally, minimal evidence was obtained for a moderating effect of training volume for outcomes generating either large (β _{Lower: Higher} = −0.12 (95% CrI: −0.43 to 0.17; p=0.614)) or small (β _{Lower: Higher} = −0.08 (95% CrI: −0.30 to 0.13; p=0.707)) effect sizes.

Limitations

We did not search Web of Science, however, together with our information scientist we are confident that our search was comprehensive and rigorous. We may have missed some potentially relevant high-quality studies by excluding those that were not from countries ranked ‘very high’ on the HDI (eg, Brazil and South Africa). However, this number was small and allowed us to generalise the findings to the UK. One of the limitations of this study, and a challenge for future evidence syntheses, is the lack of clear reporting. We found that in general, exercise volume and frequency were better reported, with reporting of intensity often poor. Similarly, our review identified that although load progression was frequently stated, studies rarely reported the actual loads or intensity used. The use of reporting guidelines such as Consolidated Standards of Reporting Trials³⁷ or Consensus on Exercise Reporting Template³⁸ in primary research would greatly enhance future reviews. Understandably, progressions are matched to individuals, however, more detailed reporting of loads prescribed and ultimately used across participants would be useful for future evidence syntheses and to better inform clinicians. Relative intensity measures such as per cent of maximum repetition are likely to provide the most useful information for future evidence syntheses and the most precise comparisons. Another limitation to note is that we extracted resistance dose data for the primary exercise, therefore other exercises prescribed as part of the wider intervention were not accounted for in analysis. This means that variables such as exercise volume may not be fully representative of the true volume of overall exercise performed, but rather specifically the dominant resistance exercise. Although most studies made clear which exercise was the focus of investigation, it was unclear in a small number and in these cases, we identified the exercise with the higher volume. The comparability of our RoB judgements with recent reviews using the updated RoB2 tool may be limited due to poor inter-tool reliability with the older version of Cochrane’s RoB tool.²² While we used RobotReviewer pragmatically for efficiency, we wish to point out that decisions were checked to improve accuracy and consistency.

An additional substantive limitation of this review includes the use of non-controlled effect sizes. This approach was adopted due to the ability to greatly increase the amount of data available and address limitations of previous reviews based largely on small sample sizes and subsequent decisions not to quantitatively synthesise results. The major limitation of this approach is the potential for unbalanced treatment moderators including a range of intervention and population characteristics to associate with different levels of the dose variables defined, thus biasing results. Given the likely interaction between intensity, volume and frequency, we attempted to control for these interactions by including a more complete meta-regression with all three variables included. While the results of the analysis supported those obtained with the individual meta-regressions, there are likely to be many other effect moderators including intervention duration, follow-up duration, adherence and baseline characteristics of patients^{39 40} that may have been imbalanced and could not be controlled. Additionally, in clinical settings there is potential for extreme values to occur and if these errors are asymmetric then regression to the mean effects can create poor estimates further limiting non-controlled effect sizes. In a previous analysis conducted with similar data we showed that effect size values are greatly influenced by outcome domains and could be summarised by a binary classification.¹⁸ As a result, we conducted analyses in the present review based on outcome domains that tended to generate large and small effects. However, it is possible that this binary classification represents too much of a simplification and imbalances in outcomes may also have biased results. Large increases in data were obtained by pooling results across the different tendinopathy locations. This approach was adopted based on our previous analysis indicating that the distribution of effect sizes following exercise therapy is likely to be similar.¹⁸ However, where sufficient data were available to conduct meta-regressions for individual tendinopathy locations, results tended to be consistent. Finally, another limitation of the review included the confidence in cumulative evidence which was most frequently identified as low and in a number of cases very low. This was predominantly due to extensive heterogeneity in studies resulting in issues of inconsistency and imprecision in effect size estimates.

Patient consent for publication

Not applicable.

Ethics approval

Not applicable.

