Methods

Study design: quasi‐randomised controlled trial.

Duration of the study: from January 2011 to December 2013.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not registered.

Funding sources: none known.

Participants

Place of study: Mazandaran University of Medical Sciences, Iran.

Number of participants assigned: 50 participants (50 fingers); 25 percutaneous surgery and 25 steroid injection.

Number of participants assessed: 50 participants (50 fingers); 25 percutaneous surgery and 25 steroid injection.

Inclusion criteria:

• Participants who presented with grade II or III trigger digit as classified by Quinnell 1980.

Exclusion criteria:

• Participants with trigger thumbs were excluded.

Age:

• Total of participants (mean/range): 48/40 to 65 years.

Gender:

• Total of participants: 20 male; 30 female.

Side:

• Total of participants: 31 dominant hand and 19 non‐dominant hand.

Digits:

• Total of participants: 9 index, 26 long, 10 ring and 5 little.

Classification of injury: trigger fingers were classified according to the Quinnell 1980 criteria (graded I to V).

Interventions

Timing of intervention: not reported.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous surgery

• Steroid injection: 1 mL of triamcinolone mixed with 1 mL of 1% lidocaine was injected into the tendon sheath and around the nodule.

Rehabilitation: the authors did not clearly describe whether all study participants received physiotherapy.

Any co‐interventions: analgesia was given for 3 days.

Outcomes

Length of follow‐up:

• Follow‐up was 9 months.

• Participants were evaluated weekly for 6 consecutive weeks.

Loss of follow‐up: not reported.

Primary outcomes:

Symptomatic relief: the authors did not clearly define what they considered as "symptomatic relief" and reported incomplete data, stating only that both groups showed improved symptoms within 2 weeks of follow‐up, and after 2 weeks the response was better in the steroid injection group.

Patient satisfaction.
Complications (adverse events):

• Partial loss of movement.

• Dysaesthesia.

• Pulley or tendon injury.

• Skin atrophy or hypopigmentation.

Secondary outcomes:

Pain:measured by Visual Analog Scale (VAS)

Recurrence of triggering: the authors did not clearly define what they considered recurrence.

Outcomes included in this review:

Pain: measured by Visual Analog Scale (VAS).

Patient satisfaction.

Recurrence of triggering

Adverse events:

• Partial loss of movement.

• Dysaesthesia.

• Pulley or tendon injury.

• Skin atrophy or hypopigmentation.

Notes

• Pain and patient satisfaction were described incompletely (no numerical data were reported), and we were unable to include these data in the results.

• We tried unsuccessfully to contact the authors to obtain further information on pain (VAS‐score) and patient satisfaction.

• The authors did not clearly report if there were losses to follow‐up.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Quasi‐randomised controlled trial.

Participants were randomised to either steroid injection or percutaneous surgery using their birth year. Those with even numbers were allocated to the steroid group and uneven numbers to the percutaneous group.

Allocation concealment (selection bias)

High risk

Quasi‐randomised controlled trial.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Not reported. The authors did not clearly described whether all study participants completed follow‐up.

Selective reporting (reporting bias)

High risk

There was no protocol published.

Outcome of interest in the review (resolution of trigger finger and functional status of the hand) were not reported. Pain was reported incompletely, without numerical values.

Other bias

Unclear risk

The authors did not report data about baseline balance, and they did not clearly report about care providers and rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: 1 May 2007 to 31 December 2008.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: Thailand.

Number of participants assigned: 142 participants (160 fingers); 80 percutaneous surgery and 80 open surgery.

Number of participants assessed: 142 participants (160 fingers); 80 percutaneous surgery and 80 open surgery.

Inclusion criteria:

• Failure non‐surgical treatment for 3 months.

• At least 1 local steroid injection.

• Grade II, III or IV trigger digit as classified by Froimson 1993.

Exclusion criteria:

• Clinically active osteoarthritis of the affected hand.

Age:

• Percutaneous surgery (mean/range): 48.6/22 to 72 years.

• Open surgery (mean/range): 46.2/24 to 76 years.

Gender:

• Percutaneous surgery: 28 male; 44 female.

• Open surgery: 30 male; 40 female.

Side: not reported.

Digits:

• Percutaneous surgery: 23 thumb, 8 index, 27 long, 20 ring and 2 little.

• Open surgery: 26 thumb, 6 index, 24 long, 23 ring and 1 little.

Classification of injury:

Trigger fingers were graded according to Froimson’s modification of Quinnell’s classification (graded I to IV) (Froimson 1993).

Interventions

Timing of intervention: at least 3 months.

Duration of treatment

• Percutaneous surgery (mean operative time): 1.8 minutes.

• Open surgery (mean operative time): 2.2 minutes.

Type of intervention:

• Percutaneous surgery: a full handle knife 45º was inserted 2 mm proximal to the proximal edge of the A1 pulley; when the distal edge was reached, the knife was moved distally to proximally, releasing the A1 pulley.

• Open surgery: transverse incision was made over the involved metacarpal head, and the A1 pulley was transected under direct observation.

Rehabilitation: not reported.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up:

• Follow‐up was 8 weeks.

• Participants were evaluated at 1, 2, 3, 4, 6, and 8 weeks.

Loss of follow‐up: none lost to follow‐up.

Primary outcomes:

Operative time.

Range of motion of finger PIP or thumb IP.

Patient satisfaction score.

Patient pain score.

Surgical complications (adverse events):

• Section of all or a portion of the A2 pulley.

Neurovascular injury.

Outcomes included in this review:

Resolution of trigger finger: considered as the relief of pain and the cessation of finger locking after the procedure. Although reported in the study, it was not considered as primary outcome by the author.

Pain (0 to 3 scale).

Satisfaction scores (0 to 3 scale).

Recurrence of triggering: the author did not clearly define what he considered recurrence. Although reported in the study, it was not considered as primary outcome by the author.

Adverse events measured by:

• Section of all or a portion of the A2 pulley.

Neurovascular injury.

Notes

• Pain and satisfaction scores were presented graphically only, and we were unable to include these data in the analyses.

