Skip to main content
. 2014 May 8;9(4):419–446. doi: 10.1007/s11552-014-9642-x

Table 1.

Characteristics of included studies

Author, year Participants Experimental groups Blinding Outcome measures Follow‐up periods Lost to follow‐up Results Side effects (number of patients affected) Conflict of interest statement and/or source of funding
Corticosteroid injection
Hay, 1999 164 patients with new episode of LE (1) Injection of 20 mg methylprednisolone plus lignocaine Patient and doctors aware of treatment but assessor was blinded Primary: participant global assessment of improvement (5-point scale) 4 weeks, 6 months, 12 months; patients not restricted from receiving additional intervention 1/164 At 4 weeks, 92 % of injection group, 57 % naproxen group, and 50 % placebo group were completely better or improved; the difference between injection and other groups was statistically significant. Results at 12 months not statistically significant Injection: local skin atrophy at the lateral epicondyle (3); naproxen: GI side effects (4), allergic reaction, edema (1) Funded by Arthritis Research Campaign; methyprednisolone provided by UpJohn and naproxen by Syntex
(2) Naproxen 500 mg twice daily for 2 weeks Secondary: pain, function, and main complaint on Likert scale
(3) Placebo tablets
Verhaar, 1996 106 patients with LE not previously treated with Cyriax physiotherapy (1) Injection of 1 mL, 1 % triamcinolone with 1 mL, 1 % lidocaine +/− 2nd and 3rd injections at 2 and 4 weeks; injections given at extensor origin Patient blinding not possible, no mention of assessor blinding Primary: Verhaar criteria 6 weeks, 52 weeks; patients not restricted from receiving additional interventions after 6 weeks 3/106 patients not assessed at 6 weeks At 6 weeks, 69 % success rate in injection group vs 27 % of physiotherapy group, this difference was statistically significant. No difference between groups at 52 weeks None No benefits received from a commercial party related directly or indirectly to subject of this article
(2) Cyriax physiotherapy, 12 treatments over 4 weeks Secondary: grip strength
Tonks, 2007 48 patients with LE not treated in prior 6 months (1) Single injection 10 mg triamcinolone plus 2 % lignocaine Patient blinding not possible, no mention of assessor blinding Primary: pain free grip strength 7 weeks 11/48; 41 % drop-out rate from physiotherapy groups Patients receiving injection statistically significantly better for all outcome measures, no significant effect of physiotherapy for any outcome measure, no significant interaction found between physiotherapy and injection Injection: skin depigmentation and atrophy and injection site (1) None reported
(2) Physiotherapy Secondary: extensor strength, PRFEQ, complications of treatment
(3) Injection plus physiotherapy
(4) No treatment; injections given at point of maximum tenderness
Smidt, 2002 185 patients with LE for at least 6 weeks, no physiotherapy or injections in prior 6 months (1) Injection of 10 mg triamcinolone with 1 mL, 2 % lidocaine +/− 2nd and 3rd injection Patient blinding not possible, research physiotherapist doing assessment blinded Primary: general improvement 6‐point Likert scale; severity of main complaint, pain during the day, and inconvenience on 11‐point scale, modified pain‐free function questionnaire (PFFQ), researcher assessed overall severity 3, 6, 12, 26, and 52 weeks 2/178 At 6 weeks, injection group significantly better in all outcome measures (92 % success vs 47 % for physiotherapy and 32 % for wait‐and‐see). At 52 weeks, injection group significantly worse than other intervention groups but no significant difference between physiotherapy and wait-and‐see Injection: facial flush (2), skin irritation (3), red swollen elbow (2), skin color change over elbow (7) No conflict of interest; funded by Health Insurance Council and Netherlands Organization for Scientific Research
(2) 9 physiotherapy treatments over 6 weeks Secondary: PFG, MGS, pain-pressure threshold Physiotherapy: skin irritation (3), red swollen elbow (3), skin color change over elbow (3)
(3) Wait-and‐see
Bisset, 2006 198 patients with LE for at least 6 weeks, no treatment for 6 months (1) Injection of 10 mg triamcinolone and 1 mL, 1 % lidocaine +/− 2nd injection at 2 weeks Patient blinding not possible, assessor and study personnel blinded Primary: global improvement on 6‐point Likert scale (completely recovered or much improved considered success), pain-free grip force, assessor's rating of severity, recurrence rates beyond 6 weeks 3, 6, 12, 26, and 52 weeks 8/198 At 6 weeks, injection group had significantly better outcomes with 78 % success vs 27 % for wait‐and‐see and 65 % for physiotherapy. Injection group had significantly more recurrences than other treatment groups. At 52 weeks, physiotherapy had significantly better outcomes than injection but no significant difference in any primary outcome compared to wait-and‐see Injection: loss of skin pigment (2), atrophy of subcutaneous tissue (1) Funded by University of Queensland and the National Health and Medical Research Council, Primary Health Care Project Grant
(2) 8 physiotherapy treatments over 6 weeks Secondary: VAS, PFFQ
(3) Wait-and‐see
Bisset, 2007 Patient data extracted from 2 RCTs: Bisset (2006) and Smidt (2002) outlined above Treatment groups of respective studies, analysis of the following subgroups performed: baseline pain on VAS (<60, >60), duration of symptoms (<14 weeks, >14 weeks), previous history of symptoms, employment status (manual, non-manual, no work) See individual studies Global improvement on 6-point Likert scale Patient data from 6- to 52-week follow-up used See individual studies Subgroups of patients determined by baseline patient characteristics do have significantly different outcomes over time See individual studies See individual studies
Coombes, 2013 165 patients with unilateral LE for at least 6 weeks (1) CSI of 10 mg/mL of triamcinolone acetonide in a 1 mL injection plus 1 mL of 1 % lignocaine Assessors blinded, patients were blinded to injection type but not to PT 6-point Likert scale, VAS, PRTEE, EuroQuol EQ‐5D, use of non-allocated treatments, adverse events 4, 8, 12, 26, and 52 weeks 2/165 patients lost to follow-up 4 weeks: Treatment with CSI or PT alone showed improvement 26; 52 weeks: CSI with lower complete recovery or improvement, no difference in groups with or without physical therapy Pain after CSI/PT, transient depigmentation or subcutaneous atrophy, hand numbness, vomiting, swelling, skin irritation 2 authors received payment for lectures and a book chapter
