Treatment options for Achilles tendinopathy: a scoping review of preclinical studies

Literature search strategy

A systematic search was performed in PubMed, Embase.com, Clarivate Analytics/Web of Science Core Collection, and the Wiley/Cochrane Library. The timeframe within the databases was from inception to the 4th of May 2023. The search included the following keywords and free text terms: (synonyms of) ‘Achilles tendinopathy’ combined with (synonyms of) ‘in vivo’ or ‘in vitro’. A full overview of the search terms per database can be found in Appendix B.

Inclusion and exclusion criteria

The review focusses in particular on Achilles tendinopathy as result of overuse. Pre-clinical models focusing on overuse AT will be included in the studies. Other Achilles tendon injuries such as ruptures will be excluded. Models based on Achilles tendon ruptures or surgically created ruptures to establish a tendinopathy model will be excluded because they are iatrogenic injuries and do not represent overuse injuries.

Furthermore, the following inclusion criteria were applied:

• Studies evaluating treatment options specifically for Achilles tendinopathy (tendinitis or tendinosis) in ‘in vivo’ AT animal models or in ‘in vitro’ AT tendon cells;

• The Achilles tendinopathy was chemically or mechanically induced;

• Reporting of biomechanical properties that include the description of tensile strength, elastic modulus and their relevance to the tendon function;

• Reporting of biomechanical properties in which key biochemical markers such as collagen content and inflammatory mediators are mentioned;

• Reporting of histological properties in which a detailed view of microscopic changes such as cell morphology and tendon fibre organizations is assessed.

The following exclusion criteria were applied:

• Clinical human studies;

• Systematic, scoping, narrative or other reviews;

• Commentaries;

• Guidelines;

• Treatment options for Achilles tendon rupture in both human and animal studies;

• Achilles tendon injuries induced by tenotomies or blunt trauma;

• Treatment options for Achilles tendinopathy caused by systemic conditions such as diabetes;

• If the article is published in a language other than English.

Article selection

After completion of the initial search, the articles were uploaded to Rayyan.ai for the title and abstract screening (Mourad Ouzzani, Fedorowicz & Elmagarmid, 2016). Duplicates were removed with the automated duplicate screening of Rayyan and verified by NAP to confirm that a duplicate was rightfully deleted. First, titles were screened for words that fit the exclusion criteria. If the title was not clear enough, the abstract was examined. The remaining articles were full-text reviewed for meeting the inclusion criteria. Endnote was used for full-text screening as the manuscript was written in Microsoft Word with an Endnote extension for citing.

Risk of bias quality assessment

The Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) risk-of-bias tool was used to evaluate the quality of animal studies (Hooijmans et al., 2014). The risk of bias was assessed on the grounds of ten points which evaluate selection bias, performance bias, detection bias, attrition bias, reporting bias, as well a category of other sources of bias that are not covered by the SYRCLE domains. Only if a specific domain is clearly stated in the article, it was classified as low risk of bias. When a domain is stated imprecisely it was classified as unclear risk of bias. When a certain domain is not mentioned or specified it is classified as a high risk of bias. Specific attention was paid to potential conflicts of interest and included as the other source of bias (McGuinness & Higgins, 2020). The tables with the individual risk of bias assessment of the included animal studies are displayed in Appendix C.

Data extraction and synthesis of results

The following data were extracted: Author and year of publication, study design, number of animals or cells, Achilles tendinopathy induction method, treatment conducted, positive and negative biomechanical, histological, and/or biochemical outcomes of the Achilles tendon before and after treatment intervention. Microsoft Word was used to create tables for data extraction. These data were labelled in the first row and the studies in the first column. During the analysis of the text, the data was extracted. After data extraction the article was analysed again to ensure data were not missed. When data was missing or an outcome measure was not reported, it was labelled as ‘Not reported’ (NR), and it was assumed that the authors did not evaluate a certain outcome measure.

The treatment types extracted were categorized as non-invasive, minimally invasive, invasive, or orally administered. Non-invasive treatments encompassed therapies not involving surgery or injections, but other than orally administered treatments. Examples are topical applications or laser therapies. Minimally invasive therapies are defined as treatments requiring application through injections. Oral therapies were administered orally such as diclofenac or green tea, while invasive therapies necessitated an incision or surgical procedure for administration.

