Introduction

Lateral epicondylitis is a common source of lateral elbow pain. Population studies have shown a prevalence of 1.3 % among those between 30 and 64 years of age, peaking between 45 and 54. It typically affects the dominant upper extremity and is associated with repetitive and forceful activity [53]. Pain is often most pronounced with wrist extension.

Lateral epicondylitis is believed to be a degenerative process, which stems from repetitive microtrauma. Typically, samples from the affected tissue demonstrate angiofibroblastic hyperplasia at the extensor origin of the forearm [60]. Activities requiring repeated contraction of the wrist extensors are implicated, with the extensor carpi radialis brevis (ECRB) tendon most commonly involved. Studies comparing cadaveric and surgical specimens indicate that lateral epicondylitis evolves through several stages, beginning with degenerative angiogenesis and ending with fibrosis and calcification [38, 49, 60].

The majority of patients diagnosed with lateral epicondylitis can be effectively managed without surgery, as it is usually a self-limited process from which up to 90 % of patients will recover by 1 year without surgical intervention [5, 12, 60]. Non-surgical approaches to treatment are numerous. The aim of this review is to examine randomized controlled trials (RCTs) of the non-surgical treatment of lateral epicondylitis.

Materials and Methods

Study Identification

Articles from Medline/Ovid, Cochrane Database of Systematic Reviews, and CINAHL were systematically searched from earliest records to February 2013 (Fig. 1). Search terms included “lateral epicondylitis,” “lateral elbow pain,” “tennis elbow,” “lateral epicondylalgia,” and “elbow tendinopathy” combined with “randomized controlled trial.” The abstracts of the resulting articles were reviewed by two of the authors for the following eligibility criteria: (1) a target population of patients with symptoms of lateral epicondylitis; (2) evaluation of treatment of lateral epicondylitis with one of the following non-operative techniques: corticosteroid injection, injection technique, iontophoresis, botulinum toxin A injection, prolotherapy, platelet-rich plasma (PRP) or autologous blood (ABI) injection, bracing, physical therapy (PT), shockwave therapy, or laser therapy; and (3) a RCT design. Articles meeting the above criteria based on the abstract were considered for inclusion, and full text was retrieved. The quality of each article was assessed according to randomization, blinding, outcome measures, and proportion of patients lost to follow-up. Only articles representing level I or II evidence for a particular intervention were included. This represents double-blind RCTs or single-blind studies for interventions in which blinding was either impossible or had little potential effect on outcome measures. Loss to follow-up was limited to 20 %, with few exceptions, which are noted in the text. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed wherever applicable. A summary of all articles included in this review can be found in Table 1.

Fig. 1
figure 1

Flow diagram representing process of study identification. RCT randomized controlled trial, LE lateral epicondylitis

Full size image
Table 1 Characteristics of included studies
Full size table

Results

Corticosteroid Injection

Injection of corticosteroid into the area over the origin of the ECRB or point of maximal tenderness has been a common treatment for relief of lateral elbow pain. Dexamethasone, betamethasone, and triamcinolone have been variously used mixed with a local anesthetic such as lidocaine or bupivacaine.

A study by Hay et al. [23] examined the management of lateral epicondylitis with corticosteroid injection, naproxen, or placebo tablets. At 4 weeks, injection showed clear advantage with 92 % of patients in this group reporting they were completely pain free or improved. Only 57 % of the naproxen group and 50 % of the placebo group reported the same. At 12 months, however, no differences were found between the groups.

Corticosteroid injection, physical therapy, and a wait-and-see approach were compared by Verhaar et al. [66] using a scoring system based on subjective pain relief, patient satisfaction, grip strength, and provoked pain to rate outcomes for treatment with corticosteroid injection or physical therapy. At 6 weeks, patients receiving corticosteroid injection had significantly (p < 0.05) more excellent and good results than the physical therapy group. There was no difference when the groups were compared at 1 year; however, at that point, 50 % of the patients had had other therapy or surgery. Another short-term study comparing corticosteroid injection, physical therapy, both injection and therapy, and no treatment was performed by Tonks et al. [62]. This study had similar results at 7 weeks, with the injection only group performing significantly (p < 0.045) better than other treatment groups. However, 23 % of participants were lost to follow-up in this study.