(2) Placebo injection 0.5 mL of 0.9 % isotonic saline
(3) CSI and 8 weekly PT sessions
(4) Placebo and 8 weekly PT sessions
Newcomer, 2001 39 patients with LE for less than 4 weeks and no prior treatment (1) Injection of 6 mg betamethasone and 4 mL, 0.25 % bupivacaine Subjects and physician performing injections were blinded to medication used VAS, functional pain questionnaire, MGS 4 and 8 weeks, telephone follow-up at 6 months (excluding grip strength) 4/39 discontinued to pursue alternative treatment Both groups improved in a statistically significant manner over time, there were no significant differences in outcome between groups with the exception of an improvement in VAS in the injection group from 8 weeks to 6 months None reported None reported
(2) Injection of 5 mL, 0.25 % bupivacaine; patients not allowed to undergo any other prescribed treatment during 6-month study period
Lindenhovius, 2008 64 patients with LE for less than 6 months, no prior treatment with CSI or iontophoresis (1) Injection of 4 mg dexamethasone and 1 mL, 1 % lidocaine Patient and treating surgeon blinded to treatment DASH, VAS, MGS 1 and 6 months 6 patients dropped from study, 6 patients missed 1-month follow‐up, 9 patients missed 6-month follow-up; included in analysis until dropout Both groups improved over time, no statistically significant difference between groups in any outcome measure at 1 or 6 months Slight discoloration of skin around injection site (2 placebo, 1 dexamethasone) None reported
(2) Injection of 2 mL, 1 % lidocaine; patients free to pursue other forms of treatment during study
Altay, 2002 120 patients with LE, no prior CSI (1) Injection of 2 mL lidocaine Patient and assessor blinded to injection type Verhaar criteria 2, 6, and 12 months 0/120 Both groups improved over time, no statistically significant difference between groups at 2 months, results were not changed at 6 and 12 months No complications observed None reported
(2) Injection of 1 mL lidocaine and 1 mL triamcinolone; peppering technique used for both groups, option for 2nd injection at 2 weeks
Dogramaci, 2009 75 patients with LE, no prior CSI (1) Injection of 1 mL triamcinolone and 1 mL lidocaine Patient and assessor blinded to injection type VAS, Verhaar criteria (excellent or good outcomes considered successful) 3 weeks and 6 months Not discussed No statistically significant differences between groups at 3 weeks; at 6 months, successful outcomes statistically higher and VAS statistically lower in peppered CS injection group than other treatment groups No complications observed None reported
(2) Peppered injection of 1 mL lidocaine
(3) Peppered injection of 1 mL triamcinolone and 1 mL lidocaine; all groups +/− 2nd injection at 3 weeks
Okcu, 2012 80 patients with LE, no prior CSI (1) Injection of 1 mL betamethasone and 1 mL prilocaine Patient could not be blinded to peppering, assessor blinded to injection technique Turkish DASH 3, 6, 12, 18 months, and final follow-up (average 21.6 months) 31/80, not included in analysis A statistically significant difference in favor of the peppered injection group was found for DASH scores at 12 and 18 months and at final follow-up None reported Authors declare no conflict of interest
(2) Peppered injection of 1 mL betamethasone and 1 mL prilocaine; injections given at point of maximum tenderness
Iontophoresis
Demirtas, 1998 40 patients with LE (1) Dermal iontophoresis of sodium diclofenac followed by infrared treatment Not discussed Pain with pressure to LE, with resisted wrist extension, during functional activity, and at rest 18 days None Statistically significant improvement in both groups None None reported
(2) Dermal iontophoresis of sodium salicylate followed by infrared treatment
Labelle, 1997 129 patients with LE, no steroid intake in prior 6 weeks, no NSAID use in prior 3 weeks (1) Long-arm cast immobilization and daily slow release diclofenac 150 mg Patients and assessor blinded to treatment Primary: MPFGS 14, 21, and 28 days 1/129 At 28 days, all outcomes improved for both groups in a statistically significant manner. The difference between groups was only statistically significant for VAS (pain) and MGS ratio, in favor of the NSAID group Headache (4 placebo, 2 NSAID), abdominal pain (6 placebo, 19 NSAID), diarrhea (13 placebo, 25 NSAID) Supported by grant form National Health and Welfare Canada. Ciba‐Geigy Canada provided both NSAID and placebo
(2) Long-arm cast immobilization and daily placebo; cast was removed at 14 days, medication given daily for 28 days; no other medication for LE allowed during study period Secondary: MPFGS ratio of affected to normal side, MGS, MGS ratio, VAS (pain and function), pain-free function index
Nirschl, 2003 199 patients with medial or lateral epicondylitis for less than 3 months, no CSI in prior year (1) Dermal iontophoresis of dexamethasone Patients and investigator blinded to treatment Primary: investigator's global evaluation of improvement (−1, worsened, to 4, no symptoms), patient's global evaluation of improvement on same scale, VAS Secondary: patient's assessment of symptoms (0, none, to 4, very severe), investigator's tenderness evaluation on same scale 2 days (in person) and 1 month (by phone) post-treatment 9/199 Dexamethasone produced a statistically significant improvement in VAS (23 vs 14 mm for placebo) and more scores of moderate or better on the investigator's global improvement scale (52 vs 33 %) at 2 days post-treatment, but outcomes were not significantly different between groups at 1 month Dexamethasone: application site vesicles (2), blister (2), localized skin reaction (2), burning sensation (1), atopic dermatitis (1), erythema (1), hypersensitivity (1), pruritus (1), skin irritation (1) Supported by a grant from Iomed, Inc, Salt Lake City, Utah, who also provided solution
(2) Dermal iontophoresis of sodium chloride; 6 treatments within 15 days, patients were not to begin other treatments until after 2 day follow-up Placebo: pruritus (3), erythema (2), arthralgia (2), application site pain (1), application site pruritus (1), burning sensation (1), elbow pain (1)