The SYRCLE risk-of-bias assessment is visualized with the use of the risk-of-bias visualization tool. A high risk of bias is pictured with a red circle, uncertain risk of bias with a yellow circle and a low risk of bias with a green circle.

Study inclusion

The literature search yielded 4,790 results after removing duplicate articles. After title and abstract screening, a total of 335 articles were included for full-text screening. The full-text screening resulted in the inclusion of 98 articles. The selection process and exclusion reasons during the abstract and full-text review are displayed in the PRISMA flow-chart Fig. 1.

Description of the included studies

Detailed tables with the characteristics of the studies are presented in Appendix D, Appendix E and Appendix F. Of the included studies, 80 were in vivo animal studies, 16 studies involved both in vivo animal and in vitro tendon specimens, and two in vitro studies. In general, the studies compared the efficacy of different treatments with each other and/or with a sham group that received saline (NaCl). After saline injection, the sham and control groups of the studies showed slightly disorganized tendon fibres which turned to normal at the time of analysis. The time of analysis and follow-up time varied between the studies. Short and long term outcomes were reported whereas the shortest analysis was done after two hours and the longest follow-up time was 24 weeks. On average 48 animals were included per study with a range of 6 to 493 animals. Mostly rats were used for the analysis. After that rabbits, mice, sheep, and horses were used. The cells tested in the in vitro models were tenocytes derived from humans (Lee et al., 2021), mice (Liu et al., 2020), rabbits (Ruan et al., 2021) sheep (Al-Shudiefat et al., 2022), or rat (Lee et al., 2021; Choi et al., 2020a; Choi et al., 2020b; Wang et al., 2020; Zhao et al., 2019; Jeong et al., 2018; Kim et al., 2018; Chen et al., 2014; Vieira et al., 2018; Chen et al., 2012) Achilles tendon cells.

A total of 65 different treatment interventions were evaluated (Fig. 2). The treatment types were either non-invasive, minimally invasive (intratendinous injections), invasive, or orally administered. The treatments investigated most frequently are platelet-rich plasma (PRP) (n = 14), low-level laser therapy (LLLT) (n = 10), and the administration of Non-Steroidal Anti-Inflammatory Drug (NSAID) (n = 8).

Methodological quality of included studies

A total of 96 in vivo animal studies were analysed with the SYRCLE risk of bias tool (Hooijmans et al., 2014). The pooled quality of all the included studies is summarized in Fig. 3. All studies had a moderate to high risk of selection bias based on the risk-of-bias criteria. None of the studies reported the exact method of the applied randomization. There was an uncertain to high risk of performance bias. A total of 28 studies did not report how the animals were housed. The other 37 mentioned that the animals were housed but did not specify how. Only nine studies specified that the researchers giving the intervention were blinded. The risk was moderate in general regarding the detection bias. A total of 52 studies reported that random samples were used for outcome assessment and 43 studies reported that researchers analysing the outcomes were blinded. In sum, these results indicate a considerable risk of bias in the majority of the articles. The individual quality assessment of the studies is presented in Appendix C.

Induction of tendinopathy model

The in vivo animal model studies describe seven methods of induction of Achilles tendinopathy. Injection with collagenase type I directly in the Achilles tendon has been done by 82 of the 97 in vivo animal Achilles tendinopathy models. Other studies induced AT with prostaglandin E, TGF-β1, carrageenan, PGE₁, betamethasone, or H₂O₂. Two studies induced the tendinopathy mechanically through intensive treadmill running for 8 or 24 weeks (Zhang et al., 2020; Ng & Chung, 2012).

Six different methods were used to establish an in vitro tendinopathy model. Five by adding TNF-α, IL-1β, H₂O₂, HMGB1, or bacterial lipopolysaccharides. One model used a mechanical method, by cyclic stretching of the tenocytes (Chen et al., 2012). Thus, while several methods exist to induce AT, induction with collagenase type I is the most common in the in vivo model. These induction methods led to disrupted collagen fibres, neovascularization, and infiltration of inflammatory cells.