Smidt et al. [54] assessed long-term outcomes in a RCT comparing corticosteroid injection, physical therapy, and wait-and-see. Corticosteroid injections were found to be the most effective treatment at 6 weeks but were surpassed by physical therapy and no treatment by 52 weeks. Results from later studies by Bisset et al. [6, 7] found comparable results using the same treatment groups. Physical therapy and corticosteroid injection were more effective than wait-and-see before 6 weeks. At 52 weeks, however, there was no difference between physical therapy and wait-and-see, and subjects receiving an injection fared worse than those in other treatment groups. A recent study by Coombes et al. [11] found similar short-term improvement with corticosteroid injections over placebo at 4 weeks; however, these patients fared worse at 26 and 52 weeks of follow-up.

Two double-blind RCTs compared the effects of corticosteroid combined with local anesthetic injection to an injection of local anesthetic alone. Newcomer et al. [37] randomized patients to receive betamethasone/bupivacaine or bupivacaine alone. All patients also received physical therapy. The only statistically significant (p < 0.04) finding was a better visual analog scale (VAS) pain score in the corticosteroid group between 2 and 6 months. Lindenhovius et al. [31] compared injection of dexamethasone and lidocaine with lidocaine alone. No difference was found in any outcome measure at 1- or 6-month follow-ups.

Injection Technique

The peppering technique was described almost 50 years ago [48] and involves many small injections delivered by inserting a needle, injecting, withdrawing without exiting the skin, repositioning, and injecting again until the sensation of crepitus felt with these injections disappears. This technique is thought to initiate healing of the tendon by creating new channels through degenerative tissue in which bleeding occurs [3].

Altay et al. [3] compared injection of corticosteroid plus local anesthetic to local anesthetic alone using the above technique. A second injection of the same type was given at a 2-week follow-up in patients whose symptoms persisted. Results were categorized as excellent, good, fair, or poor according to modified Verhaar criteria (pain relief, patient satisfaction, grip strength, presence of provoked pain on resisted wrist extension). No difference between the two groups was found up to 1 year post-injection, concluding that injection technique is more important than the substance injected.

Dogramaci et al. [16] randomized patients to receive either a single injection of corticosteroid plus local anesthetic, a peppered injection of corticosteroid plus local anesthetic, or a peppered injection of local anesthetic alone. Patients receiving peppered injection of corticosteroid plus local anesthetic had significant (p < 0.011) improvement in VAS and a higher percentage of excellent results according to the Verhaar criteria than the other groups at 6 months. Okcu et al. [41] found similar results when they compared single and peppered injections of corticosteroid, assessing the outcome with Disabilities of the Arm, Shoulder, and Hand (DASH) scores (Turkish). Statistically significant (p < 0.017) results in favor of the peppered injection group were found at 1 year, but results of this study did not include the 38 % of patients with inadequate follow-up.

Iontophoresis

Iontophoresis involves the use of a small electric current to drive charged molecules through the skin, allowing topical medications to penetrate deeper tissues. Demirtas and Oner [15] studied the short-term effects of iontophoresis of two well-known non-steroidal anti-inflammatory drugs (NSAIDs), sodium diclofenac and sodium salicylate. Significant (p < 0.001) decreases in pain were seen with both drugs after one treatment a day for up to 18 days. Sodium diclofenac produced a larger reduction in pain than sodium salicylate. A RCT of oral diclofenac and placebo was conducted by Labelle and Guibert [28]. Both groups were immobilized and the treatment group received 150 mg of diclofenac daily for 28 days. No difference was found between the groups with regard to function; however, there was a significant (p < 0.03) difference in pain reduction in the treatment group.

In Nirschl et al. [40], patients received six treatments of dexamethasone or placebo iontophoresis within 15 days. At 2 days post-treatment, those treated with dexamethasone reported significantly (p < 0.012) less pain on VAS than patients receiving placebo. There was no significant (p = 0.249) difference found between the groups at 1 month.

Stefanou et al. [57] tested the delivery of dexamethasone via a self-contained battery powered iontophoresis patch against dexamethasone injection and triamcinolone injection. All subjects also received therapy. Outcomes were assessed at the end of therapy and at 6 months. Outcomes were similar for all groups at 6 months, but only the iontophoresis group had significantly (p < 0.05) improved grip strength and higher rates of returning to work without restriction at the end of therapy.

Botulinum Toxin A Injection

Botulinum toxin A has been proposed as a non-surgical option for treatment of lateral epicondylitis. The mechanism for relief of pain is paralysis of extensor muscles, preventing further microtrauma to the tendon origins and allowing healing to occur.