Stefanou, 2012 101 patients with LE less than 2 years, no steroid medication within 6 months (1) Self-contained iontophoresis patch in place for 2 days with 10 mg dexamethasone in reservoir, +/− second patch replaced 2–10 days after removal of first Patients could not be blinded to treatment, physician who determined work status not blinded to treatment Primary: grip strength, PRTEE Completion of 8-week protocol, 6 months after completion of 8-week protocol 19/101 left the study due to non-compliance, treatment failure, or loss to follow-up and were excluded from analysis Iontophoresis group had statistically significant increase in grip strength at both follow-ups while the other 2 groups only achieved this by 6 months; all 3 groups had statistically significant improvement in PRTEE (pain and function) at 6-month follow-up; improvement in work status at the first follow-up was only statistically significantly for the iontophoresis group Injection: post-injection pain (<10 % patients) None reported
(2) Injection of 10 mg dexamethasone Secondary: work status (no restrictions, work with restrictions, no work Iontophoresis: skin irritation (1)
(3) Injection of 10 mg triamcinolone; treatment for all groups included an 8-week protocol of rest, mobility, and strengthening
Botulinum toxin
Hayton, 2005 40 patients with LE for more than 6 months; all patients received at least 1 CSI and full course of PT (1) Injection of 50 U of botulinum toxin type A in 2 mL normal saline Patient, physician administering injection, and assessor blinded to injection type MGS, VAS, QOL measured with SF-12 health status questionnaire 3 months 2 patients had an operation (1 from each experimental group), 1 patient lost to follow-up At 3 months, there were no significant differences between treatment groups for any outcome measure Transient extensor lag (12/18 patients receiving botulinum toxin injection) Authors did not receive benefits in support of their research or preparation of this manuscript. 1 or more authors received benefits from a commercial entity (botulinum toxin supplied by Allergan)
(2) Injection of 2 mL normal saline; injections administered 5 cm distal to area of maximal tenderness in line with the middle of the wrist
Wong, 2005 60 patients with LE for more than 3 months; no prior injection (1) Injection of 60 U of Dysport botulinum toxin Patient and physician blinded to injection type Primary: VAS 4 and 12 weeks No loss to follow-up At 4 weeks, the botulinum group showed a statistically significant improvement in VAS over the placebo group, which was maintained at 12 weeks; no significant difference in grip strength was found between groups at any time Botulinum: post-injection pain (2), weakness in finger extension (10), paresis of digits (4) Funded by a donation from New World Development Ltd. Sponsor had no part in design of study, data collection or interpretation, or final manuscript preparation
(2) Injection of an equivalent volume of normal saline; injections administered 1 cm from the lateral epicondyle and were aimed toward the tender spot Secondary: MGS Placebo: post-injection pain (1), post-injection nausea (1), weakness in finger extension (6)
Placzek, 2007 132 patients with LE for more than 4 months, 3 or more conservative therapies tried without success (1) Injection of 60 U botulinum toxin A in 0.6 mL normal saline Patient and assessor blinded to injection type Clinical pain score, VAS (continuous pain past 48 h, max pain past 48 h), MGS, global assessment of improvement by patient and physician (4, substantially better, to 0, substantially worse) 2, 6, 12, and 18 weeks 2/132 Improvement was more statistically significant for the botulinum group than the placebo group in the clinical pain score at every follow‐up and in VAS for continuous pain and global assessments of improvement from the 6th week onward Strength of extension of the third finger was decreased significantly in the botulinum group at 2 weeks, this weakness resolved by week 18 1 or more authors received outside funding or grants in excess of $10,000 from Ipsen Pharma in support of their research for or preparation of this work
(2) Injection of 0.6 mL normal saline; injections administered 3–4 cm distal to tender epicondyle
Espandar, 2010 48 patients with LE for at least 6 months and a previous course of PT and/or CSI (1) Injection of 60 U botulinum toxin A in 1 mL of normal saline Patient and physician blinded to injection type Primary: VAS 4, 8, and 16 weeks No loss to follow‐up Intensity of pain at rest and during max pinch decreased more in the botulinum group than the placebo group, this difference was statistically significant Botulinum: tingling sensation around injection site (5) subjective feeling of muscle spasm around injection site (8), pain at injection site (7) Study funded by Vice Chancellor of Research at Tehran University of Medical Sciences. No competing interest declared
(2) Injection of 1 mL of normal saline; injection given at 1/3 the length of the forearm from the tip of the lateral epicondyle on the course of the posterior interosseus nerve Secondary: pain during max grip and max pinch, grip strength Placebo: pain at injection site (2)
Lin, 2010 16 patients (19 elbows), no CSI in prior 3 months (1) Injection of 50 U botulinum toxin type A in 1 mL normal saline Patient, and assessor blinded to injection type VAS pain, pain‐free grip strength, WHO QOL‐BREF questionnaire to assess quality of life 4, 8, and 12 weeks 1 patient did not receive post‐treatment evaluations, excluded from analysis At 4 weeks, both groups reported decreased pain with a significantly greater reduction reported by the steroid group, this pain remained reduced at 8 and 12 weeks but difference between groups were not significantly significant. No statistically significant differences in QOL were observed between groups at any time All patients in botulinum toxin A group experienced mild weakness in wrist and middle digit extension that resolved by 12 weeks Study funded by Council of Labor Affairs of Taiwan government. Allergan Inc. provided Botox. Authors report no conflict of interest
(2) Injection of 1 mL, 40 mg triamcinolone; injections into the ECRB muscle near common extensor origin
Prolotherapy