Reported outcomes

Detailed tables of outcomes of the included studies are presented in Appendix D, Appendix E and Appendix F. An improvement of Achilles tendon properties after treatment is reported by 80 of 98 articles. Improvement of AT such as better organization of collagen fibres, a lower amount of inflammatory cells and an improved maximum load of the tendon were mentioned. Worsening of AT after treatment was reported by seven studies. These studies evaluated the following treatments: Percutaneous augmented soft tissue mobilization (ASTM) (Imai et al., 2015), extracorporeal shockwave therapy (Çınar et al., 2013), diclofenac (Marcos et al., 2011; Marsolais, Côté & Frenette, 2003), passive stretching combined with laser therapy (Ng & Chung, 2012), early exercise after injury (Godbout, Ang & Frenette, 2006) and heparin (Tatari et al., 2001). Five studies reported that the evaluated treatment did not have effect on AT. Six studies reported mixed results (Fig. 4). Mixed results imply that an intervention positively affected a certain outcome measure, but did not affect or negatively affected another outcome measure.

Furthermore, 33 of the included studies reported the effect of their intervention on all three outcome modalities (biomechanical, histological, and biochemical). The outcomes of these 33 studies are presented in Table 1. Eighteen of these studies reported positive outcomes with their interventions in all three outcome measures and are marked with an “*” in Fig. 2 (Choi et al., 2020a; Choi et al., 2020b; Jeong et al., 2018; Kim et al., 2018; Kim et al., 2022; Gundogdu et al., 2019; Ge et al., 2023; Ma et al., 2023; Chen et al., 2011; Jiang et al., 2022; Vieira et al., 2015b; Lee et al., 2019; Ko et al., 2022; Gundogdu et al., 2023; Wu et al., 2022; Oshita et al., 2016; Lacitignola et al., 2014; Jiang et al., 2020). Other studies evaluated only one or two of these outcomes (Fig. 5). In summary, only 7% reported negative treatment effects, 12% reported mixed results, while 81% of the studies reported positive results.

Biomechanical outcomes

The biomechanical outcomes of the treatment interventions that reported significant changes are summarized in Table 2. Generally, biomechanical properties were evaluated using rupture force, maximum load, stiffness, tensile stress, and Young’s modulus. Overall, after injecting collagenase type I lower stiffness and maximum loads were reported. Several studies reported increased rupture force after intervention with 3J LED therapy (Marcos et al., 2012), PRP (Chen et al., 2014), Triamcinolone combined with PRP (Ruan et al., 2021) glycine diet (Vieira et al., 2015a), green tea administration (Vieira et al., 2015b), recombinant human platelet-derived growth factor-BB (Solchaga et al., 2014), and treadmill exercise (Bell et al., 2013) compared to control groups. Three studies reported increased tensile modulus after intervention with treadmill exercise (Bell et al., 2013), HUCPVCs (Emrani & Davies, 2011), and rhPDGF-BB (Solchaga et al., 2014) compared to control and sham groups. Interestingly, early exercise after AT seems to worsen the biomechanical properties while late exercise seems to improve the biomechanical properties (Godbout, Ang & Frenette, 2006). The biomechanical characteristics of the Achilles tendon are significantly improved by the majority of treatments such as PRP and Low level laser therapy. However, the results also show that several treatment interventions have no effects or in fact worsen the mechanics of the Achilles tendon with therapies such as ibuprofen administration.

Table 2:

Biomechanical effects on Achilles tendon tissue.

