The use of platelet-rich plasma therapy in treating tennis elbow: A critical review of randomised control trials

Tennis elbow (TE) is a painful and debilitating condition of the elbow. Recently, the use of orthobiologics, such as platelet-rich-plasma (PRP), has been proposed to promote tendon regeneration. Despite their popularity, there is a paucity of updated reviews on the use of PRP compared with other treatment modalities for treating TE. The aim of this review is to summarise high quality studies that compare the use of PRP therapy with other therapies for TE and to identify areas where further research is warranted. This systematic review was performed in accordance to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. A comprehensive literature search of PubMed, Scopus and Cochrane Library databases was undertaken in May 2021. Articles were screened for the following criteria: randomised control trials (RCTs) involving PRP in at least one of the treatment arms for tennis elbow. The quality of the RCTs included were analysed for their risk of bias using the modified Cochrane Collaboration Risk of Bias Assessment Tool for randomised trials. A total of 20 RCTs of which 1520 TE patients were analysed. The RCTs included in this review compared PRP with various treatment modalities routinely used in clinical practice such as physiotherapy, steroid injections, Autologous Whole Blood (AWB) and surgical interventions. With regards to the quality of RCTs, collectively, selection bias was found to be low risk however, performance bias in terms of blinding of participants and personnel performed poorly. Of the 20 RCTs, only 5 studies were classified as low risk of bias. In these 5 studies, 2 RCTs compared PRP with steroids and reported contrasting results, 1 RCT compared PRP with AWB injections which reported both to be similarly efficacious, 3 RCTs included a placebo group and only 1 reported superior effects with PRP. There are 2 main types of PRP classified according to the number of pro-inflammatory leukocyte i.e. leukocyte-rich and leukocyte-poor PRP. However, only 8 studies documented the formulation of PRP used. While the heterogeneity of PRP formulations could in-part explain the reported differences in outcomes, overall there is limited robust evidence to recommend PRP therapy for TE. Further research is required to establish the optimal formulation and administration of PRP injections. Proper documentation of TE patients need to be standardised before concrete recommendations on the use of PRP therapy may be offered.


Introduction
Tennis elbow (TE), or lateral epicondylitis, is a painful and debilitating condition of the elbow caused by angiofibroblastic hyperplasia of the tendinous origin of wrist extensors especially the extensor carpi radialis brevis (ECRB) muscle. 1 TE is a common elbow pathology affecting 4e7 people per 1000 annually and is particularly common amongst racquet sports athletes due to overuse or repetitive stress. 2,3 TE typically presents as pain over the lateral epicondyle extending distally over the mobile wad that may be exacerbated by forearm supination and radial deviation. TE was previously thought to be a self-limiting condition, 4 however chronic cases of TE have been reported to persist for years and associated to complications such as muscle wasting and weak grip strength. 5 In the first instance, the current national treatment guidelines in England proposes a 6-week long plan of analgesia and rest with a reduction or cessation in the aggravating activities. During this period, the application of an orthosis such as a wrist or elbow brace may also be considered. After 6 weeks, other forms of conservative management such as physiotherapy with eccentric loading, strengthening exercises and massage may be considered. 6,7 Finally, if the patient shows no improvement after 6e12 months of treatment, referral to an orthopaedic clinic should then be considered. 9 Surgical debridement of tendinosis including the release or repair of the damaged extensor tendon may then be considered as a last resort for these patients with refractory pain or severe functional impairment. Percutaneous needle tenotomy and injections of corticosteroid into the tendon sheath, platelet-rich plasma (PRP) or autologous whole blood (AWB) are also alternative therapeutic options before resorting to surgical decompression, debridement with or without tendon repair. 3,6e9 Lately, in vitro studies have proposed that PRP can stimulate tendon regeneration as high concentrations of growth factors and cytokines have been found at the site of tendon healing, suggesting a potential mechanism of action. 10,11 Additionally, multiple studies demonstrated PRP to be able to positively influence angiogenic factors as well as tendon cell proliferation. 12,13 Despite their popularity, there is a paucity of recent and updated reviews on the use of PRP comparing with other available treatment modalities for treating TE. Given that surgical intervention for the treatment of TE is usually only offered after a trial of conservative and medical injection therapy, most TE patients would have undergone a form of injection therapy. Given their importance, it is crucial to scrutinise these injection therapies.
This study aims to summarise the findings of robust studies with high levels of evidence which compare PRP therapy with other available treatment modalities for TE. Based on the current basic science evidence, we hypothesise that PRP therapy is therapeutically efficacious in managing pain and functionality in TE patients.