Scarpone, 2008 24 patients with LE for at least 6 months and failure of conservative treatments, no CSI in prior 6 weeks (1) 0.5 mL prolotherapy injections (50 % dextrose, 5 % sodium morrhuate, 4 % lidocaine, 0.5 % sensorcaine) Patient and physician blinded to injection type Primary: resting elbow pain on 11‐point Likert scale 8 and 16 weeks, telephone interview at 52 weeks 4/24 patients dropped out of study, excluded from analysis The prolotherapy group, but not the control group, showed a statistically significant improvement in pain scores at 8 and 16 weeks. Isometric strength for the prolotherapy group improved significantly at 8 and 16 weeks, the difference between groups were statistically significant at 16 weeks Prolotherapy: local erythema, irritation, and discomfort 1 day after injection (2), all patients experienced self‐limited post‐injection pain Authors have no financial interest in the products mentioned in this article
(2) Injections of 0.5 mL normal saline; 3 injections administered into supracondylar ridge, lateral epicondyle, and annular ligament for total of 1.5 mL at baseline, 4, and 8 weeks Secondary: resting grip strength, isometric resistance strength
Carayannopoulos, 2001 24 patients with LE from 3 months to 2 years, no CSI in prior 6 months (1) Injection of 1 mL procaine, 0.9 mL of P2G (phenol 1.2 %, glycerine 12.5 %, dextrose 12.5 % in sterile water), and 0.1 mL sodium morrhuate Patient and physician blinded to injection type Primary: VAS 1 and 3 months, telephone follow‐up at 6 months 7/24, lost to follow‐up, excluded from analysis The prolotherapy group experienced a statistically significant improvement in VAS, QVAS, and DASH at 3 and 6 months, while the steroid group only saw this improvement in QVAS at 6 months and in DASH at 3 and 6 months. No significant differences between groups were observed at any time 1 patient from prolotherapy group withdrew from study because of excessive pain, no other adverse reaction reported by patients Peer reviewers and all others who control content have no relevant financial relationships to disclose
(2) Injection of 1 mL procaine and 1 mL methylprednisolone; above solutions split between 3 injection sites: annular ligament, origin of common extensor tendon, radial collateral ligament at tubercle of radius; repeat injections at 1 month Secondary: quadruple VAS (QVAS), DASH, grip strength, pain free grip strength
Zeisig, 2008 32 patients (36 elbows) with LE for more than 3 months, no interventions in prior 3 months (1) Injection of 0.5 mL polidocanol Patient and physician blinded to injection type Primary: patient satisfaction with treatment (yes/no), VAS (elbow pain during grip activities of daily life) 3 months; patients in group 2 offered injection of polidocanol after initial 3-month follow-up, all patients followed up 12 months after 1st injection 2/36 withdrew from trial before first follow‐up, not included in analysis There were no statistically significant differences between treatment groups for any outcome measures at any time. Patients in both groups had statistically significant decreases in VAS at both follow‐ups Patients in group 2 noted increased elbow pain and stiffness during first week after injection, no other adverse effects Authors declare no conflict of interest
(2) Injection of 0.5 mL lidocaine plus epinephrine; injections delivered to target vessels inside extensor origin under ultrasound guidance Secondary: MGS
Platelet-rich plasma/autologous blood injection
Peerbooms, 2010 100 patients with LE for more than 6 months, failed conservative therapy (1) PRP injection (PRP collected with Recover system by Biomet Biologics) Patient and assessor blinded to injection type VAS, DASH; successful treatment defined as more than 25 % reduction in VAS or DASH without a reintervention at 1 year 4, 8, 12, 26, and 52 weeks 8/100 lost to follow-up, included in analysis with last data point carried forward Success rates were 49 and 73 % according to VAS and 51 and 73 % according to DASH for the steroid group and PRP group respectively. These differences were statistically significant in favor of the PRP group Apart from local inflammation causing increased pain 3–4 weeks after PRP injection, no systemic or other local reactions were seen Sponsored by Biomet, the funding source had no involvement in study design; collection, analysis, and interpretation of data; writing of report; or decision to submit for publication
(2) Injection of triamcinolone 40 mg/mL with epinephrine; for both groups 1 mL was injected directly into area of maximum tenderness, remaining solution was injected into common extensor tendon using peppering technique; use of NSAIDs prohibited
Gosens, 2011 Continuation of Peerbooms (2010) According to Peerbooms (2010) Patient and assessor blinded to injection type VAS, DASH; successful treatment defined as more than 25 % reduction in VAS or DASH without a reintervention at 2 years 2 years 8/100 lost to follow-up, included in analysis with last data point carried forward Differences in success rates according to VAS and DASH were statistically significant in favor of the PRP group. Compared with baseline VAS and DASH scores, a number of patients deteriorated at 2 years, with more patients being from the steroid group. This difference was statistically significant for both VAS and DASH scores Apart from local inflammation causing increased pain 3–4 weeks after PRP injection, no systemic or other local reactions were seen Biomet supplied the Recover system used at a discounted rate but did not have any influence on the collection and analysis of data
Kazemi, 2010 60 patients with a new episode of LE, no CSI in prior 3 months (1) Injection of AB with 2 % lidocaine Patients not blinded because blood taken for ABI but assessors blinded to injection type Primary: VAS (within last 24 h) 4 and 8 weeks No loss to follow‐up Differences in outcome measures between groups was statistically significant in favor of ABI for VAS, pain in grip, pressure pain threshold, and Quick DASH (not significant for other outcomes) at 4 weeks and in all outcome measures at 8 weeks There were no noticeable or reported side effects of treatment in either group Authors report no conflict of interest