Hayton et al. [24], Wong et al. [69], Placzek et al. [47], and Espandar et al. [18] compared 50–60 U of botulinum toxin A or saline injected 1–5 cm distal to the point of maximum tenderness or one third the distance along the forearm from the tip of the lateral epicondyle. Primary outcomes were pain score and grip strength at intervals up to 18 weeks. Hayton’s group did not find any difference between the treatment and placebo groups at any time. However, studies by Wong, Placzek, and Espandar found significant (p < 0.006, 0.003, and 0.01, respectively) differences in pain score but not in grip strength. All studies documented patients in the treatment groups who experienced transient weakness in finger extension.

Lin et al. [30] compared botulinum toxin A to corticosteroid injection and assessed VAS pain score, quality of life, and grip strength at 4, 8, and 12 weeks. The botulinum treatment was significantly (p < 0.02) less effective in reducing pain and resulted in significantly (p < 0.01) reduced grip strength when compared to the corticosteroid group at 4 weeks. The reduced grip strength continued at 8 weeks. There was no difference in quality of life at any time.

Prolotherapy

Prolotherapy consists of an injection composed of osmotics/irritants (dextrose or polidocanol) and/or chemotactics (sodium morrhuate), which promote inflammation in the target tissue. The iatrogenic local inflammation is proposed to induce fibroblastic growth and collagen synthesis, ultimately leading to stronger repair of damaged fibers at the lateral epicondyle.

A double-blind RCT was performed by Scarpone et al. [52] to assess the efficacy of prolotherapy vs. placebo injection in the treatment of refractory lateral epicondylitis. The study group received three injections of dextrose, sodium morrhuate, lidocaine, and sensorcaine at 4-week intervals, while the control group received saline injections. Significantly (p < 0.05) improved pain and grip strength were observed in the prolotherapy group at intervals up to 52 weeks.

Carayannopoulos et al. [8] compared prolotherapy and corticosteroid injections for the treatment of lateral epicondylitis in a later RCT. The prolotherapy group received two injections consisting of dextrose, glycerin, phenol, sodium morrhuate, and procaine, whereas the corticosteroid group received an injection of methylprednisolone and procaine. Both groups demonstrated significant (p < 0.04) improvement at 3 or 6 months, but results of this study did not include the 29 % of patients with inadequate follow-up. No difference could be found between the treatments. Zeisig et al. [70] studied ultrasound guided injections of polidocanol vs lidocaine/epinephrine in a double-blind RCT and reported improvement in grip strength and VAS in both groups at 3 months with no difference between groups.

Platelet-Rich Plasma and Autologous Blood Injection

Injections of PRP or ABI have been proposed to facilitate healing through release of growth factors directly at the site of interest. Specifically, PRP contains high levels of various growth factors including platelet-derived growth factor, transforming growth factor beta, and vascular endothelial growth factor [35]. PRP is created by withdrawing the patient’s own blood and centrifuging it to isolate a platelet-rich fraction before injecting into the patient at the site of interest. In the case of ABI, autologous blood is withdrawn from the patient but is not treated before reinjection.

Several studies have compared the injection of PRP or ABI to injection of corticosteroids. Gosens et al. and Peerbooms et al. [20, 44] performed double-blind RCTs demonstrating a significant (p < 0.005) difference in favor of the groups receiving PRP injections over those receiving corticosteroid injections with follow-up intervals up to 2 years. Similarly, Kazemi et al. [25] compared ABI to corticosteroid injection in a RCT where assessors were unaware of treatment. At 8 weeks, ABI was found to be more effective in all outcomes. A recent study by Krogh et al. [27] did not find PRP superior to corticosteroid injection at 1 month or to either corticosteroid injection of placebo at 3 months of follow-up; however, tendon thickness was found to be reduced on ultrasound examination in corticosteroid injection over PRP or placebo test subjects.

Wolf et al. [68] attempted to account for any placebo effect associated with injection by adding a control group to their RCT study design. Patients were randomized to receive one of three treatments including ABI, corticosteroid/lidocaine injection, and saline/lidocaine injection. Outcomes improved over time in all three groups with no differences at 2- and 6-month follow-ups.

PRP has also been compared directly to ABI. Creaney et al. [14] conducted a double-blind RCT with patients receiving injections at 1 month intervals of either autologous blood or PRP. Both preparations produced improvement over 6 months, but there were no differences between groups. Thanasas et al. [61] also compared the two in a RCT in which assessors were unaware of treatment allocation. Ultrasound guidance and peppering injection technique was used in both groups. Statistically significant (p < 0.05) improvement in pain for PRP over ABI was present only at 6 weeks.