• We tried unsuccessfully to contact the authors to obtain further information on absolute numerical values ​​and standard deviations for pain (VAS‐score) and satisfaction scores.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

High risk

No protocol was published.

Pain and satisfaction score were measured by non‐validated instruments and the results were exhibited in figures, with inaccurate values and no measure of variance, compromising the assessment.

Other bias

Low risk

There was no baseline imbalance, and no risk of bias was associated with care providers or differences in rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: January 2007 to May 2007.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: Varese, Italy.

Number of participants assigned: 30 participants (30 fingers); 15 open surgery and 15 steroid injection plus hyaluronic acid injection ultrasound‐guided.

Number of participants assessed: 30 participants (30 fingers); 15 open surgery and 15 steroid injection plus hyaluronic acid injection ultrasound‐guided.

Inclusion criteria:

• Age between 35 and 70 years old.

• Clinical signs and symptoms of stenosing tenosynovitis of the flexor tendons and in whom diagnosis was confirmed by ultrasound.

Exclusion criteria:

Trigger finger grade IV.

Comorbidities:

• Diabetes mellitus.

• Rheumatoid arthritis.

• Hypercholesterolemia.

• Hypotension.

• Hypertension.

Age:

• Open surgery (mean/range): 52.13/40 to 70 years.

• Injection group (mean/range): 52.86/35 to 69 years.

Gender:

• Open surgery: 4 male; 11 female.

• Injection group: 6 male; 9 female

Side: not reported.

Digits: 16 thumb, 7 ring and 7 long.

Classification of injury:

Trigger fingers were graded according to Froimson’s modification of Quinnell’s classification (graded I to IV) (Froimson 1993).

Interventions

Timing of intervention: average period: 3.5 months of symptoms (range 1 to 6 months).

Duration of treatment: the duration of the surgical procedure was not reported. Hyaluronic acid injection was injected 10 days after steroid injection (injection group).

Type of intervention:

• Open surgery: patients underwent open surgery by conventional technique under locoregional anaesthesia and a haemostatic pressure cuff inflated around the upper arm. The procedure was carried out on a day surgery basis and patients were discharged in the evening with a compression dressing to be kept in place for 4 days until it was changed in the outpatient clinic. Sutures were removed 2 weeks after surgery.

• Injection group: inject methylprednisolone acetate 40 mg/1 mL with 0.8 mL lidocaine chlorhydrate 2% into the sheath of the flexor tendons, distally to the A1 pulley, under ultrasound guidance. 10 days later, 1 mL 0.8% hyaluronic acid was injected using the same technique.

Rehabilitation:

• Open surgery: the patients started mobilisation of the finger after 4 days.

• Injection group: therapy was not reported.

Any co‐interventions: 10 patients in open surgery group needed physiotherapy and local or oral analgesics for complete resolution of symptoms, which was approximately 30 to 40 days post surgery.

Outcomes

Length of follow‐up: 1 year. Participants were evaluated before intervention, at 6 weeks and at 3, 6, and 12 months.

Loss of follow‐up: none lost to follow‐up.

Primary outcomes:

Resolution of trigger finger: considered as the remission of symptoms within 6 weeks, with no recurrence within 6 months.

Recurrence of triggering: was considered the return of any degree of triggering after a full remission period of trigger finger.

Secondary outcomes:

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Functional status of the hand: was used DASH (0 to 100%).

Satisfaction scores: measured by Satisfaction Visual Analog Scale (SVAS: 0 to 10 scale).

Outcomes included in this review:

Resolution of trigger finger

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Functional status of the hand: measured by DASH (0 to 100%).

Satisfaction scores: measured by Satisfaction Visual Analog Scale (SVAS: 0 to 10 scale).

Recurrence of triggering:

Adverse event (Although reported in the study, it was not considered as primary outcome by the authors):

• Partial loss of movement.

• Algodystrophic syndrome.

Notes

• The authors did not report standard deviations on VAS, DASH and SVAS scores.

• We tried unsuccessfully to contact the authors to obtain further information on standard deviations for pain (VAS‐score), functional status of the hand (DASH‐score) and patient satisfaction (SVAS‐score).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

High risk

No protocol was published.

Although outcomes of interest in the review were reported, the authors failed to report any measure of variance for the validated instruments (VAS, DASH and SVAS).

Other bias

High risk

There were different rehabilitations in 2 groups.

In both groups, at the first follow‐up visit, participants were advised to mobilise the finger, depending on the level of pain experienced, but 10 participants in open surgery group needed physiotherapy, and local and/or oral analgesics for complete resolution of symptoms.

Methods

Study design: randomised controlled trial.

Duration of the study: January 2005 to February 2007.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: China.

Number of participants assigned: 86 participants (97 trigger thumbs); 42 participants (47 thumbs) in percutaneous surgery group and 44 patients (50 thumbs) in steroid injection group.

Number of participants assessed: 83 participants (93 trigger thumbs); 41 participants (46 thumbs) in percutaneous surgery group and 42 patients (47 thumbs) in steroid injection group.

Inclusion criteria: idiopathic adult trigger thumbs with grade III–V on the Quinnell classification.

Exclusion criteria: patients who had rheumatoid arthritis, diabetes mellitus or chronic systemic disease.

Age:

• Percutaneous surgery: the average was 48 years ( 27 to 65).

• Steroid injection: the average age was 49 years (28 to 72).

Gender:

• Percutaneous surgery: 29 female and 12 male.

• Steroid injection: 28 female and 14 male.

Side:

• Percutaneous surgery: 33 right and 13 left.

• Steroid injection: 36 right and 11 left.

Digits: 97 thumbs.

Classification of injury:

The trigger thumb was graded according to Quinnell classification (graded I to V) (Quinnell 1980).

Interventions

Timing of intervention: 4 months' duration of the symptoms in both groups.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous release with a new instrument called mini scalpel‐needle (MSN), based on acupuncture.

• Injection with 1 ml triamcinolone acetonide (10 mg/ml). Previous injection of the 0.5 ml of 1% lidocaine was infiltrated into the skin and tissue around the tendon sheath. When necessary a second injection at 1 week was realised.