(2) Injection of 20 mg methylprednisolone with 2 % lidocaine; injections given proximal to lateral epicondyle; no other therapy allowed for duration of study Secondary: pain‐free function questionnaire, pain with max grip, quick DASH, modified Nirschl score, MGS, pressure pain threshold
Krogh, 2013 60 patients with LE for at least 3 months, tendinopathy on US (1) Injection of 3–3.5 mL PRP (prepared with Biomet GPS II); injections delivered deep to extensor origin with a peppering technique Assessors and patients blinded PRTEE 4, 12, 26, and 52 weeks 0/60 at 3 months, 34/60 at 6 months, 44/60 at 12 months CSI superior at 1 month, PRP or CSI not superior to placebo at 3 months with regard to pain reduction. CSI reduces tendon thickness on US more than PRP or placebo Rash, post-injection pain, skin atrophy, loss of pigmentation None reported
(2) CSI of 1 mL triamcinolon 40 mg/mL and 2 mL lidocaine 10 mg/mL
(3) Placebo injection 3 mL
Wolf, 2011 34 patients with LE, no CSI in prior 6 months (1) Injection of saline with lidocaine Patient but not physician giving injection blinded to injection type DASH, PRFEQ, VAS 2 weeks, 2 and 6 months 6/34 patients lost to follow‐up, not included in analysis There were no statistically significant differences between groups for any outcome measure at any time None reported Funded by Clinical Research Grant from the American Society for Surgery of the Hand. No benefits related to the article received
(2) Injection of CS with lidocaine
(3) Injection of AB with lidocaine; injections given with multiple passes of needle in fanlike fashion under the extensor origin
Creaney, 2011 150 patients with LE for at least 6 months failing conservative measures, no previous injection (1) Injection of PRP (isolated from AB spun in a centrifuge with siphoning of buffy coat layer) Patient and assessor but not investigator blinded to injection type PRTEE; success defined as an improvement in PRTEE of 25 points at final analysis 1, 3, and 6 months 20/150 lost to follow-up, not included in analysis At 6 months, both groups experienced a reduction in PRTEE pain scores. The difference in success rate between the 2 groups was not statistically significant. 12/17 patients in the AB group compared to 7/24 patients in the PRP group elected to proceed to surgery None reported No competing interests
(2) Injection of AB; tendons first bathed with bupivacaine before PRP/AB ultrasound guided injection into extensor origin under minimal pressure, second injection given at 1 month
Thanasas, 2011 28 patients with LE for at least 3 months (1) Injection of 3 mL AB Assessors blinded, patients aware of treatment as blood was drawn from ABI group VAS, Liverpool elbow score 6 weeks, 3 and 6 months 1/28 patients lost to follow‐up Improvement in VAS was larger in PRP group at every follow‐up but only the difference at 6 weeks was statistically significant. No significant difference in Liverpool elbow score was noted at any time With the exception of local pain and discomfort post-injection, no other complications noted Authors report no conflict of interest
(2) Injection of 3 mL PRP (prepared with Biomet GPS III); injections delivered deep to extensor origin with a peppering technique
Bracing
Van De Streek, 2004 43 patients with LE for at least 3 weeks, no CSI in prior month (1) Thamert Epi‐med elbow band Patient blinding not possible, further blinding not discussed MGS with pain score (1–10), PRFEQ 6 weeks 3 patients dropped out of study All outcome measures improved in a statistically significant manner for both groups, there were no statistically significant differences between groups None None reported
(2) Forearm/hand splint (combination of Thamert Epi-med band, orthoflex brace, and aluminum bar from elbow to palm with 30° dorsiflexion at wrist; both groups instructed to wear as much as possible for 6 weeks during all daily activities
Garg, 2010 70 patients (74 elbows) with LE, no treatments in prior 6 months (1) Hely & Weber Velcro wrist extension splint Patient blinding not possible, further blinding not discussed ASES Elbow Assessment Form derived score, MEP 6 weeks 25 patients lost to follow‐up, 3 did not complete outcome measurement forms; 42 patients (44 elbows) included in final analysis The improvement in the pain component of ASES score was statistically more significant for the wrist extension group at 6 weeks, both groups improved over time but there were no other significant differences in outcome measures None No benefits from any commercial entity related to the subject of this article received
(2) Hely & Weber forearm counterforce strap brace; patients instructed to wear brace during all daytime hours for 6 weeks and given standard home icing and stretching exercise
Altan, 2008 50 patients with LE for less than 12 weeks, no prior treatment with CSI or physical therapy during that period (1) Lateral epicondyle bandage Patient blinding not possible, researcher unaware of brace type VAS pain during rest and with movement, pain with pressure (0–3), grip strength, response to treatment (1, bad, to 4, excellent) 2 and 6 weeks 1/50 Both groups improved in a statistically significant manner over time in all parameters, improvement in resting pain was statistically more significant in the wrist extension splint group at 2 weeks compared to, there was no difference for other parameters between groups at either evaluation None This study was not funded by any institution and there are no conflicts of interest
(2) Wrist extension splint; patients instructed to wear braces continuously for 6 weeks, removal for no more than 1 h at a time
Struijs, 2004 180 patients with LE for at least 6 weeks, no treatment in prior 6 months (1) 9 physiotherapy (PT) treatments over 6 weeks (pulsed ultrasound, friction massage, strengthening exercises) Patient blinding not possible, assessors blinded Primary: global improvement (1, completely recovered, to 6, much worse), severity of patient's complaints (11‐point scale), pain intensity (11‐point scale), modified PFFQ 6 weeks, 1 year 10/180, included in analysis No significant differences between groups at 1 year. At 6 weeks, statistically significant benefits of PT over brace were found for pain, disability, and satisfaction. Brace superior only for inconvenience during daily activities. No significant benefit to combination therapy compared to PT, but compared to brace, severity of complaints and satisfaction were statistically significant in favor of combination None None reported