Biomechanical Outcomes	Rupture force	Maximum load	Stiffness	Tensile stress	Young’s modulus
Docosahexaenoic (Gundogdu et al., 2021)		+	+	+
AAMG (Palumbo Piccionello et al., 2021)	+
PRP (Chen et al., 2014; Jiang et al., 2022; Solchaga et al., 2014; Li et al., 2020; Yan et al., 2017; Calandruccio et al., 2015)		+	+	+
rHuPH20 (Rezvani et al., 2020)		↓	↓		↓
Ibuprofen (Bittermann et al., 2018)	=	=	↓	=	↓
hADSC+ rhPDGF-BB (Chen et al., 2018)		+	+	+
Glycine (Vieira et al., 2015a)		+
Green tea (Vieira et al., 2015b)		+
ASTM (Gehlsen, Ganion & Helfst, 1999; Davidson et al., 1997)					↓
rhPDGF-BB (Solchaga et al., 2014)		+	+	+	+
hMSC (Ma et al., 2023; Machova Urdzikova et al., 2014)	=		+
LLLT (Marcos et al., 2012; Naterstad et al., 2018; Marcos et al., 2014)		+/ ↓	+/ ↓
Uphill treadmill (Gundogdu et al., 2023; Bell et al., 2013)	+	+	+	+	+
Extracorporeal shockwave (Chen et al., 2004; Yoo et al., 2012)		↓	+/ ↓
Autologous tenocytes (Chen et al., 2011)		+	+
HUCPVCs (Emrani & Davies, 2011)				+	+
Early exercise (Godbout, Ang & Frenette, 2006)	↓			↓
LF/Hep-PLGA NPs (Choi et al., 2020a)			+	+
DEX/PMSs PLGA (Choi et al., 2020b)			+	+
Aspirin (Wang et al., 2020)				+	+
Diclofenac (Marcos et al., 2011; Naterstad et al., 2018)	↓	↓
Cur/PMSs PLGA (Kim et al., 2018)				+
SIM/PMSs PLGA (Jeong et al., 2018)			+	+
Irreversible Electroporation (Wang et al., 2023)		+	↓	+
US + Ximenia Americana L (Leal et al., 2016)	+

DOI: 10.7717/peerj.18143/table-2

Notes:

+

Improvement of biomechanical properties

↓

Deterioration of biomechanical properties

+/ ↓

Both improvement and deterioration reported

=

No difference reported Blank boxes implies that the biomechanical outcome is not reported for the treatment

Preclinical models of Achilles tendinopathy

Currently, there is no animal model that accurately mimics AT in humans (Perucca Orfei et al., 2016; Warden, 2007). Considering the disadvantages and advantages of the different animal models, the sheep and rabbit models appear to be the better overall option. Rabbit’s models may be more comparable to humans as their cellular and tissue physiology approximates that of humans. Sheep models have weight bearing of the tendon similar to humans (Zhang et al., 2022). The higher representation of rats (66% of the included studies) could be explained by their easy availability and low cost. The disadvantage is the small size of their tendons which makes intratendinous injections more challenging. Due to their size histological and biomechanical analyses are also more complex (Lui et al., 2011). Additionally, rat tendons have a much higher surface/volume ratio, which may overestimate the effect of treatment. It is important to highlight that the majority of evaluated animal models are quadruped models which differs greatly from bipedal models. The lack of a low-cost valid animal model for AT hinders representative high-quality pre-clinical research, which can eventually be translated into clinical practice.

The variations of the induction methods in animal studies contributes to the heterogeneity of the included studies. The predominant chemical induction (86% of studies) involves intratendinous injection of collagenase type I which disrupts collagen bundles that mimic human AT. Currently, a comprehensive protocol for establishing AT with collagenase is lacking, leading to inconsistency of applies collagenase doses. This is characterized by the diverse doses applied (1 mg/mL–10 mg/mL) among rats in the included studies of this scoping review. Perucca Orfei et al. (2016) reports that higher doses (3 mg/mL) demonstrate a closer resemblance to human disease, displaying features like fatty deposits and morphological changes similar to human AT at day 15. Additionally, a disadvantage of induction with collagenase type I is that tendon damage is immediately apparent after induction which may not resemble the pathophysiology of overuse AT in humans (Warden, 2007). Mechanical overloading, a potentially more valid induction method, is less frequently applied (two studies) possibly due to time and resource limitations (Zhao et al., 2019; Ng & Chung, 2012). To enhance the validity and comparability of preclinical AT models, there is a need for a validated animal model and standardized induction methods.