Bracing

Two commonly used braces are the proximal forearm strap (counter force brace) and wrist extension splint. By holding the wrist in extension, the splint is thought to unload the extensor origin and relax the wrist extensors. There are several theories as to how the forearm strap reduces the force exerted on the common extensor origin, one being that its compressive force limits expansion and thereby the force generated by the extensors. Another is that the strap serves as a secondary origin for the ECRB muscle, reducing the force conducted proximally [39]. Both mechanisms would theoretically decrease stress on the damaged tendons, allowing healing.

Several authors have compared the effects of the proximal forearm strap and wrist extension splint on lateral epicondylitis symptoms. One study found no difference between the two orthotics [64], and two found a substantial difference in favor of the wrist splint in reduction of pain [2, 19]. Both studies were limited to 6 weeks of follow-up and lacked a control group; one study [19] did not include 40 % of patients with inadequate follow-up, making interpretation of the true benefit of either splint difficult to determine in these studies.

Struijs et al. [59] randomly assigned 180 patients to forearm strap, physical therapy, or combination and found no differences at 26- or 52-week follow-ups. At 6 weeks, the strap-only group was significantly (p < 0.05) better in ability to perform daily activities. Luginbuhl et al. [32] compared use of a forearm strap for at least 3 months, strengthening exercises, and a combination with the addition of a corticosteroid injection at the beginning of treatment. All groups experienced significant (p < 0.0001) improvement in pain at 6 weeks and 1 year, and there was no difference between groups at any time.

Physical Therapy

PT is often employed for non-surgical treatment of lateral epicondylitis. There are many studies comparing the efficacy of PT to other modalities of treatment that are included elsewhere in this review. There are also several RCTs which have compared PT to wait-and-see or examined different methods of PT. Peterson et al. [45] and Park et al. [43] compared PT to wait-and-see, with mixed findings. Peterson’s group found significant (p < 0.0016) improvement in pain for the treatment group while Park saw significantly (p < 0.01) better pain scores for the control group. Viswas et al. [67] evaluated a stretching/strengthening program against Cyriax friction massage, finding no difference in outcomes between the methods. Tyler et al. [63] and Martinez-Silvestrini et al. [34] examined various combinations of eccentric and concentric exercises compared to standard PT with inconsistent results; Tyler’s study found a significant difference (p < 0.011) in all measures for the eccentric group while Martinez-Silvestrini found no difference in outcomes between any group. The recent study examining corticosteroid injection and physical therapy by Coombes et al. [11] previously discussed found short-term improvement with physical therapy alone; however, this effect disappeared by 26 and 52 weeks of follow-up.

Shockwave Therapy

Extracorporeal shock wave therapy (ESWT) applies energy to the interface of two substances with differing acoustic impedance. It has been proposed that this energy promotes tissue healing and possibly reduces pain by stimulating nerve fibers to produce analgesia [56].

Of the several double-blind RCTs done to evaluate the efficacy of ESWT, only two authors have produced significant (p < 0.001) findings. Rompe et al. conducted two studies [50, 51] looking at the efficacy of ESWT in the treatment of lateral epicondylitis. Both studies were double-blind RCTs looking at patients with refractory lateral epicondylitis of at least 12 months. Patients received three weekly treatments of either active or sham ESWT. In the earlier study, significant (p < 0.001) improvement in pain and grip strength was observed in the active treatment group at 3, 6, and 24 weeks. In the later study, significant (p < 0.001) difference in favor of the active treatment group was found up to 12 months after treatment. Pettrone and McCall [46] also saw a positive effect of ESWT on pain and function 12 weeks post-treatment in patients with refractory lateral epicondylitis for more than 6 months. Other investigators, Haake et al. [21], Speed et al. [55], Melikyan et al. [36], Chung et al. [10], and Staples et al. [56] did not find any substantial difference between treatment and placebo groups with any outcome measure.

Laser Therapy

Low-level laser therapy (LLLT) is proposed to have a biostimulatory effect on tissue, reducing levels of TNF alpha [1] and reducing cell apoptosis [9] although the precise method by which this occurs is incompletely understood. The effect of LLLT on lateral epicondylitis has been studied with conflicting results. Early double-blind RCTs did not show an advantage to LLLT. Lundeberg et al. [33] studied two different types of laser (pulsed Ga–As and continuous He–Ne), with no difference between treatment and placebo groups up to 3 months after treatment. Four more RCTs studied the effect of either a Ga–As or Ga–Al–As laser vs. sham laser therapy. Varying levels of energy were delivered per point in each study, and follow-up periods ranged from 7 weeks to 1 year. Three studies did not demonstrate a difference in results between laser therapy and placebo [22, 26, 42] (although one of these studies [26] did not include 25 % of subjects lost to follow-up) whereas a fourth study did [65].