• 1. Fu FH, Wang J-C, Rothrauff BB. Best practice tendinopathy. London: BMJ; 2019. [Google Scholar]
• 2. Ackermann PW, Renström P. Tendinopathy in sport. Sports Health 2012;4:193–201. 10.1177/1941738112440957 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 3. Hopkins C, Fu S-C, Chua E, et al. Critical review on the socio-economic impact of tendinopathy. Asia Pac J Sports Med Arthrosc Rehabil Technol 2016;4:9–20. 10.1016/j.asmart.2016.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 4. Scott A, Squier K, Alfredson H, et al. Icon 2019: international scientific tendinopathy symposium consensus: clinical terminology. Br J Sports Med 2020;54:260–2. 10.1136/bjsports-2019-100885 [DOI] [PubMed] [Google Scholar]
• 5. Alizadehkhaiyat O, Fisher AC, Kemp GJ, et al. Pain, functional disability, and psychologic status in tennis elbow. Clin J Pain 2007;23:482–9. 10.1097/AJP.0b013e31805f70fa [DOI] [PubMed] [Google Scholar]
• 6. Abat F, Alfredson H, Cucchiarini M, et al. Current trends in tendinopathy: consensus of the ESSKA basic science Committee. Part I: biology, biomechanics, anatomy and an exercise-based approach. J Exp Orthop 2017;4:18. 10.1186/s40634-017-0092-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 7. Alfredson H. The chronic painful Achilles and Patellar tendon: research on basic biology and treatment. Scand J Med Sci Sports 2005;15:252–9. 10.1111/j.1600-0838.2005.00466.x [DOI] [PubMed] [Google Scholar]
• 8. O’Neill S, Watson PJ, Barry S. Why are eccentric exercises effective for Achilles tendinopathy? Int J Sports Phys Ther 2015;10:552–62. [PMC free article] [PubMed] [Google Scholar]
• 9. Young JL, Rhon DI, Cleland JA, et al. The influence of exercise dosing on outcomes in patients with knee disorders: a systematic review. J Orthop Sports Phys Ther 2018;48:146–61. 10.2519/jospt.2018.7637 [DOI] [PubMed] [Google Scholar]
• 10. Wasfy MM, Baggish AL. Exercise dose in clinical practice. Circulation 2016;133:2297–313. 10.1161/CIRCULATIONAHA.116.018093 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11. Silbernagel KG. Does one size fit all when it comes to exercise treatment for Achilles tendinopathy? J Orthop Sports Phys Ther 2014;44:42–4. 10.2519/jospt.2014.0103 [DOI] [PubMed] [Google Scholar]
• 12. Meyer A, Tumilty S, Baxter GD. Eccentric exercise protocols for chronic non-insertional Achilles tendinopathy: how much is enough? Scand J Med Sci Sports 2009;19:609–15. 10.1111/j.1600-0838.2009.00981.x [DOI] [PubMed] [Google Scholar]
• 13. Krämer R, Lorenzen J, Vogt PM, et al. Systematic review about eccentric training in chronic Achilles tendinopathy. Sportverletz Sportschaden 2010;24:204–11. 10.1055/s-0029-1245820 [DOI] [PubMed] [Google Scholar]
• 14. Young JL, Rhon DI, de Zoete RMJ, et al. The influence of dosing on effect size of exercise therapy for musculoskeletal foot and ankle disorders: a systematic review. Braz J Phys Ther 2018;22:20–32. 10.1016/j.bjpt.2017.10.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 15. Alexander L, Shim J, Harrison I, et al. Exercise therapy for tendinopathy: A scoping review mapping interventions and outcomes. SportRχiv [Preprint] 2021. 10.51224/SRXIV.90 [DOI]
• 16. Lewis J. Rotator cuff related shoulder pain: assessment, management and uncertainties. Man Ther 2016;23:57–68. 10.1016/j.math.2016.03.009 [DOI] [PubMed] [Google Scholar]
• 17. Requejo-Salinas N, Lewis J, Michener LA, et al. International physical therapists consensus on clinical descriptors for diagnosing rotator cuff related shoulder pain: a Delphi study. Braz J Phys Ther 2022;26:100395. 10.1016/j.bjpt.2022.100395 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 18. Swinton P, Shim J, Pavlova A. Empirically derived guidelines for interpreting the effectiveness of exercise therapy for tendinopathies: a meta-analysis. SportRχiv 2022. 10.1136/bmjsem-2022-sportskongres.2 [DOI] [Google Scholar]
• 19. United Nations Development Programme . Human development reports. New York: United Nations Development Programme; 2020. Available: https://hdr.undp.org/data-center/human-development-index#/indicies/HDI [Google Scholar]
• 20. Alfredson H, Pietilä T, Jonsson P, et al. Heavy-load eccentric calf muscle training for the treatment of chronic Achilles tendinosis. Am J Sports Med 1998;26:360–6. 10.1177/03635465980260030301 [DOI] [PubMed] [Google Scholar]
• 21. Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928. 10.1136/bmj.d5928 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 22. Challoumas D, Millar NL. Risk of bias in systematic reviews of tendinopathy management: are we comparing apples with oranges? Transl Sports Med 2021;4:21–37. 10.1002/tsm2.196 [DOI] [Google Scholar]