Rehabilitation: did not have rehabilitation.

Any co‐interventions: topical NSAIDs were administered for 3 days with the occasional use of paracetamol for pain control in both groups when necessary.

Outcomes

Length of follow‐up: 1 year

Loss of follow‐up: 1 patient (1 thumb) in percutaneous surgery group and 2 patients (3 thumbs) in steroid injection group were lost to follow‐up at 12 months and were excluded.

Primary outcomes:

Successful procedure or satisfaction (resolution of trigger finger): "satisfactory" was considered to be participants who progressed with pain score lower than or equal to 1 (VAS scale) and cessation of triggering.

Pain:measured by Visual Analog Scale (VAS:0 to 10 scale).

Adverse events measured:

• Infection.

• Tendon bowstringing (tendon or pulley injury).

Neurovascular injury.

Outcomes included in this review:

Resolution of trigger finger: "satisfactory" was considered to be participants who progressed with pain score lower than or equal to 1 (VAS scale) and cessation of triggering.

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Recurrence of triggering: the authors did not evaluate it as a study outcome, but they published indirect data on recurrence.

Adverse events measured:

• Infection.

• Tendon bowstringing (tendon or pulley injury).

Neurovascular injury.

Notes

• The follow‐up data were collected by clinical examination (69 participants) or by telephone interview (14 participants).

• In the study, the authors use the words "satisfaction", "success of procedure" and "satisfactory results" as synonyms for trigger finger resolution.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

The patients were randomly assigned by the selection of number 1 or 2 from sealed envelopes in the presence of a witness.

Allocation concealment (selection bias)

Low risk

Sealed envelopes were used.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Missing outcome data were balanced in numbers across intervention groups; 1 thumb 1/47 (2%) in percutaneous release group and 3 thumbs 3/50 (6%) in steroid injection group were lost to follow up of 12 months and excluded; however they were reported.

Selective reporting (reporting bias)

High risk

No protocol was published.

Functional status (as primary outcome using validated instruments to measure hand function) was not evaluated by the authors.

Other bias

Unclear risk

There was no baseline imbalance, but the authors did not describe about care providers or rehabilitation.

Methods

Study design: quasi‐randomised controlled trial.

Duration of the study: during the year 2003.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not registered.

Funding sources: none known.

Participants

Place of study: Oldenburg, Germany.

Number of participants assigned: 36 participants (36 fingers); 20 percutaneous surgery and 16 open surgery.

Number of participants assessed: 36 participants (36 fingers); 20 percutaneous surgery and 16 open surgery.

Inclusion criteria: patients with primary trigger finger and age between 18 and 80 years.

Exclusion criteria: patients with trigger thumb, more than 1 trigger finger, previous operations of the upper extremity, evidence of symptomatic synovitis, or diseases possibly influencing pain scores or hand function (e.g. nerve entrapments, neuropathy, diabetes, and rheumatoid arthritis), or patients with any joint extension lag.

Age:

• Percutaneous surgery (mean): 62 years (41 to 79).

• Open surgery (mean): 64 years (39 to 88).

Gender:

• Percutaneous surgery: 9 male and 11 female.

• Open surgery: 7 male and 9 female.

Side: not reported.

Digits: not reported.

Classification of injury: not reported.

Interventions

Timing of intervention:

• Percutaneous surgery (mean): 7 months (1 to 36).

• Open surgery (mean): 12 months (1 to 60).

Duration of treatment:

• Percutaneous surgery: 26 seconds.

• Open surgery: 4 min 17 seconds.

Type of intervention:

• Percutaneous surgery: the affected digit was placed in extension. At the proximal level of the A1 pulley, a L15 blade scalpel was put through the skin and pushed palmar‐ward with the backside of the knife. Then the knife was positioned on top of the distal end of the A1 pulley centred at the palmar axis of the tendon sheath, and the sharp side of the knife was directed dorsally. External pressure from the surgeon’s finger on the skin is performed to put the knife through the A1 pulley.

• Open surgery: a longitudinal incision was placed in a skin crease at the level of the A1 pulley. The neurovascular structures were preserved by Langenbeck hooks. The A1 pulley was opened longitudinally. The skin was closed with 4.0 Ethylon sutures.

Rehabilitation:

A direct postoperative mobilization protocol was used in both groups. The authors did not describe which protocol was used, nor for how long it was used.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up: 12 weeks.

Loss of follow‐up: none.

Primary outcomes:

Range of motion (ROM) of the PIP joint.

Grip strength.

Pain: mean score assessed using a scale from 1 to 6; 1 = no pain and 6 = extreme pain.

Time of surgery.

Postoperative complications (adverse events and neurovascular injury).

Costs of the surgical techniques.

Outcomes included in this review:

Resolution of trigger finger: considered as complete relief of symptoms (the authors did not evaluate it as a study primary outcome, but they published indirect data on resolution).

Pain: mean score assessed using a scale from 1 to 6; 1 = no pain and 6 = extreme pain.

Recurrence of triggering: the authors did not evaluate it as a study primary outcome, but they published indirect data on recurrence.

Adverse events:

• Inflammation.

Neurovascular injury.

Notes

• The authors did not report on functional status of the hand and patients' satisfaction.

• The authors reported inclusion criterion of participants aged 18 to 80 years, but they presented results of participants aged 39 to 88 years.

• Open surgical technique was more expensive. The cost difference (personnel costs excluded) was € 7 (7 euros).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Quasi‐randomised controlled trial.

Participants were randomised to either open or percutaneous surgery using their patient numbers. When their numbers started with an uneven number, they were treated percutaneously, but if their numbers started an even number, they were treated by open surgery.

Allocation concealment (selection bias)

High risk

Quasi‐randomised controlled trial.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

High risk

No protocol was published.

Functional status (as primary outcome using validated instruments to measure hand function) was not evaluated by the authors and pain was measured by non‐validated instrument.

Other bias

Low risk

There was no baseline imbalance, and no risk of bias was associated with care providers or differences in rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: between February 1993 and October 1994.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: Rotterdam, the Netherlands.

Number of participants assigned:

96 participants (100 trigger digits); 54 percutaneous surgery and 46 open surgery.