Methods
This systematic review was performed in accordance to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. In May 2021, the Cochrane Library, PubMed and Scopus databases were thoroughly searched with the following search string: ("elbow tendon injury" OR "elbow tendinopathy" OR "elbow tendonitis" OR "tennis elbow" OR "lateral epicondylitis" OR "lateral epicondylosis" OR "lateral epicondylalgia" OR "epicondylitis") AND ("platelet-rich plasma" OR "autologous whole blood"). Each keyword is connected via a Boolean operator "AND" or "OR".
Articles were screened for the following criteria: randomised control trials (RCTs) involving PRP in at least one of the treatment arms for TE as well as a treatment group without PRP therapy. Articles not in English language, incomplete data, meeting summaries, case reports, retrospective studies, randomised and nonrandomised prospective cohort studies, or review articles with no original data were excluded.
Details of each RCT were retrieved individually. Information on the different treatment groups, number of patients included, duration of symptoms, length of follow up and blinding, key outcomes for each RCT are documented in Table 1. The details of the treatment recorded include the type of PRP used and how the PRP was prepared where available. All quantitative and qualitative outcomes, including pain and performance scores, of the RCTs were recorded where available. The primary outcomes compared was improvements in pain and function. Secondary outcomes included any complications or side effects of the treatments. Attempts to contact the corresponding authors of studies which required clarification of certain details were made. The outcomes of the RCTs were largely evaluated qualitatively due to inadequate quantitative data for the collective pooling of the data quantitatively. Additionally, a meta-analysis was not attempted as multiple studies were at risk of bias.
The quality of the RCTs included were analysed for their study design. The risk of bias for each RCT was evaluated using the modified Cochrane Collaboration Risk of Bias Assessment Tool for randomised trials (adapted from Higgins and Altman). 14 2 reviewers individually assigned each RCT to have a classification of either "high risk", "low risk", or "unclear risk" in various aspects of each study as summarised in Fig. 1. Further discussion with the senior reviewers were done to resolve any discordance between the 2 initial reviewers. A RCT was classified with an overall low bias risk if the criteria for "low risk" was met for the following aspects of the study: participant blinding (D3), selection bias (D1 and D2), attrition bias (D6), and reporting bias (D7). 15

Results
The PubMed, Scopus and Cochrane database searches yielded 299 manuscripts collectively. These manuscripts underwent a twostage selection process by two independent reviewers (Fig. 2). In the first step, all 299 abstracts were screened against the criterions for inclusion and exclusion. Abstracts without original data, not utilising PRP as therapy for TE, not in English language and duplicates were excluded. Secondly, the manuscripts were retrieved in full and subsequently analysed for eligibility. Studies not randomised in a control trial or failed to record comparison outcomes of PRP and the control group were excluded. Bibliographies of relevant published papers were also reviewed with the same two-stage process (Fig. 2) to detect other relevant studies not captured by the primary search.

Overview of studies
In this review, a total of 20 RCTs were included with a total of 1520 TE patients analysed and followed up for a minimum of 6 weeks (range 6e104 weeks). 16e35 The average number of patients recruited for each RCT was 76 (range 24e230 patients) with the largest RCT conducted being a multicentre study including 230 patients by Mishra et al. 16 The RCTs included in this review compared PRP with various treatment strategies for TE ranging from placebo saline injections to conservative physiotherapy to surgical release. In the patients who underwent PRP injection to the ECRB tendon or tendon sheath, 8 studies (40%) administered PRP under ultrasound guidance while the remaining 12 studies (60%) performed manual palpation. The most common comparison group within these RCTs was corticosteroid injections. Only 3 RCTs compared PRP with multiple treatments modalities within the same study.
With regards to preparation, PRP can be formulated to be either leukocyte enriched or deficient. Of the 20 RCTs, only 8 studies documented the formulation of the PRP therapy used. 5 RCTs utilised leukocyte-rich PRP whereas the other 3 RCTs used leukocytepoor PRP. Not one study compared the different formulations of PRP as separate treatment groups.
The included studies recruited TE patients with varying duration of symptoms. 14 RCTs exclusively involved TE patients who were chronically symptomatic for longer than 3 months. 4 RCTs did not document the duration of symptoms of their patients.
All RCTs recorded the length of follow up post treatment. 16 RCTs followed up before 4 weeks and had multiple subsequent    follow ups. 2 RCTs followed up significantly later at 90 days and 6 months respectively. 2 RCTs followed up just once after 6 weeks and 2 years. In terms of blinding, only 1 RCT completely blinded all 3 parties involved with the study i.e. patient, assessor and radiologist. 9 RCTs arranged for single blinding while 6 RCTs performed double blinding. 4 RCTs failed to document any form of blinding.