In vitro studies of AT are mostly done at an early phase to evaluate the toxicity, cell differentiation, pathophysiology, and different AT pathways. Interestingly, only one study was found that used human tenocytes to establish an in vitro AT model by administering TNF-β (Lee et al., 2021). However, the article did not state whether these tenocytes were derived from patients with AT or healthy Achilles tendons. It would be interesting to investigate if the use of human tenocytes gives more representative outcomes than animal tenocytes. Furthermore, the specific type of human tenocytes utilized in studies is crucial, as differences exist between energy-storing tendons and positional tendons (Birch, 2007). The Achilles tendon is an example of an energy storing tendon which is required to stretch and recoil to provide an adequate return of energy. While positional tendons such as the tibitalis anterior tendons are composed of more stiff matrix. The differences between energy-storing tendons such as the Achilles tendon and positional tendons such as the tibialis anterior may be interesting to study as they have different functions. However, it is essential to clearly distinguish these aspects in order to accurately assess the results of such studies.

This scoping review excluded one study that used an ex vivo model of freshly harvested bovine superficial digital flexor tendons to evaluate the use of genipin injections for collagenase D-induced AT (Tondelli et al., 2020). The use of ex vivo models is interesting as animals have multifunctional use and potentially do not have to be bred specifically for AT research. As ethical and moral subjects about the necessity and efficiency of the use of animal models are ongoing, the use of validated ex vivo AT models could be an answer to this issue.

Evaluated therapies in animal models

Interestingly, there is heterogeneity in the most assessed treatments (PRP, LLLT, and NSAID) within themselves. For example, the method of preparing PRP and its composition differs per study. Jiang et al. (2020) and Li et al. (2020) compared the use of leukocyte-rich PRP and leukocyte-poor PRP with each other. Both studies found that the use of leukocyte-rich PRP achieved better results than leukocyte-poor PRP at the early stage of AT. However, others found that late-stage leukocyte-poor PRP is more beneficial for AT (Yan et al., 2017). Furthermore, photo-bio modulation therapies are delivered in different wavelength intensities and frequencies in the included studies. Additionally, pulsed-based therapies are delivered in different wave- and radio frequencies. There are also studies reporting conflicting results with intratendinous injection with heparin. Intratendinous injections with heparin were reported to result in more pronounced collagen fibers, less cellularity, and less neovascularization (Williams et al., 1986). However, others found that heparin had a degenerative effect on the Achilles tendon (Tatari et al., 2001). All in all, even though the same treatment types are used, outcomes may vary because the specifics of the intervention differ.

Interestingly, treatment methods for AT which are contraindicated in clinical practice because of their adverse events such as glucocorticoids and NSAID, were used in innovative administration tools. Choi et al. (2020b) used porous microspheres to administer dexamethasone, which resulted in healing and anti-inflammatory effects and resulted in no sign of degeneration. Lee et al. (2019) implanted a diclofenac-immobilized polycaprolactone fibrous sheet through surgery. This resulted in higher collagen content, anti-inflammatory effects, and improved mechanical strength. These innovative options may create opportunities for methods that were not applicable because of their initially adverse effects, to be reconsidered and expand the treatment options.

Translation and comparability of preclinical studies

In 2018 and 2021 the AT management guidelines of The Orthopaedic Section of the American Physical Therapy Association (APTA) and Dutch multidisciplinary guidelines on Achilles tendinopathy were published (Martin et al., 2018; de Vos et al., 2021). There is not much overlap when comparing these guidelines with the preclinical treatment options included in this scoping review. Both guidelines advocate the use of exercise therapies to manage AT. However, only about 4% of the included treatments in this review evaluated exercise therapies for AT. That said, the lack of exercise based animal studies could be attributed to the challenges of implementing exercise regimens in animal models. These numbers indicate that the translation of preclinical to clinical concepts lacks synchronization. More consideration should be drawn to improving the translatability of these studies. Embracing open science initiatives where clear guidelines for pre-clinical research are developed may contribute to the translatability of this research field (Diederich et al., 2022).