Number of participants assessed:

96 participants (100 trigger digits); 54 percutaneous surgery and 46 open surgery.

Inclusion criteria:

Patients had to be older than 18 years and have symptoms of a trigger digit for at least 1 month.

Exclusion criteria:

• Evidence of inflammation at the time of presentation (to exclude active rheumatoid arthritis or other connective tissue disease).

• Previous surgery on the affected digit.

Age:

• Percutaneous surgery (mean): 62 years (24 to 88).

• Open surgery (mean): 60 years (24 to 81).

Gender:

• Percutaneous surgery: male/female = 1/1.5.

• Open surgery: male/female = 1/2.3.

Side: not reported.

Digits: although the authors assessed the results of 100 fingers, they reported data on 99 fingers.

• Percutaneous surgery: 23 thumb, 2 index, 17 long, 7 ring and 5 little.

• Open surgery: 17 thumb, 4 index, 12 long, 12 ring and 0 little.

Classification of injury: not reported.

Interventions

Timing of intervention:

Mean duration of symptoms:

• Percutaneous surgery: 6 months (1 to 24).

• Open surgery: 12 months (1 to 144).

Duration of treatment:

• Percutaneous surgery: 7 minutes.

• Open surgery: 11 minutes.

Type of intervention:

Percutaneous surgery:

• The affected digit was placed in extension. At the level of the A1 pulley an 18‐gauge hypodermic needle was inserted into the flexor sheath with the opening directed distally. After ascertaining that the flexor tendon was clear (active flexion shows no movement of the needle), the needle was turned 90° to direct the bevelled edges longitudinally. By moving the needle, using the skin as a pivot point, the flexor sheath and the A1 pulley were divided, resulting in a typical grating sound. When necessary, a second insertion of the needle was made to obtain full release of the trigger digit. Any residual triggering was tested by active flexion of the affected digit. A compressive bandage was applied.

Open surgery:

• A transverse incision was placed in a skin crease at the level of the metacarpal head. The flexor sheath was opened longitudinally, incorporating the A1 pulley. Any residual triggering was tested by active flexion of the affected digit. The skin was closed with a 4.0 nylon suture. A compressive bandage was applied.

Rehabilitation:

• Exercises were started immediately in both groups.

• The authors did not describe which protocol was used, nor for how long it was used.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up:

Follow‐up was 12 weeks.

Patients were examined 10 days, 6 weeks, and 12 weeks after surgery.

Loss of follow‐up: not reported.

Primary outcomes:

Mean duration of surgery (minutes).

Mean duration of postoperative pain (days).

Recovery of motor function (days).

Return to work (days).

Success rate (resolution of trigger finger): considered as the cessation of triggering, with no recurrence during follow‐up (3 months).

Complications.

Outcomes included in this review:

Resolution of trigger finger: considered as the cessation of triggering, with no recurrence during follow‐up (3 months).

Pain: reported the average postoperative pain duration in days.

Recurrence of triggering: was reported in the study although it was not considered as primary outcome by the authors; the authors did not clearly define what they considered recurrence.

Adverse event:

• Oedema.

• Hematoma and adherence.

Neurovascular injury.

Notes

• One adverse event that occurred in the percutaneous surgery group was not clearly specified by the author.

• We tried unsuccessfully to contact the authors in order to obtain further information to clearly specify what this adverse event was.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Low risk

Sealed envelopes were used.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

High risk

No protocol was published.

Pain and functional status of the hand (as primary outcomes, measured by validated instruments) were not evaluated by the authors.

Other bias

Low risk

There was no baseline imbalance, and no risk of bias was associated with care providers or differences in rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: January 2012 to May 2015.

Protocol was published before recruitment of patients: yes.

Details of trial registration: ClinicalTrials.gov, www.clinicaltrials.gov NCT01486420

Funding sources: the authors have declared no conflicts of interest.

Participants

Place of study: Center for Planned Surgery, Regional Hospital Silkeborg, Silkeborg, Denmark.

Number of participants assigned: 165 participants (165 fingers); 84 open surgery and 81 steroid injection.

Number of participants assessed: 153 participants (153 fingers); 76 open surgery and 77 steroid injection.

Inclusion criteria:

• Patients were ≥ 18 years old.

• Patients with trigger finger classified as Quinnell Type IIb (history of uneven movement with or without pain or discomfort) or greater.

Exclusion criteria:

Patients were excluded if they had insulin‐dependent diabetes mellitus, rheumatoid arthritis, amyloidosis, mucopolysaccharidosis, previous treatment of trigger finger in the included digit, Dupuytren disease affecting the included digit, or medical contraindications to corticosteroid injection.

Age:

• Open surgery (mean): 60 years.

• Steroid injection (mean): 60 years.

Gender:

• Open surgery: 27 male and 57 female.

• Steroid injection: 27 male and 54 female.

Side: 97 right and 68 left.

Digits: the allocated digit was the thumb in 39%, index in 5%, middle in 25%, ring in 25%, and little in 6%.

Classification of injury:

The trigger finger was graded according to Quinnell' classification modified by adding a history of uneven movement with pain or discomfort at the A1 pulley (graded I to V) (Quinnell 1980).

Interventions

Timing of intervention:

• Open surgery (mean): 4 months.

• Steroid injection (mean): 5 months.

Duration of treatment: not reported.

Type of intervention:

• Open surgery was performed under local anesthesia with a tourniquet placed at the upper arm. At the level of the A1 pulley, an incision was made and blunt dissection was done down to the A1 pulley. A small, round‐tipped dissection scissor was used to split the A1 pulley. After securing free movement of the flexor tendons, the skin was closed with nonabsorbable sutures, a light bandage was applied, and the tourniquet was released. As standard procedure, digital nerves were not explored in any digits.

• Ultrasound‐guided corticosteroid injection with a 23‐gauge needle was performed with a linear transducer placed in an axial direction to identify the flexor tendons. All patients were injected with a corticosteroid solution containing 1 ml triamcinolone acetonide, 40 mg/ml (Kenalog; Bristol‐Myers Squibb AB, New York) and 1ml Llidocaine, 10mg/ml. After intra‐sheath placement of 1 ml of the corticosteroid solution, the needle was withdrawn just superficially to the A1 pulley, and the last 1 ml was injected outside the sheath in close proximity to the A1 pulley.