Bias risk assessment of studies
All 20 RCTs included in this review underwent a thorough risk of bias assessment of which a summary of the analysis is reflected in Fig. 1. Across the RCTs assessed, 5 studies (25%) were classified as low risk of bias, 5 studies (25%) were unclear, whilst the remaining 10 studies (50%) had a high risk of bias.

PRP vs steroid therapy
A total of 8 RCTs compared the use of PRP with steroids as therapy for TE. 17e24 4 RCTs found PRP to be more effective than corticosteroid injection of which 1 RCT had a low risk of bias. 17,19,20,23 Only 1 of the 4 RCTs reporting similar efficacy between PRP and steroids had a low risk of bias. 18

PRP vs autologous whole blood (AWB) therapy
A total of 4 RCTs have been performed to compare PRP with AWB therapy. 25e28 The only low risk of bias RCT reported PRP to be similarly efficacious as AWB at treating TE. 25 The only RCT suggesting both PRP and AWB had no therapeutic effect was at high risk of bias. 28

PRP vs surgical management
In summary, only 2 RCTs were found to have compared PRP with surgical interventions. Patients in these studies found surgical management to be superior in terms of pain relief compared to PRP, however both studies had a high risk of bias. 29,30

PRP vs physiotherapy
Only 1 RCT was found to have compared PRP with physiotherapy which showed superior pain and functional results in the PRP group however, this study had a high bias risk. 31

PRP vs anaesthesia
A total of 3 RCTs were found to have compared PRP with anaesthesia however all 3 studies had either a high or unclear risk of bias. 22,32,33 The only RCT reporting PRP to be superior in terms of pain relief and functional scores had an unclear risk of bias. 32

PRP vs placebo
In summary, 5 RCTs were found to have compared PRP with various types of placebo including percutaneous needle tenotomy and saline injections. 4 RCTs compared PRP with saline injections however all 4 studies reported no difference in pain or functional scores. 18,28,34,35 The only RCT that compared PRP with dry needling as an active control group reported superior effects with PRP. 16 This RCT had a low risk of bias.

