Methodological characteristics, inclusion and exclusion criteria, and follow-up completion rates of the included studies
| Author | Study type | Randomisation method | Blinding method | Allocation concealment | Statistical power calculation | Baseline comparison | Inclusion criteria | Exclusion criteria | Follow-up completion |
| Berrazueta et al 13 | Double-blind placebo-controlled RCT. | Not stated. | Not stated. | Not stated. | No. | Not done—greater pain and joint restriction in GTN group compared with control group as reported by Cumpston et al. 14 | Acute supraspinatus tendintis (symptoms of <7 days) with tenderness and limitation of motion with increased pain with abduction. | Chronic shoulder pain, cervical spain/nerve root/brachial plexus lesions, cardiac/pulmonary/systemic disease, bone lesions or calcification at tendon insertion on X-ray, glaucoma, hypersensitivity to nitrates and NSAIDs. | 100% |
| Pons et al 32 | Double-blind placebo-controlled RCT. | Random number table. | Not stated. | Not stated. | No. | No difference. | Acute supraspinatus tendintis (symptoms of <6 weeks, pain on moving arm 60°–120° abduction, positive impingement test, Jobe, Gerbe or Pate tests, no response to 1-week NSAIDs treatment. | Adhesive capsulitis, biceps tendinitis, allergy/intolerance to GTN and those receiving GTN patches for heart disease. | 100% |
| Steunebrink et al 35 | Double-blind placebo-controlled RCT. | Sealed, coded envelopes; block randomisation (four participants per block). | Not stated. | Coordinated by independent physician; data after randomisation stored at secret location. | Yes; sample size adequate for 80% power. | Significantly higher VAS score during activities in control versus treatment group. | 18–40 years, clinical diagnosis patellar tendinopathy (pain with activity or tenderness/thickening). | Symptoms of >24 months, VISA-P score >80, previous surgery or injection, previous eccentric programme in last 2 years, serious illness, pregnancy and contraindications to GTN. | 82.5% |
| Giner-Pascual et al 33 | Double-blind placebo-controlled RCT. | Allocation based on hour and date of clinic appointment. | Not stated | Not stated | No | GTN group older than control group; no baseline comparison of outcome measures. | >18 years, complete motor paraplegia, full-time wheelchair user, symptoms >3 months, diagnosis on MRI or US. | Tetraplegia, incomplete paraplegia, treatment with NO drugs and heart disease or hypotension. | 91% (only 66.7% completed treatment) |
| Paoloni et al 29 | Double-blind placebo-controlled RCT. | ‘Coded randomisation’. | Active and placebo patches indistinguishable from each other. | Randomisation supervised by senior pharmacist. | Yes; sample size adequate for 90% power. | No difference. | >18 years, supraspinatus tendinopathy clinically (signs of impingement and pain in empty can position) and on MRI. | Symptoms of <3 months, pregnancy, previous surgery, previous shoulder dislocation, distal neurology, IHD and steroid injection last 3 months. | 91% |
| Kane et al 30 | Single-blinded, non-placebo-controlled RCT. | ‘Sealed envelopes’. | Not stated. | Not stated. | No. | No difference. | Achilles tendinopathy confirmed on US and MRI. | Symptoms of <3 months, previous surgery or injection and contraindications to GTN. | 90% |
| Paoloni et al 28 | Double-blind placebo-controlled RCT. | Not stated. | Not stated. | Randomisation controlled by senior pharmacist. | Yes; sample size adequate for 90% power. | Significantly lower ankle plantar flexor mean total work in treatment than placebo group. | >18 years, diagnosis of non-insertional Achilles tendinopathy both clinically (insidious onset Achilles pain, tender nodule 2–6 cm proximal to calcaneal insertion) and radiologically (US scan with no tear). | Symptoms of <3 months, previous surgery, previous ankle dislocation, distal neurology, steroid injections in last 3 months and pregnancy. | 89% |
| Ozden et al 34 | Double-blind placebo-controlled RCT. | Not stated. | Not stated. | Not stated. | Yes; sample size adequate for 80% power. | No difference. | Tenderness and pain over lateral epicondyle, positive Mill’s sign, positive chair lift test, symptoms of >3 months and resisted wrist extension. | Surgery, effusion, radiculopathy, ulnar entrapment, fracture, infection, high ESR and injections. | 100% |
| Paoloni et al 31 | Double-blind, placebo-controlled RCT. | Stratified computer-generated randomisation. | Placebo and active patches identical in appearance. | Not stated. | Yes; sample size adequate for 80% power. | No difference in demographics; statistical analyses of outcome measures between groups not performed/stated. | 18–70 years, symptoms of >3 months, VAS score >4/10 with provocative testing (ORI-TETS). | BMI >38, requirement of oral or topical analgesia, injection last 3 months, worker’s compensation cases, previous use of GTN, cardiac disease, pregnancy and previous surgery/fracture/dislocation. | 88% |
| Paoloni et al 27 | Double-blind placebo-controlled RCT. | ‘Coded randomisation’. | Not stated. | Randomisation supervised by senior pharmacist. | Yes; sample size adequate for 90% power. | No difference in dropouts or outcome measures; statistical comparison of demographics between groups not reported. | >18 years. | Surgery, previous dislocation of wrist/elbow, injection in last 3 months and distal neurology. | 86% |
BMI, body mass index; ESR, erythrocyte sedimentation rate; GTN, glyceryl trinitrate; IHD, ischaemic heart disease; NO, nitric oxide; NSAIDs, non-steroidal anti-inflammatory drugs; ORI-TETS, Orthopaedic Research Institute Tennis Elbow Testing System; RCT, randomised controlled trial; US, ultrasound; VAS, visual analogue scale; VISA-P, Victorian Institute of Sports Assessment – Patella.