Rehabilitation: not reported.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up: 12 months.

The patients were prospectively assessed after 3 and 12 months.

Loss of follow‐up: 1 patient in both groups were lost to follow‐up at 3 months, and 1 participant in open surgery group and 2 participants in steroid injection group were lost to follow‐up at 12 months and were excluded.

Six participants in open surgery group and one participant in steroid injection group did not receive allocated intervention.

Primary outcomes:

Resolution of trigger finger (cure): the authors considered normal movement with or without pain or discomfort after 12 months of follow‐up.

Secondary outcomes:

Topical pain: defined as patient‐reported pain when pressure was applied on the palmar side of the hand at the level of the A1 pulley, assessed by a numerical rating scale from 1 to 10 (1 = no pain, and 10 = worst imaginable pain).

Complications (adverse events):

• Infection.

• Neuropathy.

• Bowstringing.

• Flare (defined as the worsening of symptoms in a short time after the injection).

• Failure (reported by patients at the 1‐year follow‐up interview).

• Fat necrosis at the injection site.

• Tendon rupture.

Outcomes included in this review:

Resolution of trigger finger.

Pain: assessed by a numerical rating scale from 1 to 10 (1 = no pain, and 10 = worst imaginable pain).

Recurrence of triggering.

Adverse event:

• Infection.

• Bowstringing.

• Flare (defined as the worsening of symptoms in a short time after the injection).

• Fat necrosis at the injection site.

• Tendon rupture

Neurovascular injury.

Notes

• If patients presented with more than one affected finger, a single finger was arbitrarily chosen for inclusion based on the patient’s request.

• Outcomes were assessed at 3 months by personnel in the outpatient clinic, and at 12 months in a telephone interview by the first author.

• The authors did not report the mean and standard deviations on pain score. They reported only the median and interquartile range (IQR).

• We tried unsuccessfully to contact the authors to obtain further information on mean and standard deviations for pain. In the analyses of results we using the median as an approximate for the mean, and we calculate an approximate standard deviations (SD) based on IQR (using the formula: SD = IQR/1.35).

• In the open surgery group eight allocated participants did not complete the follow‐up (six patients were allocated but they did not receive the treatment, and two patients were operated but they were lost to follow‐up at 12 months). In the steroid injection group four allocated participants did not complete the follow‐up (one patient was allocated but he did not receive the treatment, and three patients were operated but they were lost to follow‐up at 12 months).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

The schedule for randomization was generated by the randomization software Research Randomizer (http://www.randomizer.org). The authors used a block randomization with blocks of 5 patients.

Allocation concealment (selection bias)

Low risk

Opaque sealed envelopes were used.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Missing outcome data were balanced in numbers across intervention groups; eight participants 8/84 (9%) in open surgery group and four participants 4/81 (5%) in steroid injection group were allocated to treatment, but they did not complete to follow up of 12 months and they were excluded; however they were reported.

Selective reporting (reporting bias)

Low risk

The study protocol was previously published. No changes in methods were applied after the trial commenced.

Other bias

Unclear risk

The groups were similar at baseline except for lower alcohol consumption in the open surgery group.The authors did not describe about differences in care providers or rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: between January 2013 and February 2014.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: the authors have declared no conflicts of interest.

Participants

Place of study: Nijmegen, the Netherlands.

Number of participants assigned:

30 participants (32 trigger digits);

Open surgery by transversal incision of the skin in the distal palmar crease: 11 trigger fingers.

Open surgery by transversal incision of the skin about 2–3 mm distally from distal palmar crease: 10 trigger fingers.

Open surgery by longitudinal incision of the skin at the level of the A1‐pulley without crossing the distal palmar crease proximal: 11 trigger fingers.

Number of participants assessed:

30 participants (32 trigger digits).

Open surgery by transversal incision of the skin in the distal palmar crease: 11 trigger fingers.

Open surgery by transversal incision of the skin about 2–3 mm distally from distal palmar crease: 10 trigger fingers.

Open surgery by longitudinal incision of the skin at the level of the A1‐pulley without crossing the distal palmar crease proximal: 11 trigger fingers.

Inclusion criteria:

• Patients had to be 18 years or older.

• Diagnosis of at least grade II trigger finger according to the Quinnell classification (Quinnell 1980).

• Duration of symptoms for at least 3 months.

• Absence of surgical treatment of the affected finger.

Exclusion criteria:

• Participants with trigger thumbs were excluded.

• More than one finger affected in one hand.

• History of severe hand trauma.

Age (mean): 61.77 years.

Gender: 13 male and 17 female.

Side:

• Dominant hand: 14.

• Non‐dominant hand: 15.

• Unknown: 3.

Digits: 4 index, 17 long, 8 ring and 3 little finger.

Classification of injury: The trigger finger was graded according to Quinnell classification (Quinnell 1980).

Interventions

Timing of intervention: at least 3 months.

Duration of treatment: not reported.

Type of intervention:

All participants were submitted to open surgery, and was randomised to one of three kinds of skin incision:

• Transversal in the distal palmar crease.

• Transversal 2‐3 mm distally from distal palmar crease.

• Longitudinally at the level of the A1‐pulley without crossing the distal palmar crease proximal.

All surgeries were performed under local anesthesia. Tourniquet was placed at the
Forearm at 250 mmHg and a incision pattern was carried out over a defined length of 15 mm. The A1‐pulley was identified and fully opened by a longitudinal incision over the pulley. Approximately 2‐3 mm in width of the A1‐pulley were resected to reduce the risk for recurrence. The skin was then closed with Prolene 4‐0.

Rehabilitation:

• Directly after surgery, all patients were instructed to use the hand without any specific limitations.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up:

Follow‐up was 12 months.

Patients were examined 1, 3 and 12 months after surgery.

Loss of follow‐up: not reported.

Primary outcomes:

Functional status of the hand: was used DASH score.