Preclinical treatments that show positive effects on the biomechanical, histological, and biochemical outcome measures warrant consideration for clinical trial investigation. Mesenchymal stem cells (Ma et al., 2023; Ahrberg et al., 2018) and substances such as dexamethasone (Choi et al., 2020b) and simvastatin (Jeong et al., 2018) administered via porous microspheres, though not frequently explored in human trials, present interesting possibilities for evaluation. Previously hindered by adverse effects, innovative approaches, such as dexamethasone and diclofenac, may now offer possible options for clinical studies. Similarly, interventions involving green tea leaves in combination with glycine (Vieira et al., 2015b), farrerol (Wu et al., 2022), and friedelin (Jiang et al., 2022), stand out due to their accessibility. However, despite positive outcomes in preclinical models, treatments like PRP after many human trials have yet to demonstrate satisfactory clinical efficacy (de Vos et al., 2021). Furthermore, surgical treatments were underrepresented in the included studies. As the literature shows that patients who fail conservative treatment may have an indication for surgery more preclinical studies investigating the surgical treatments may be warranted (Maffulli, Sharma & Luscombe, 2004).

The clinical symptom severity associated with Achilles tendinopathy is not necessarily correlated with their current tendon structure or the extent of tendon damage (Warden, 2007). This poses a challenge for translating pre-clinical studies to clinical practice. Pre-clinical studies predominantly focus on the effects of treatments on tendon characteristics, biochemical effects and histological changes. In clinical practice, factors such as patient expectations, education, concomitant chronic diseases, and coping strategies may play a role in pain severity and quality of life.

Limitations of the scoping review

Using the ROB tool of SYRCLE, no study was identified that specifically reported how the randomization of the animals were conducted. In systematic reviews that analyse clinical randomized control trials, the specification of the randomization process is an important factor for the strength of the review article (Sterne et al., 2019). However, this is not standard practice in animal studies (Hooijmans et al., 2014). Some studies did not report crucial data regarding the age or gender of the included animals, which hampers the methodological quality of these studies. Although there is no certainty that the researchers did not consider these points, the absence of reporting these important characteristics diminishes the overall validity of these studies and thus the findings of this review.

Furthermore, a limitation of this review is that a high number of studies were excluded because the treatment was not evaluated on a chemically or mechanically induced AT model but rather on a model where the tendon was either ruptured or surgically damaged. As a consequence, some interesting and innovative treatments and interventions were not included.

Additionally, the literature section, data extraction, and risk of bias evaluation were performed by a single reviewer. This raises the risk of missing relevant articles and misinterpretations during the risk of bias assessment (Waffenschmidt et al., 2019). Due to substantial heterogeneity among studies regarding animal and breed differences, initiation methods of the Achilles tendinopathy model, therapeutic application and type, treatment frequency and dosage, follow-up time, and reported outcome measures, no clear conclusion about the best treatment intervention can be reported.

Lastly, a limitation of this scoping review is the absence of preregistration. As mentioned in the method sections, at the time of conducting the review, the authors were not aware of the requirement for preregistration of scoping reviews. Registration of scoping reviews is now recognized as a best practice to enhance transparency and minimize bias. Despite this, we followed the guidelines for scoping reviews, ensuring a systematic and thorough approach of the available literature. Future reviews by our team will integrate preregistration to follow these best practices and further strengthen the transparency of our future studies.

Implications for future research

This scoping review provides an overview of treatments in vivo and in vitro studies and may provide thoughts for future preclinical research. Additionally, it enables comparing innovative pre-clinical treatment options with established clinical practices considering AT management. Despite the promising preclinical treatment outcomes, the translation to clinical practice falls behind when comparing the preclinical treatments with common clinical AT guidelines. All the included studies used quadruped animal models for analysis. A justification is needed how quadruped models can be translated to bipedal animal models. Future studies should focus more on standardizing protocols to establish a valid in vitro and in vivo animal model and thus strive for less heterogeneity among the studies and in our opinion endeavour to have higher reporting standards to minimize the risk of bias. Less heterogeneity could be achieved by aiming for a universal agreement concerning the histopathological, biochemical, and biomechanical changes. The risk of bias could be minimized using a tool such as the SYRCLE ROB tool to design the animal studies. This may lead to more uniformity of baseline characteristics and a lower risk of bias allowing for comparisons across studies, enhancing their quality, and better understanding of the reported outcomes.