• 23. Marshall IJ, Kuiper J, Wallace BC. RobotReviewer: evaluation of a system for automatically assessing bias in clinical trials. J Am Med Inform Assoc 2016;23:193–201. 10.1093/jamia/ocv044 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24. Higgins JPT, Green S, eds. Chapter 7.7.3 data extraction for continuous outcomes. In: Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration, 2011. Available: www.handbook.cochrane.org [Google Scholar]
• 25. Marinho VCC, Higgins JPT, Logan S, et al. Fluoride mouthrinses for preventing dental caries in children and adolescents. Cochrane Database Syst Rev 2003:CD002284. 10.1002/14651858.CD002284 [DOI] [PubMed] [Google Scholar]
• 26. Swinton P, Burgess K, Hall A, et al. A Bayesian approach to interpret intervention effectiveness in strength and conditioning Part 1: a meta-analysis to derive context-specific thresholds. SportRχiv 2021. Available: 10.51224/SRXIV.9 [Google Scholar]
• 27. Schünemann H, Brożek J, Guyatt G, et al. Grade Handbook. In: Handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach. Available: https://gdt.gradepro.org/app/handbook/handbook.html [Google Scholar]
• 28. Challoumas D, Millar NL. Do we need to improve the reporting of evidence in tendinopathy management? A critical appraisal of systematic reviews with recommendations on strength of evidence assessment. BMJ Open Sport Exerc Med 2021;7:e000920. 10.1136/bmjsem-2020-000920 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29. Malliaras P, Kamal B, Nowell A, et al. Patellar tendon adaptation in relation to load-intensity and contraction type. J Biomech 2013;46:1893–9. 10.1016/j.jbiomech.2013.04.022 [DOI] [PubMed] [Google Scholar]
• 30. Bohm S, Mersmann F, Arampatzis A. Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults. Sports Med Open 2015;1:7. 10.1186/s40798-015-0009-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 31. Brumitt J, Cuddeford T. Current concepts of muscle and tendon adaptation to strength and conditioning. Int J Sports Phys Ther 2015;10:748–59. [PMC free article] [PubMed] [Google Scholar]
• 32. Allison GT, Purdam C. Eccentric loading for Achilles tendinopathy -- strengthening or stretching? Br J Sports Med 2009;43:276–9. 10.1136/bjsm.2008.053546 [DOI] [PubMed] [Google Scholar]
• 33. Maenhout AG, Mahieu NN, De Muynck M, et al. Does adding heavy load eccentric training to rehabilitation of patients with unilateral subacromial impingement result in better outcome? A randomized, clinical trial. Knee Surg Sports Traumatol Arthrosc 2013;21:1158–67. 10.1007/s00167-012-2012-8 [DOI] [PubMed] [Google Scholar]
• 34. Stasinopoulos D, Stasinopoulos I, Pantelis M, et al. Comparison of effects of a home exercise programme and a supervised exercise programme for the management of lateral elbow tendinopathy. Br J Sports Med 2010;44:579–83. 10.1136/bjsm.2008.049759 [DOI] [PubMed] [Google Scholar]
• 35. Stasinopoulos D, Stasinopoulos I. Comparison of effects of eccentric training, eccentric-concentric training, and eccentric-concentric training combined with isometric contraction in the treatment of lateral elbow tendinopathy. J Hand Ther 2017;30:13–9. 10.1016/j.jht.2016.09.001 [DOI] [PubMed] [Google Scholar]
• 36. Malliaras P, Johnston R, Street G, et al. The efficacy of higher versus lower dose exercise in rotator cuff tendinopathy: a systematic review of randomized controlled trials. Arch Phys Med Rehabil 2020;101:1822–34. 10.1016/j.apmr.2020.06.013 [DOI] [PubMed] [Google Scholar]
• 37. Rennie D. Consort revised -- improving the reporting of randomized trials. JAMA 2001;285:2006–7. 10.1001/jama.285.15.2006 [DOI] [PubMed] [Google Scholar]
• 38. Slade SC, Finnegan S, Dionne CE, et al. The consensus on exercise reporting template (CERT) applied to exercise interventions in musculoskeletal trials demonstrated good rater agreement and incomplete reporting. J Clin Epidemiol 2018;103:120–30. 10.1016/j.jclinepi.2018.07.009 [DOI] [PubMed] [Google Scholar]
• 39. Littlewood C, Malliaras P, Chance-Larsen K. Therapeutic exercise for rotator cuff tendinopathy: a systematic review of contextual factors and prescription parameters. Int J Rehabil Res 2015;38:95–106. 10.1097/MRR.0000000000000113 [DOI] [PubMed] [Google Scholar]
• 40. Voleti PB, Buckley MR, Soslowsky LJ. Tendon healing: repair and regeneration. Annu Rev Biomed Eng 2012;14:47–71. 10.1146/annurev-bioeng-071811-150122 [DOI] [PubMed] [Google Scholar]
• 41. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:71. 10.1136/bmj.n71 [DOI] [PMC free article] [PubMed] [Google Scholar]

Supplementary Materials