Scar volume: was measured using an high‐resolution ultrasound.

Outcomes included in this review:

Functional status of the hand: was used DASH score.

Notes

• The authors did not report standard deviations on DASH score or data about how many participants evolved with hypertrophic scar or keloid, but they reported the mean standard error on DASH score.

• We tried unsuccessfully to contact the authors to obtain further information on standard deviations for functional status of the hand (DASH‐score), and about how many participants presented hypertrophic scar or keloid in yours hands. So we used the mean standard error in the forest plots calculator to get the standard deviations in analyses about DASH score.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

Unclear risk

The authors did not assess or report any objective outcome.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Not reported.

Selective reporting (reporting bias)

High risk

No protocol was published.

Outcomes of interest in the review (resolution of trigger finger and pain) were not reported.

Other bias

High risk

There was baseline imbalance.

The authors reported there was a significant difference in baseline DASH scores between groups.

Methods

Study design: randomised controlled trial.

Duration of the study: October 1998 to December 2001.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: Bangkok, Thailand.

Number of participants assigned:

• 115 participants (127 triggering thumbs).

• Percutaneous surgery with steroid injection: 66 thumbs in 60 participants.

• Steroid injection alone: 61 thumbs in 55 participants.

Number of participants assessed:

• 125 triggering thumbs in 113 participants.

• Percutaneous release with steroid injection: 65 thumbs in 59 participants.

• Steroid injection alone: 60 thumbs in 54 participants.

Inclusion criteria:

Idiopathic adult trigger thumb grade II (actively correctable) III (passively correctable) or IV (fixed in flexion) according to the Quinnell classification (Quinnell 1980).

Exclusion criteria:

• Participants with history of trauma.

• Participants with carpal tunnel syndrome were excluded.

Age:

• Percutaneous release with steroid injection (mean): 52 years (31 to 68).

• Steroid injection alone (mean): 53 years (31 to 76).

Gender:

• Percutaneous release with steroid injection: 61 female and 4 male.

• Steroid injection alone: 55 female and 5 male.

Side:

Percutaneous release with steroid injection:

• Thumb affected (R/L)
  : 31/34.

• Hand dominance (R/L): 57/8.

Steroid injection alone:

• Thumb affected (R/L): 28/32.

• 
  Hand dominance (R/L): 55/5.

Digits: 127 thumbs.

Classification of injury:

The trigger thumb was graded according to Quinnell classification (graded 0 to IV) (Quinnell 1980).

Interventions

Timing of intervention:

Percutaneous release with steroid injection (mean): 4 months (1 to 36).

Steroid injection alone (mean): 4 months (1 to 20).

Duration of treatment: not reported.

Type of intervention:

Percutaneous release with steroid injection:

• The release was done with an 18‐gauge needle inserted at a point 1 mm to 2 mm distal to metacarpophalangeal joint crease, keeping the thumb in hyperextension position. After release the flexor tendon sheath was injected with 1 ml triamcinolone acetonide, 10 mg/ml.

Steroid injection alone:

• Participants were treated simply by injection of 1 ml triamcinolone acetonide and 1 ml of 1% lidocaine.

Rehabilitation: not reported.

Any co‐interventions:

• All participants in both groups received 20 paracetamol tablets for home medication. They were told to take the medicine only if they felt pain.

• The mean paracetamol requirement in the first 2 weeks was 4 tablets in percutaneous release with steroid injection and 3 tablets in steroid injection.

Outcomes

Length of follow‐up: range 23 months (6 to 42 months).

Participants were evaluated at 2 and 6 weeks, and 6 or more months.

Loss of follow‐up:

1 participant (1 thumb) was lost to follow‐up at 6 months in both groups.

Primary outcomes:

Resolution of trigger finger: "satisfactory" was considered to be participants who progressed with pain score lower than or equal to 1 (VAS scale) and cessation of triggering.

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale) and paracetamol requirement in the first 2 weeks.

Outcomes included in this review:

Resolution of trigger finger

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Adverse events: were reported in the study although not considered as primary outcomes by the author.

Superficial infection (cellulitis).

Partial loss of movement.

Neurovascular injury: was reported in the study although not considered as primary outcome by the author.

Notes

• The authors did not report standard deviations on VAS score.

• We tried unsuccessfully to contact the authors to obtain further information on standard deviations for pain (VAS‐score).

• The follow‐up data were collected by clinical examination (78 participants) or by telephone interview (35 participants) between 6 and 42 (mean, 23) months.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Missing outcome data were balanced in numbers across intervention groups. 1 thumb was lost in both groups at 6 months: in percutaneous surgery plus steroid injection group 1/66 (1.5%); and in steroid injection 1/61 (1.6%).

Selective reporting (reporting bias)

High risk

No protocol was published.

The authors did not report standard deviations on VAS score, and the functional status of the hand (using validated instruments) was not evaluated.

Other bias

Unclear risk

There was no baseline imbalance and no risk of bias was associated with care providers, but the author did not report about rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: not reported.

Protocol was published before recruitment of patients: yes.

Details of trial registration: clinicaltrials.gov/ct2/show/study/NCT02830672.

Funding sources: the authors have declared no conflicts of interest.

Participants

Place of study: National and Kapodistrian University of Athens, Greece.

Number of participants assigned: 32 participants (32 fingers); 16 ultrasound‐guided percutaneous surgery and 16 open surgery.

Number of participants assessed: 32 participants (32 fingers); 16 ultrasound‐guided percutaneous surgery and 16 open surgery.

Inclusion criteria:

• Participants affected by trigger finger or trigger thumb clinically and ultrasonographically examined.

• Failure non‐surgical treatment for 3 months.

• Grade II, III or IV trigger digit as classified by Froimson 1993.

Exclusion criteria:

• Individuals under 18 years old.

• Patients who were treated with a previous operation or a corticosteroid injection.

• Individuals with inflammatory arthritis, tumor or autoimmune disease.

• Patients with multiple trigger fingers.

Age: mean age of 32 patients: 45.5 years old.

• Percutaneous surgery (mean/range): not reported.

• Open surgery (mean/range): not reported.