(2) Epipoint elbow band worn during the day for 6 weeks Secondary: inconvenience during daily activities (11‐point scale), pain‐free grip strength, MGS, pressure pain threshold, satisfaction of patient (11‐point scale)
(3) Combination of above 2 treatments
Luginbuhl, 2008 35 patients (36 elbows) with LE, no prior treatment with forearm band or strengthening, no more than 3 CSI in prior 6 months (1) Forearm support band worn during the day for at least 3 months Patient blinding not possible, no mention assessor blinding ROM, MGS, Nirschl–Pettrone score 6 weeks, 3 months, and 1 year 6/36 elbows lost to follow‐up, not included in analysis; 16 of remaining patients did not have follow‐up at 1 year but were included in other time points All groups had a statistically significant improvement in Nirschl–Pettrone score at 6 weeks and latest follow-up, no significant differences between groups were found at any time point None reported None reported
(2) Strengthening exercises for 3 months
(3) Combination of above 2 treatments; all patients initially treated with injection of 2–3 mL Scandicain and 40 mg triamcinolone
Physical therapy
Peterson, 2011 81 patients with LE for more than 3 months (1) Daily exercise program for 3 months Patient blinding not possible, data collection not blinded Primary: VAS (during MVC, MME) 1, 2, and 3 months 2/81, included in analysis The exercise group had greater and faster regression of pain during MVC and MME than wait‐and‐see, this difference was statistically significant Not discussed Funding sources had no involvement in study. Authors declare no conflict of interest
(2) Wait‐and‐see Secondary: extensor strength, DASH, Gothenburg QOL Instrument
Park, 2009 31 patients with LE for at least 6 weeks, no CSI in prior 3 months (1) Immediate physical therapy Patient blinding not possible VAS, modified Nirschl–Pettrone score, MEPS 1, 3, 6, and 12 months 5/31 At 1 month, the difference between groups for the modified Nirschl–Pettrone score and VAS was statistically significant in favor of the immediate therapy group. No other significant differences were found during the study period Not discussed Authors declare no conflict of interest
(2) Physical therapy started after 4 weeks oral NSAIDS
Viswas, 2012 20 patients with LE for 8–10 weeks, no previous therapy (1) Supervised exercise program Patient blinding not possible, assessor blinded VAS, TEFS 4 weeks No loss to follow‐up Both groups improved over time, but the group undergoing the supervised exercise program improved more in both outcomes than the Cyriax group, this difference was statistically significant Not discussed Authors declare no conflict of interest
(2) Cyriax Physiotherapy; 3 sessions per week for 4 weeks
Tyler, 2010 21 patients with LE for more than 6 weeks, no CSI in prior 6 weeks (1) Standard physical therapy Patient blinding not possible, assessor blinded DASH, VAS, wrist and middle finger extension strength, pain with pressure At completion of treatment No loss to follow‐up Improvements in all outcomes were greater for the eccentric group, this difference was statistically significant None Hygenic Corporation donated flexbars to study. Authors did not receive benefits from any commercial entity related to the article
(2) Standard physical therapy plus isolated eccentric wrist extensor strengthening (Thera‐Band FlexBar); treatments continued until symptoms resolved or patients were referred back to their physician with continued symptoms
Martinez, 2005 94 patients with LE for more than 3 months (1) Stretching Patient blinding not possible PFG, PRFEQ, DASH, VAS, pain and function assessment with SF‐36 patient's subjective assessment of results 6 weeks 13/94, included in intent‐to‐treat analysis At 6 weeks, all groups improved in a statistically significant manner for PFG, VAS, DASH, and SF‐36. No significant difference was found between groups Not discussed Not reported
(2) Stretching plus concentric strengthening program
(3) Stretching plus eccentric strengthening program
Extra‐corporeal shock wave therapy
Rompe, 1996 115 patients with LE for more than 12 months, unsuccessful conservative therapy in prior 6 months (1) ESWT 1000 pulses per treatment Patient and assessor blinded to treatment allocation VAS, MGS, patient assessed overall outcome (1, excellent, to 4, poor) 3, 6, and 24 weeks after last treatment 15 patients discontinued treatment during first 6 weeks, not included in analysis In the active treatment group there was a statistically significant decrease in VAS compared to placebo group at each follow‐up. There were 24 good or excellent results in the active group compared with only 3 in the placebo group No complications observed No benefits received from a commercial party related to the subject of this article
(2) Subtherapeutic ESWT; 3 treatments at weekly intervals; no other therapy allowed during course of treatment
Rompe, 2004 78 recreational tennis players with LE for at least 1 year unresponsive to conservative treatments including CSI, no therapy in prior 2 months (1) ESWT 2,000 pulses per treatment Patient and assessor blinded to treatment allocation, but 29 of 38 active treatment patients guessed allocation correctly Primary: VAS during resisted wrist extension 3 and 12 months after last treatment 4 patients withdrew to pursue additional therapy, 2 patients refused follow‐up, 2 patients only had phone follow‐up; included in intention to treat analysis At 3 months, the difference in results between treatment groups was statistically significant for all outcome measures, in favor of the active treatment group. This significant difference was not maintained at 12 months Temporary reddening after treatment and pain during treatment were common in both groups. 8 patients from the active group and 1 patient from placebo group experienced nausea during treatment No author or related institution received financial benefit from research study
(2) Placebo ESWT with shock waves deflected; 3 treatments at weekly intervals; no other therapy allowed during 3 month follow‐up Secondary: 50 % reduction in primary outcome measure, Roles and Maudsley score (1, excellent, to 4, poor), Upper Extremity Function Scale, grip strength, overall patient satisfaction