Results of more recent studies conflict with those found previously. Basford et al. [4] conducted a double-blind RCT with a Nd-YAG laser and placebo which did not demonstrate a difference in outcome at 4 weeks. However, a study by Stergioulas [58] combined plyometric exercise with Ga–As laser or placebo laser and found a significant (p < 0.05) improvement in VAS and strength at 8 and 16 weeks in the active treatment group. A similar study by Lam and Cheing [29], looking only at short-term outcomes of Ga–As laser treatment at 3 weeks found comparable results (p < 0.0125). Emanet et al. [17] also found positive results of LLLT in their double-blind RCT using a Ga–As laser and additional physical therapy for both groups with a statistically significant (p < 0.05) difference with respect to improved pain, grip strength, and functional assessment in favor of the treatment group at 12 weeks.

Discussion

Many non-surgical treatments for lateral epicondylitis have been developed and studied. Corticosteroid injection, injection technique, iontophoresis, botulinum toxin A, prolotherapy, platelet-rich plasma/autologous blood injections, bracing, physical therapy, extracorporeal shock wave therapy, and laser therapy have been subjected to RCTs. When taken in aggregate, we propose there are several conclusions to be drawn from the collection of studies reviewed herein.

Given that lateral epicondylitis is thought to result from repetitive microtrauma, rather than an inflammatory process, it is not surprising that corticosteroid injections do not provide long-term relief. Patients receiving corticosteroid injections have improved pain relief over physical therapy or a wait-and-see approach for 1 to 2 months post-treatment, although these effects do not remain when examined at 1 year. The several studies testing the effect of a peppering technique of corticosteroid injection show the technique improves results at 6 months to 1 year of follow-up. RCTs of iontophoresis have shown some benefit to this technique of corticosteroid delivery, at least on a very short-term basis of days to weeks.

Other types of injection have been studied in addition to injection of corticosteroids. Treatment with botulinum toxin A may also provide pain relief when compared to placebo, but not to corticosteroid injection, and the accompanying transient extensor weakness may not be acceptable to patients. Prolotherapy is superior to saline placebo but does not appear to be more effective than corticosteroid or lidocaine injection. PRP and ABI have been found to be both less and more effective than corticosteroid injections, and PRP has been shown to also be either more or equally effective than ABI in studies.

Bracing with a proximal forearm strap or wrist extension splint, however, seems to provide little to no additional benefit when combined with physical therapy. The studies reviewing physical therapy without bracing or other concurrent treatment did not show a clear advantage to PT.

Non-invasive techniques such as laser therapy and ESWT have been the subject of RCTs. Early studies of laser therapy did not show an effect of treatment whereas more recent investigations did show substantial improvement for patients treated with laser therapy over those who received placebo therapy. These beneficial effects were most apparent at 1 to 4 months of follow-up. The preponderance of studies of extracorporeal shockwave therapy appears to indicate that this treatment is no more effective than placebo.

A weakness of this systematic review is that which is inherent in many studies, even those which are undertaken as RCT, blinded trials: Design of the perfect trial is probably impossible, as there are always unanticipated variables [13]. Patients are lost to follow-up, have subjective responses to test criteria, or there may be unintended investigator bias. However, given a large number of investigations, valuable conclusions may be drawn.

In summary, lateral epicondylitis is a condition that is usually self-limited, resolving over a 12- to 18-month period without treatment. There are many options for non-surgical treatment for pain relief. Corticosteroids, whether delivered with a standard or peppered injection technique or via iontophoresis, may provide short-term pain relief on the order of several months; however, this effect is not sustained in the long term. Injections of botulinum toxin A, prolotherapy, PRP, or ABI may also provide varying degrees of improvement in pain, although botulinum toxin A can have unacceptable side effects. Studies of LLLT do not show a clear benefit at this time. Therapies that do not include injection, such as bracing, PT, or ESWT, do not appear to provide substantial benefit in terms of pain relief or improved function. In conclusion, there are multiple randomized controlled trials for non-surgical management of lateral epicondylitis, but the existing literature does not provide conclusive evidence that there is one preferred method of non-surgical treatment for this condition.