Gender: 12 male; 20 female.

• Percutaneous surgery (mean/range): not reported.

• Open surgery (mean/range): not reported.

Side: not reported.

Digits: not reported.

Classification of injury:

Trigger fingers were graded according to Froimson’s modification of Quinnell’s classification (graded I to IV) (Froimson 1993).

Interventions

Timing of intervention: at least 3 months.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous surgery: under continuous sonographic imaging of the digit, an ophthalmic corneal/scleral V‐Lance knife (Alcon, Novartis company) was inserted percutaneously, over flexor tendons proximally to the A1 pulley and towards their longitudinal axis. Then, the knife was advanced distally, just below A1 pulley and pressed palmary so as to loosen the thicken pulley. Thus, after having withdrawn the V‐Lance knife, a thin hook with a long neck was introduced under the ‐ now extended ‐ A1 pulley. The hook penetrated the annular ligamentous structure facing palmary in order to protect the flexor tendons and subsequently removed proximally (in a steady quick move) carrying along and dissecting the A1 pulley. Intraoperatively and right after the performed dissection, each patient was clinically and sonographically evaluated for the achieved resolution of the triggering.

• Open surgery: the section of the A1 pulley was done through a 10‐15 mm skin incision.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up:

• Follow‐up was 12 weeks.

• Participants were evaluated at 2, 4 and 12 weeks.

Loss of follow‐up: none lost to follow‐up.

Primary outcomes:

Resolution of triggering was expressed as the “success rate” per digit.

The time for taking postoperative pain killers.

QuickDASH score.

Range of motion recovery.

Return to normal activities (including work).
Complications (adverse events).

Cosmetic results.

Outcomes included in this review:

Resolution of triggering was expressed as the “success rate” per digit.

Pain: postoperative pain duration (measured by mean time in days for taking postoperative pain killers).

Functional status of the hand: QuickDASH score.

Adverse events measured by:

• Infections.

• Partial loss of movement.

Notes

• The authors did not report standard deviations on QuickDASH score.

• We tried unsuccessfully to contact the authors to obtain further information on standard deviations for functional status of the hand (QuickDASH‐score).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Low risk

Closed envelopes were used.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

Low risk

The self‐reported subjective outcomes assessors were blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

Unclear risk

Although results were evaluated and explained according to protocol published in July 6, 2016 and the study was published in February 18, 2017, the paper was received in the World Journal of Orthopedics' editorial in July 4, 2016 before date of the protocol publication.

Although the protocol published by authors estimated the enrollment of 60 patients in the study, the sample of the paper was only 32 participants.

Other bias

Unclear risk

The authors did not report data about baseline balance, and they did not clearly report about care providers and rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: October 2005 to March 2006.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none known.

Participants

Place of study: University of Milan, Italy.

Number of participants assigned:

200 participants (231 fingers);

Endoscopic surgery: 114 trigger fingers in 100 participants.

Open surgery: 117 trigger fingers in 100 participants.

Number of participants assessed:

200 participants (231 fingers).

Endoscopic surgery: 114 trigger fingers in 100 participants.

Open surgery: 117 trigger fingers in 100 participants.

Inclusion criteria:

Participants affected by trigger finger, despite other concomitant disease.

Exclusion criteria: not reported.

Age:

• Endoscopic surgery (mean): 56 years (23 to 82).

• Open surgery (mean): 61 years (33 to 79).

Gender:

• Endoscopic surgery: 33 male and 67 female.

• Open surgery: 27 male and 73 female.

Side:

The authors reported data incompletely; in the endoscopic surgery group they reported the affected side in relation to the fingers, while in the open surgery group they reported the affected side in relation to the participants.

Endoscopic surgery: 71 (fingers) right hand and 43 (fingers) left hand.

Open surgery: 69 (participants) right hand and 31 (participants) left hand.

Digits:

• Endoscopic surgery: 12 thumb, 23 index, 37 long, 34 ring and 8 little.

• Open surgery: 13 thumb, 25 index, 31 long, 38 ring and 10 little.

Classification of injury: not reported.

Interventions

Timing of intervention: not reported

Duration of treatment:

Endoscopic surgery (mean): 4 min and 30 seconds (range: 2 to 9 min).

Open surgery (mean): 5 min (range: 2 to 7 min).

Type of intervention:

Endoscopic surgery:

• Under local anaesthesia, with 2 sites of transverse incision that were distally at the level of the digital‐palmar crease of the finger and proximally at the level of the palmar crease, corresponding to the metacarpophalangeal joint of the finger. A 2.7‐mm diameter endoscope was introduced through the proximal incision while a retrograde knife was introduced from the distal incision. Skin suture was done with steri‐strips and a moderate compressing dressing was applied.

Open surgery:

• Under local anaesthesia, with longitudinal incision of 1 cm length made on the volar aspect of the hand, in the palmar crease overlying the metacarpophalangeal joint of the involved digit. The incision was closed in a single layer with 4.0 Vycril and a moderate compressive dressing was applied.

Rehabilitation:

• Post‐operative rehabilitation protocol consisted of 2 therapy sessions: 1 on the day after surgery, in which tendon gliding exercises and oedema control were taught, and the other after removal of the dressing, after 7 days for the endoscopic surgery group and after 12 days for the open surgery group. The participants were briefed on how to treat the operative site, with massage of the scar and tendon gliding exercises. Both groups used a dynamic extension splint for the proximal interphalangeal and metacarpophalangeal joints at night for 1 month.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up: 90 days.

Follow‐up at 7, 30 and 90 days. The first follow‐up was excluded because the dressing still present in open surgery group did not allow a proper evaluation.

Loss of follow‐up: none.

Primary outcome:

Resolution of trigger finger: considered as the disappearance of triggering after the procedure.

Pain: reported the average postoperative pain duration in days.

Adverse event:

• Infection.

• Dysesthesia.

Neurovascular injury.

Outcomes included in this review:

Resolution of trigger finger.

Pain: reported the average postoperative pain duration in days.

Adverse event:

• Infection.

• Dysesthesia.

Neurovascular injury.