Pettrone, 2005 114 patients with LE for at least 6 months unresponsive to 2 conventional therapy; no CSI in prior 6 weeks, PT in prior 4 weeks, or NSAIDS in prior week (1) ESWT 2000 pulses per treatment Patients and evaluating physicians blinded to treatment allocation Primary: 50 % reduction of VAS during resisted wrist extension 1, 4, 8, 12 weeks and 6 and 12 months after last treatment; patients not responding to treatment could be unblinded and placebo patients offered active treatment 6 patients discontinued study prior to 12 weeks, 34 crossover patients considered lost to placebo group, 13 patients discontinued study prior to 12 months; patients included in intention to treat analysis At 12 weeks, a statistically significant difference between groups was observed for reduction in pain, improvement in upper extremity function, and overall impression of disease state in favor of the active treatment group. Most of the placebo group was lost to crossover at 12 weeks and not available for comparison past this point Active ESWT: transient treatment‐related pain (28), nausea during treatment (10) 1 or more authors received outside funding from Siemens Medical in support of their research
(2) Placebo ESWT with shock waves deflected; 3 treatments at weekly intervals targeting point of max tenderness Secondary: Upper Extremity Functional Scale, grip strength, patient overall impression of disease state Placebo ESWT: transient treatment‐related pain (13)
Haake, 2002 272 patients with LE, at least 6 months unsuccessful conservative therapy and no treatment prior 2 weeks (1) ESWT 2,000 pulses per treatment Patients and assessor blinded to treatment allocation Primary: success rate after 12 weeks (defined as score of 1 or 2 on Roles and Maudsley score and no additional treatment during study period) 6 and 12 weeks after last treatment 26/272, included in intention to treat analysis There were no significant differences between groups in any outcome measure at any time No severe adverse events noted. Most common side effects were reddening of skin (42 ESWT, 11 placebo) and pain (15 ESWT, 6 placebo) Dornier Medizintechnik provided shock wave equipment. Benefits were directed to an institution with which 1 or more authors are associated
(2) Placebo ESWT with shock waves deflected; 3 treatments received, injection of 3 mL 1 % mepivacaine administered before each treatment Secondary: Roles and Maudsley score, VAS, MGS
Speed, 2002 75 patients with LE for at least 3 months, no treatment in prior 6 weeks (1) ESWT 1,500 pulses per treatment with Sonocur Plus Unit (Siemens) Patients and assessor blinded to treatment allocation VAS 1 month after last treatment 4 patients withdrew from study Both groups had statistically significant improvement over time, no significant differences existed between groups at any time 2 patients in active treatment group withdrew due to worsening symptoms, no other adverse effects reported Funded by Cambridge Arthritis Research Endeavour
(2) Placebo ESWT, subtherapeutic doses without skin contact; 3 treatments at monthly intervals directed to site of maximum pain; no other treatments permitted during study
Melikyan, 2003 86 patients with LE who had received extensive conservative treatment and were awaiting surgery (1) ESWT starting at a low energy and gradually increased, total of 1,000 mJ/mm2 energy delivered Patient and assessor blinded to treatment allocation DASH, MGS, analgesic requirement 1, 3, and 12 months after last treatment 12/86, not included in analysis None of the outcome measures showed a statistically significant difference between groups at any time. All patients improved significantly over time None reported Supported by BUPA research fund. Benefits received from a commercial party related article are directed to a research fund or other non‐profit institution with which 1 or more author is associated
(2) Placebo ESWT with shock waves deflected; 3 treatments targeting common extensor tendon under ultrasound guidance
Chung, 2005 60 patients with untreated LE for between 3 weeks and 1 year (1) ESWT 2,000 pulses, 0.03–0.17 mJ/mm2 energy delivered Patient and assessor blinded to treatment allocation VAS (overall pain, resting pain, pain during sleep, pain during activity, pain at its worst, pain at its least), QOL on EuroQol 5D instrument (EQ5D), max PFGS, pain medication used (success defined as 50 % reduction in overall elbow pain, max overall elbow pain of 4 cm, and no use of pain medication for 2 weeks prior to 8‐week evaluation) 4 and 8 weeks after last treatment 1 patient dropped out of study, 3 were lost to follow‐up, included in intention to treat analysis At 8 weeks, there was no statistically significant difference in success rate between treatment groups No severe or moderate adverse events were observed Funded by Sport Science Association of Alberta and Alberta Provincial Canadian Institutes of Health. No author received financial benefit from research in this study
(2) Placebo ESWT with shock waves deflected; 3 treatments at weekly intervals targeting point of maximum pain, all patients provided a stretching protocol
Staples, 2008 68 patients with LE for at least 6 weeks, no oral/topical NSAID in prior 2 weeks, CSI in prior month, or oral CS in prior 6 weeks (1) ESWT 240 pulses/min, 2,000 pulses per treatment Patient and assessor blinded to treatment allocation VAS (pain, function), discomfort in daily activities, DASH, QOL using Medical Outcome Study SF-36, max PFG, Problem Elicitation Technique 6 weeks, 3 months, and 6 months after completion of 3-week treatment 2 patients withdrew after treatment, 3 patients did not attend 6‐week follow-up, 3, did not attend 3‐month follow-up, and 5 did not attend 6‐month follow-up; all patients included in intention to treat analysis There were no significant differences between groups in any outcome measures at 6‐week and 6‐month follow‐up visits. At 3 months, there was a statistically significant difference between groups in favor of the active group for DASH (work and sport), but neither section was completed by all patients Active ESWT: bruising after 1st treatment (1), pain/tenderness in arm (4), burning sensation (1) Placebo ESWT: pain and/or lump in treated area (2) Supported by Australian National Health and Medical Research Council Practitioner Fellowship.
