Dunkow 2004.

Methods	Randomised controlled trial. Pre‐specified sample size calculation not reported. Overall validity of results: high risk of bias.
Participants	45 participants (47 elbows); 24 elbows in the open group, with mean age of 43 years (range 30 to 58), and 23 elbows in the percutaneous group, mean age of participants was 46 years (range 32 to 58). Inclusion criteria: previous conservative treatment for 12 months (2 injections of 80mg hydrocortisone into the common extensor origin, and modification of activity); diagnosis of lateral epicondylitis‐ resisted extension of the middle finger and pinch grip with the wrist in extension provoking pain over the common extensor region. Exclusion criteria: none reported.
Interventions	Group 1 (24 elbows), 'Open group': open surgical procedure consisting of 7cm incision over the common extensor origin, reflection of the extensor carpi radialis longus (ECRL) to expose the origin of the extensor carpi radialis brevis (ECRB), removal of the damaged portion of the tendon, and 3 drill holes in the lateral epicondyle. Group 2 (23 elbows), 'Percutaneous group': percutaneous surgical technique involving a 1cm incision over the mid‐point of the lateral epicondyle to expose the common extensor origin, which was divided, and a 1cm gap created at the common extensor origin with the wrist flexed. In both treatment groups, operations performed under tourniquet control, elbows wrapped in a wool and crepe bandage for 7 days; then bandage removed to begin post‐operative physiotherapy regime which was identical for both groups and supervised by the same physiotherapist.
Outcomes	Assessed 12 months post‐operatively (range 12 to 14 months and 11 to 13 months in the percutaneous and open groups respectively. 1) American Academy of Orthopaedic Surgeons Disability of Arm, Shoulder and Hand (DASH) score: 100 point scale, 100 = maximum disability, 0 = minimum disability; assessed pre‐operatively and post‐operatively, and change in score from pre‐ to post‐operatively. 2) DASH score ‐ sports function section, and high‐performance work score (0 to 100 point scale, 100 = maximum disability) were also reported separately. 2) Participant satisfaction: self‐assessment of outcome rated as very pleased, satisfied or not satisfied with result 3) Median time to return to work.
Sources of funding	Authors state, 'No benefits in any form have been received, or will be received from a commercial party related directly or indirectly to this article'.
Notes	Co‐morbidity at baseline was reported to be 2 (range 1 to 3) and 2 (range 0.25 to 2) in the percutaneous and open groups although it is not clear what this was a measure of and whether the score reported is a mean or median. Post‐operative physiotherapy regime not described. The trial authors reported median and interquartile ranges DASH scores and time to return to work; for the purposes of data extraction into RevMan, we assumed the median approximates the mean, and calculated the standard deviation assuming the width of the interquartile range will be approximately 1.35 standard deviations. We did not extract the participant satisfaction data, as the trial authors used a 3 point categorical scale that could not be logically dichotomized.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Comment: sequence generation not described
Allocation concealment?	Unclear risk	Quote: 'Patients were randomised to one of two groups using sealed envelopes'. Comment: unclear if envelopes were sequentially numbered and opaque, and unclear who opened the envelopes
Blinding? All outcomes	Unclear risk	Comment: unclear for participants and outcome assessor, surgeon unblinded (single surgeon performed all procedures)
Incomplete outcome data addressed? All outcomes	Low risk	Comment: no missing outcome data, and appears no participants crossed over to the other treatment
Free of selective reporting?	Unclear risk	Comment: pain not reported (may not have been measured); adverse events not reported, unclear if these were measured
Free of other bias?	Unclear risk	Comment: the trial does not report the number of participants (if any) screened and excluded prior to randomisation