Discussion
Overall, the primary finding of the current study demonstrate that PRP when compared to a variety of other treatment modalities, including placebo, physiotherapy, injections containing corticosteroid, PRP and AWB as well surgical interventions fail to demonstrate a significant improvement in pain or functional outcomes in the management of TE.
The most common treatment group that PRP was compared with was corticosteroid injections. Gosens et al. reported superior relief of pain and restoration of function more effectively with PRP than steroid injections in patients with chronic TE. 17 This study had a low bias risk. Yadav et al. reported both PRP and steroids to be effective at pain relief and improving functional outcomes but the PRP group enjoyed significantly better improvements at the 3 months follow up period. 23 This study has an unclear risk of bias. Conversely, Krogh et al. reported that while PRP and steroid therapy were similar to saline in terms of alleviating pain at 3 months follow up, steroid therapy provided superior pain relief in comparison to the other treatment groups at 1 month follow up. 18 This study had a low bias risk. Gupta et al. reported greater short term pain management and performance scores at 6 weeks with steroid injections, but patients receiving PRP injections enjoyed superior results in the longer term. 19 However, 2 studies, Varshney et al. and Omar et al. found no difference between the PRP and steroid groups during the follow up period. 20,24 Both of these studies had a high risk of bias. Gautam et al. reported that both PRP and steroids helped with significant improvement in pain and functional scores. 21 However, PRP enabled biological healing for patients with chronic symptomatic TE, whereas patients treated with corticosteroids are at increased risk of tendon degeneration. There was an unclear level of bias risk with this study. Palacio et al. reported no statistical difference between PRP, steroid and anaesthetic therapy. 22 This study had an unclear risk of bias.
In the past, steroid injections were popular and even considered to be the gold standard treatment for TE. However, recent studies have recommended against steroid therapy due to the short-lived efficacy as well as potential detrimental effects including high recurrence rates, dermal depigmentation and subcutaneous atrophy. 36e38 Physiologically, steroids have been found to inhibit tenocyte proliferation and progenitor cell recruitment leading to reduced collagen synthesis and greater fatty tissue changes. 39 It is therefore prudent to interpret studies that compare the efficacy of PRP and steroid therapy with caution. Instead, further studies should consider using anaesthetic injections as a more appropriate comparison treatment group to mitigate the steroids interference with healing mechanisms in TE patients.
Both AWB and PRP injections contain platelets which possess strong growth factors and granules vital to the healing process of chronic injuries. Theoretically, PRP contains a higher concentration of platelets than in AWB hence a superior effect in the repair process of tendinopathies. 26 However, our findings show PRP to be similarly efficacious to AWB in treating TE. Creaney et al. reported a prospective single blinded RCT and found that both PRP and AWB provided comparable efficacy with a significant reduction in pain scores at 6 months. 25 This study had a low bias risk. Raeissadat et al. performed a single blinded RCT and found that both PRP and AWB injection therapies were similarly efficacious with comparable pain and functional scores at every follow up interval. 26 This study had a high bias risk. Thanasas et al. reported that PRP therapy was superior at pain reduction in TE patients compared to AWB in the short term at 6 weeks. 27 However, in the longer term at 3 and 6 months, both AWB and PRP were statistically similar. This study had a high risk of bias. However, most recently in 2020, Linnanm€ aki et al. documented a RCT comparing PRP, AWB as well as placebo with saline as therapy for TE patients. 28 Both PRP and AWB therapies did not help with pain or function in comparison to the saline group. This study had a high bias risk.
Surgical release or debridement of the damaged extensor tendon is typically considered after a failed trial of conservative treatment. Previous studies 40 have documented high success rates of up to 97.7% however, only 2 RCTs were found to have directly compared surgical intervention with PRP. Our findings show that patients found surgery to be superior in terms of pain relief compared to PRP, the RCTs included were at high risk of bias especially because it was technically impossible to blind both patients and researchers. Merolla et al. compared the effects of PRP with arthroscopic debridement. 29 They found that both were effective at improving both pain and performance initially. Conversely, surgical debridement was superior in the long-term. Watts et al. compared PRP with surgical release for refractory TE. 30 They reported no statistical difference functionally however, surgical patients reported superior pain scores compared to the PRP group. Both studies had a high risk of bias.
Physiotherapy for the treatment of tennis elbow typically includes eccentric strengthening exercises for the wrist extensors and static stretching of the ECRB. 41 While only up to 10% of TE patients eventually opt for surgical management, 42 majority if not all patients would be offered a course of physiotherapy. However, only 1 RCTs compared PRP with physiotherapy and reported PRP led to superior pain and functional scores in TE patients. 31 These improvements were reported to be sustained longer than 6 months of follow up, without complications in any patient.
Needle tenotomy can be performed as a standalone procedure ("dry needling") or part of a combined intervention such as with local anaesthetic agents. Theoretically, tenotomy in itself may induce healing via microtrauma. 32 However, multiple studies have performed needle tenotomy as a control or placebo group to compare against PRP. Palacio et al. compared PRP, steroid and anaesthetic therapy but reported no evidence of any statistical difference between these treatment modalities. 22 This study had an unclear risk of bias. Martin et al. compared PRP therapy with lidocaine as tenotomy adjuvants in recalcitrant TE patients. 32 They reported no evidence of statistical differences in terms of pain and functional scores. This study had a high risk of bias. Behera et al. compared PRP therapy with bupivacaine. 33 They reported superior results with PRP injections in terms of pain relief and functional scores. This study had an unclear risk of bias. Linnanm€ aki et al. documented that PRP and other treatment modalities were similar to saline in terms of alleviating pain and function in TE patients. 28 This study had a high bias risk. 3 studies examined the effects of PRP as compared to saline injections and reported no statistical difference in both pain and functional outcomes. 18,34,35 2 studies 18,34 had a low risk of bias while 1 had a high risk of bias. 35 Mishra et al. compared the effects of PRP with simple tendon needling alone. 16 They reported a greater efficacy with PRP compared to dry needling alone as treatment for TE. This study had a low bias risk.