(2) Subtherapeutic ESWT; 3 treatments at weekly intervals targeting area of maximal tenderness over common extensor tendon; no other forms of therapy including NSAIDS allowed during first 6 weeks of study
Laser
Lundeberg, 1987 57 patients with LE for at least 3 months, no CSI in prior 6 months (1) Placebo laser Patients, therapists, and assessors blinded to laser type VAS, pain and diminished power with resisted wrist dorsiflexion (0, no pain, to 3, severe pain and absent power), wrist extension strength, grip strength, patient and physician assessment of treatment (satisfactory/unsatisfactory) Every other week during treatment and at least 3 months after treatment Not discussed The 3 treatment groups showed no significant difference in any outcome measure at any time None reported Work supported by grants from The Royal Swedish Academy of Sciences and Stiftelsen Clas Groschinskys minnesfond
(2) Ga–As pulse laser, 904 nm wavelength
(3) He‐Ne continuous‐wave laser, 632.8 nm wavelength; 2 treatments per week given over 5 to 6 weeks with laser stimulating 10 acupuncture points
Haker, 1990 49 patients with LE for at least 1 months (1) Ga–As Mid 1500 IRRADIA laser, wavelength 904 nm Patients, therapist, and assessors blinded to laser type Subjective physician assessment (1, excellent, to 5, unchanged or worse), max pain free grip After last treatment and 3 months post‐treatment, 1 year follow-up with questionnaire 2 patients withdrew after treatment, 7 additional patients withdrew after 3 months follow‐up At 3 months, there were no significant differences between treatment groups for any outcome measure None reported Work supported by grants from Foundations of Karolinska Institutet
(2) Placebo laser; 10 treatments, 2–3 times weekly targeting the extensor group of muscles
Krasheninnikoff, 1994 48 patients with LE for at least 1 month, no CSI in prior 3 weeks or other simultaneous treatment (1) Active Ga–Al–As laser (30 mW/830 nm) Patients, therapists, and assessors blinded to laser type Duration of pain, pain relief, symptoms, patients subjective impression of effect Telephone interview at 10 weeks post‐treatment 12/48; excluded from analysis At 10 weeks, there were no significant differences between treatment groups for any outcome measure None reported None reported
(2) Placebo laser; 8 treatments 2 times per week targeting tender points on lateral epicondyle and in forearm extensors
Papadopoulos, 1996 29 patients (31 elbows) with LE (1) Ga–Al–As laser (50 mW/820 nm) Patients, operator, and assessor blinded to treatment VAS, pain threshold with exercising forearm extensors 1st, 4th, and 6th treatment visit No loss to follow‐up There were no statistically significant differences between groups for any outcome measure at any time None reported Loan of equipment by Omega Technologies
(2) Placebo laser; 3 treatments per week for 2 weeks targeting most tender point
Vasseljen, 1992 30 patients with LE, no other treatments in prior 3 weeks (1) Ga–As laser (18 mW/904 nm) Patients, operator, and assessor blinded to treatment MGS, extension weight test, point of pain with wrist flexion, VAS, patient subjective assessment of improvement Post‐treatment and 4 weeks after last treatment 1 patient had alternative treatment during study, dropped and was replaced by another patient Statistically significant improvement in laser compared to placebo was found for VAS and MGS 4 weeks after treatment Not discussed Supported by grant from Norwegian Fund for Post‐Graduate Education in Physiotherapy. Schreuder Instrumenten supplied lasers
(2) Placebo laser; no other treatments allowed during study; 8 treatments were given at a frequency of 3 times per week
Basford, 2000 47 patients with LE for more than 30 days, no CS administered for any reason in prior 30 days (1) Nd:YAG CW 1.06 μm continuous wave laser (204 mW/cm2) Patient, therapists, and physician blinded to laser type Primary: VAS for max pain in last 24 h, max tenderness on palpation, and patient perception of change 1st, 6th, and 12th treatment visit and 1 months post‐treatment 1 patient did not return for 1 months follow‐up There were no statistically significant differences between groups for any outcome measure at any time No significant treatment side effects were noted Authors have not received a benefit from any commercial party having a direct financial interest in the results of the research supporting this article
(2) Inactive laser; 3 treatments per week for 4 weeks targeting 7 points along the symptomatic forearm Secondary: grasp and pinch strength, pain with grasp and pinch
Stergioulas, 2007 62 patients with LE for at least 5 weeks, no treatment in prior month (1) Ga–As continuous wave laser (40 mW, 904 nm) plus plyometric exercises Patients and physical therapist performing assessments blinded to laser type VAS (pain at rest, with palpation of lateral epicondyle, during resisted extension), ROM, grip strength, and free weight test Post‐treatment and 8 weeks after last treatment 12/62 patients left study to pursue other treatments Relative to the placebo group, the active treatment group had a statistically significant decrease in pain at rest, with palpation, during resisted middle finger extension, and during grip strength testing and an increase in grip strength and weight testing at both follow‐up intervals No complications reported by patients None reported
(2)Placebo laser plus plyometric exercise, 12 treatments over 8 weeks
Lam, 2007 39 patients with LE, no previous episode of lateral elbow pain on same side or any prior treatment (1) Ga–As laser (25 mW, 904 nm) plus exercise program Patient and physician blinded Mechanical pain threshold, MGS, VAS, DASH 5th and 9th treatment session and 3 weeks post‐treatment Not discussed A statistically significant difference in improvement for all outcome measures were found between the laser group and the placebo group in favor of the active laser group, except in 2 subsections of DASH None reported None reported
(2) Placebo laser plus exercise program; 3 treatments per week for 3 weeks targeting all tender points
Emanet, 2010 49 patients (50 elbows) with LE for less than 3 months, no PT or CSI in prior 2 months (1) Ga–As laser (100 W/905 nm) with exercise program Patients and assessor blinded to laser type VAS, pain with pressure over lateral epicondyle, DASH, PRTEE, pain free grip strength, Nottingham Health Profile (NHP) 3 and 12 weeks 3/49, excluded from analysis At 3 weeks, a statistically significant improvement in all outcomes was observed in both groups. At 12 weeks, a statistically significant improvement in favor of the active laser group was found for pain with resisted wrist extension, pain with pressure, DASH, PRTEE, and NHP (pain) None reported No competing financial interests exist
(2) Placebo laser with exercise program; 5 treatments a week for 3 weeks targeting the 2 most sensitive points around the lateral epicondyle

AB autologous blood, ASES American Shoulder and Elbow Society, CSI corticosteroid injection, ESWT extra-corporeal shock wave therapy, LE lateral epicondylitis, MEPS Mayo Elbow Performance Score, MGS maximum grip strength, MME maximum muscle elongation, MPFGS Maximum Pain Free Grip Score, MVC maximum voluntary contraction, PFFQ Pain Free Function Questionnaire, PFG pain free grip, PRFEQ Patient Rated Forearm Evaluation Questionnaire, PRP platelet-rich plasma, PRTEE Patient Rated Tennis Elbow Evaluation, PT physiotherapy, QVAS Quadruple Visual Analogue Scale, QOL quality of life, ROM range of motion, SF-12 Short Form 12, SF-36 Short Form 36, TEFS Tennis Elbow Function Score, VAS Visual Analogue Scale (for pain unless otherwise specified)