Gupta 2019.

Study characteristics
Methods	Design: single‐centre parallel‐group 2‐arm open‐label randomised controlled trial Setting: Department of Sports Medicine, Sir Ganga Ram Hospital, New Delhi, India Timing: July 2016 to June 2017 Interventions: platelet‐rich plasma injection vs glucocorticoid injection Sample size: 80 participants (40 + 40);initially calculated at 66 (power 90%, α = 0.05, s ~ 1.5), assuming a 20% dropout rate; total sample size was finally set at 80 (40 per group) based on 1.2‐point difference in VAS Analysis: intention‐to‐treat
Participants	Number of participants Number of participants screened for eligibility: 98. Number excluded: 18 (did not meet inclusion criteria = 12; declined to participate = 6) Number randomised: 80 (40 to platelet‐rich plasma group, 40 to glucocorticoid group) Number included in analysis: 80 (40 in platelet‐rich plasma group, 40 in glucocorticoid group) Inclusion criteria Aged 18 to 55 years Diagnosed lateral epicondylitis Unresponsive to conservative therapy for > 3 months Exclusion criteria Other elbow condition (cervical radiculopathy, osteochondritis dissecans) Systemic conditions (rheumatoid disorder, diabetes) Baseline characteristics PRP group (n = 40) Mean age (years): 42.4 No. male:female: 22:21 Mean duration of symptoms (weeks): 15.2 No. with vigorous manual occupation: 10 No. with non manual occupation: 30 Mean VAS: 81 Mean DASH: 87.15 Mean Mayo Elbow Performance Scale score (MEPS): 49.5 Mean grip strength: 67.7 Glucocorticoid group (n = 40) Mean age (years): 39.4 No. male:female: 12:25 Mean duration of symptoms (weeks): 16.5 No. with vigorous manual occupation: 8 No. with non manual occupation: 32 Mean VAS: 77.5 Mean DASH: 85.9 Mean Mayo Elbow Performance Scale score (MEPS): 54 Mean grip strength: 58.9 Pretreatment group differences: no baseline differences between groups
Interventions	PRP group Patients were kept anti‐inflammatory‐analgesic‐free for 2 weeks (to allow for relative washout of the drugs). Out of 20 mL whole venous blood, 18 mL was transferred into 4 red‐capped plain tubes (labelled 1, 2, 3, and 4) with 4.5 mL each, and 2 mL was used for cell counts. Two‐spin centrifugation was adopted, with the first at 160 g for 12 minutes at room temperature. The sample was segregated into 3 layers ‐ upper plasma, middle buffy coat, and lower red cell layer. Supernatant plasma and buffy coat from each tube (total 10 to 12 mL) were pipetted into another set of red‐capped plain tubes labelled A and B (5 to 6 mL per tube) under laminar flow. A second spin at 460 g for 18 minutes was then provided at room temperature. Platelet pellets were collected at the bottom of each tube. Around 3 mL of platelet‐poor plasma was discarded, and roughly 2 mL of plasma with platelet pellets was thoroughly mixed in each tube. The final PRP thus obtained was around 4 mL and was transferred from both tubes into 1 plain tube labelled “PRP”, was stored for 15 minutes at room temperature, and was subjected to a platelet count. On a 90‐degree flexed elbow and pronated forearm (passively stretched ECRB allowed clearer identification), the common extensor origin was identified, painted, and draped. Bony landmarks (lateral epicondyle, supracondylar ridge, olecranon, and radial head) were palpated; a 22‐gauge needle was introduced along the supracondylar ridge (proximal to lateral epicondyle) and was gently advanced into the undersurface of the ECRB and the common extensor tendon via a peppering technique: single skin penetration and 10 to 20 tendon penetrations (without emerging from the skin) Glucocorticoid group 40 mg of triamcinolone with 2% xylocaine was injected after peppering Co‐intervention Sterile dressings were removed 2 days later. Discharged after a brief 30‐minute rest, all patients followed standardised rehabilitation (limb rest ‐ 3 days, need‐based cold fomentation and oral paracetamol). Additional requirements were noted on subsequent office visits. Gentle range of movement (ROM) and isotonic exercises were prescribed after a week. Resistive training of wrist extensors using TheraBand (THERABAND, Akron, OH) and rotator cuff and periscapular muscle exercises were started at 3 weeks post intervention
Outcomes	Outcomes were measured at 6 weeks, 3 months, and 1 year Outcomes VAS pain 0 to 100 (higher is worse) DASH score 0 to 100 (higher is worse) Mayo Elbow Performance Score (MEPS) 0 to 100 Grip strength Requirements of pain relief medications Feeling of “complete relief from pain” (Yes/No) Outcomes used in this review Mean pain VAS 0 to 100 (higher is worse) Mean function DASH 0 to 100 (higher is worse) Grip strength Treatment success (complete pain relief)
Notes	Funding: not reported Trial registration: none found Withdrawal: 0 withdrawals in both groups Adverse events: Study authors report, "no major adverse events were reported in any patient" Serious adverse events: none reported Analysis: we transformed VAS to 0 to 10 scores. Table 2 reports S values, which are unusually small for SD and large for SE. We used them as SD (Analysis 2.2; Analysis 2.3), and we conducted a sensitivity analysis using imputed SD. We contacted study authors to clarify whether variance measures were SD or SE; however they did not respond
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A computer‐generated random sequence was utilised
Allocation concealment (selection bias)	Low risk	Quote: "letters “A” (PRP) and “B” (CS) placed in identical, opaque, sealed and stapled envelopes by an independent researcher (not involved with the care of the patients) to minimise selection bias. The allocation sequence was concealed from the surgeon, and the envelopes were only opened at the time of allocation of intervention"
Blinding of participants and personnel (performance bias) All outcomes	High risk	The need for patient identification during blood withdrawal for PRP injection made the study a non‐blinded one
Blinding of outcome assessment (detection bias) for self reported subjective outcomes (pain, function, treatment success, quality of life)	High risk	Study authors did not attempt blinding
Blinding of outcome assessment (detection bias) objective outcomes	High risk	Study authors did not attempt blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study
Selective reporting (reporting bias)	Unclear risk	No trial registration or protocol is available. Study authors report data at 6 weeks, 3 months, and 1 year for all outcomes specified in the methods section
Other bias	Low risk	None is apparent