Analysing the methodology and quality of the 20 RCTs, only 5 studies (25%) met the criteria to be classified having a low bias risk. The bias risk in the remaining RCTs severely limits the interpretability of the results obtained. Notably, there was high risk of bias with regards to the blinding of participants, personnel including clinicians and researchers, outcome of assessment in almost half of the RCTs. Whilst this can be partially attributed to the nature of treatment groups i.e. surgical intervention compared to PRP injections, multiple studies were performed with ambiguous blinding methodology. For instance, multiple studies recorded single, 26 or triple blinding, 22  Consequently, inconsistent blinding amongst these RCTs led to the classification of high risk for bias. The vast spectrum of treatment options available demonstrates that TE encompasses a range of heterogeneous conditions which affect the common extensor tendons with varying severity. Physiologically, tendons possess the capability to stretch in response to increasing forces. However, a microtear will ensue if this force exceeds the tendon's elastic threshold. Repetitive stress and overuse may result in multiple microtears leading to degenerative changes also known as tendinosis. Eventually, a full-thickness tendon tear may ensue with gradual degradation of the tendon. In general, there are 4 distinct grades of tendinopathy as succinctly summarised by Bhabra et al. 6 Using basic science principles, they proposed a treatment algorithm to help clinicians manage TE appropriately depending on severity of tendinopathy. 6 Briefly, grade 1 tendinopathy only requires rest and activity modification to avoid further damage. Grade 2 tendinopathy is classified according to the presence of immature vascular hyperplasia hence indicating PRP or AWB injections. In grade 3 tendinopathy, the loss of cells due to apoptosis and autophagy indicates replacing with autologous cell therapy. Finally, in grade 4 tendinopathy, surgical repair may be required to mechanically restore the collagen matrix. Consequently, the efficacy of PRP in TE will depend on the grade of tendinopathy as well as healing stage. Given this broad spectrum of severity in TE patients, RCTs should aim to identify and specify the grade of tendinopathy in their patient cohort to avoid bias in the interpretation of results. However, multiple studies failed to document the chronicity of symptoms or if any prior intervention was trialled in their TE patients let alone the grade of tendinopathy. This heterogeneity of patients in the RCTs makes their findings additionally difficult to generalise.
The results of this review must be considered in light of some limitations. Firstly, the heterogeneity of patients and the severity of their condition was not accounted for. Furthermore, only 7 studies documented the preparation and formulation of the PRP therapy used. The 2 main types of PRP include leukocyte-rich and leukocyte-poor PRP of which the former consists of a high number of leukocytes and is pro-inflammatory, whilst the latter is devoid of neutrophils making it anti-inflammatory in nature. 43 Given that proteases secreted by leukocytes can affect growth factors secreted by platelets, superior results might be achieved using pure PRP instead. 44 This was supported in a recent review which also suggested leukocytes in PRP may affect the efficacy provided by PRP. 45 It is therefore imperative that the formulation of PRP therapy used is documented so that further research will be able to accurately compare such studies. It is however unfortunate that not one RCT, in this review, was found to compare the different formulations of PRP in separate treatment groups. Additionally, it should also be noted that only 7 studies documented the brand and type of machine that was used to prepare PRP therapy. This apparent lack of homogeneity of PRP preparations could in-part justify the reported differences in outcomes amongst the RCTs in this review.
Despite the heterogeneity of RCTs included in this review, several high level evidence studies including systematic reviews and meta-analysis have been previously published to analyse the efficacy of PRP in TE patients. In 2017 and 2018, multiple studies found that while steroid therapy may be associated with greater initial benefit, PRP therapy have superior long term outcomes. 46,47 Interestingly, a recent systematic review published in 2021 concluded that PRP therapy provided similar improvements to surgical interventions in terms of functional scores and pain relief for TE patients in the initial phase. 48 These abovementioned studies demonstrate a high level of evidence suggesting therapeutic efficacy of PRP therapy in TE patients. However, the literature is riddled with an abundance of mixed and contrasting evidence as well. A recent meta-analysis reported insufficient evidence to recommend the usage of PRP for TE. 49 Another meta-analysis found PRP to be only minimally effective when compared to other forms of minimally invasive therapies for TE such as steroids, AWB injections, local anaesthetics, dry needle and even saline injections. 50 The insufficient evidence of PRP therapy coupled with the relatively high costs of PRP and its accompanying preparational equipment makes it difficult to justify PRP therapy for TE currently.

Conclusion
Overall, our primary finding is that there is limited evidence to recommend the use of PRP therapy in terms of pain relief and function for TE. Further research with larger cohorts and longer duration of follow up are required to establish the optimal formulation and administration of PRP injections. Additionally, proper documentation of TE patients and the chronicity of their symptoms, pain and functional evaluation scores need to be standardised before concrete recommendations can be made regarding the therapeutic efficacy of PRP.

Ethical approval
No Institutional Review Board approval was required for this study.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declarations of competing interests
None.