Notes

• Included were participants with carpal tunnel syndrome (28 cases in open surgery group and 12 in endoscopic surgery group), de Quervain’s syndrome (4 cases in open surgery group and 2 in endoscopic surgery group) and carpometacarpal joint arthritis (6 cases in open surgery group).

• The authors treated trigger finger and others associated pathologies in the same surgical procedure.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method of generating the random sequence was not reported.

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

High risk

No protocol was published.

Pain and functional status of the hand (as primary outcomes, measured by validated instruments) were not evaluated by the authors.

Other bias

High risk

There was baseline imbalance.

Several participants in both groups had associated pathologies in hand and underwent surgery.
In the open surgery group 38% (38/100 participants) had associated pathologies in hand: 28 had carpal tunnel syndrome, 4 had de Quervain's syndrome and 6 rizartroses. In the endoscopic surgery group 14% (14/100 participants) had associated pathologies in hand: 12 had carpal tunnel syndrome and 2 had de Quervain's syndrome.

Methods

Study design: randomised controlled trial.

Duration of the study: November 2002 to March 2007.

Protocol was published before recruitment of patients: yes.

Details of trial registration: Current Controlled Trials, www.controlled‐trials.com/ ISRCTN19255926.

Funding sources: the authors have declared no conflicts of interest.

Participants

Place of study: Federal University of São Paulo, Brazil.

Number of participants assigned:

137 participants (150 fingers); 45 percutaneous surgery, 56 open surgery and 49 steroid injection.

Number of participants assessed:

137 participants (150 fingers); 45 percutaneous surgery, 56 open surgery and 49 steroid injection.

Inclusion criteria:

• Patients were ≥ 15 years old.

• Patients with trigger finger who had not undergone previous treatment of any type and were classified as Quinnell Types II‒IV.

Exclusion criteria:

• Individuals with type I trigger fingers.

• Congenital trigger fingers.

• Secondary to the partial lesion to the tendon.

Age:

• Percutaneous surgery (mean): 54.40 years.

• Open surgery (mean): 53.4 years.

• Steroid injection (mean): 55.29 years.

Gender:

• Percutaneous surgery: 5 male and 40 female.

• Open surgery: 10 male and 46 female.

• Steroid injection: 3 male and 46 female.

Side: not reported.

Digits: 31 thumb, 4 index, 77 long, 35 ring and 3 little.

Classification of injury:

The trigger finger was graded according to Quinnell classification (graded 0 to IV) (Quinnell 1980).

Interventions

Timing of intervention:

• Percutaneous surgery (mean): 14.96 months.

• Open surgery (mean): 10.5 months.

• Steroid injection (mean): 11.80 months.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous surgery consisted of release of the A1 pulley with a 40 × 12 needle, using longitudinal movements, in the direction of the axis of the flexor tendon, and this release was introduced at the site corresponding to the A1 pulley.

• Open surgery consisted of an incision of 2 cm in the skin transverse to the axis of the finger at the palmar skin fold, followed by subcutaneous dissection and longitudinal opening of the A1 pulley.

• Steroid injection consisted of an injection of 2 ml of methylprednisolone acetate 40 mg/ml at the site corresponding to the A1 pulley, attempting to inject the solution within the osteofibrous canal.

Rehabilitation: not reported.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up: 6 months.

The patients were prospectively assessed after 1, 2 weeks, 1, 2, 4 and 6 months.

Loss of follow‐up: none.

Primary outcomes:

Resolution of trigger finger (cure): the authors considered the remission of symptoms with the cessation of blockage with no recurrence within 6 months.

Recurrence of triggering: the authors defined recurrence (relapse) as the return of finger locking within 6 months of follow‐up.

Secondary outcomes:

Topical pain.

Articular pain.

Total active motion (TAM) of the fingers.

Complications (adverse events):

• Infection.

• Tendon injury.

Neurovascular injury.

Outcomes included in this review:

Resolution of trigger finger.

Pain: assessed through presence or not presence of the pain (topical or articular) in the hand.

Recurrence of triggering.

Adverse event:

• Infection.

• Tendon injury.

Neurovascular injury.

Notes

• For the injection group, all participants received a single injection initially, then a second injection was performed if there was no resolution after the first infiltration. Those that received a second injection were followed for 6 months starting from the second intervention.

• In cases in which the patient presented with 2 trigger fingers, each finger received its own order number regardless of whether the fingers were on the same or different hands; 11 patients participated in the study on 2 occasions, and 1 participated on 3 occasions.

• After contacting the main author of the study, we were informed about the exact number of thumbs, index, long, ring and little fingers included in each comparison group of the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

The randomisation was done by means of a 6‐sided die with 2 sides representing 1 of the 3 treatments.
The draw was conducted before the study by a person independent of the research.

Allocation concealment (selection bias)

Low risk

The result of each draw was placed in an opaque envelope, which was then sealed; envelopes were numbered from 1 to 150. None of the project participants had prior access to the envelope contents.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

The authors did not report missing data.
All participants received treatment and were followed.

Selective reporting (reporting bias)

Unclear risk

Although results were evaluated and explained according to previous protocol published in October 2010, the study started in November 2002 and finished in March 2007.

Other bias

Unclear risk

There was no baseline imbalance or differences in care providers, but the authors did not describe rehabilitation.

Methods

Study design: quasi‐randomised controlled trial.

Duration of the study: from January 2005 to June 2005.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not registered.

Funding sources: none known.

Participants

Place of study: Penang General Hospital, Malaysia.

Number of participants assigned: 26 participants (26 fingers); 14 percutaneous surgery and 12 steroid injection.

Number of participants assessed: 26 participants (26 fingers); 14 percutaneous surgery and 12 steroid injection.

Inclusion criteria:

• Participants who presented with grade II or III trigger digit as classified by Quinnell.

Exclusion criteria:

• Participants with trigger thumbs were excluded.

Age:

• Total of participants (mean): 57 years.

Gender:

• Total of participants: 9 male; 17 female.

Side:

• Total of participants: 14 dominant hand and 12 non‐dominant hand.

Digits:

• Total of participants: 4 index, 11 long and 11 ring. There was no involvement of the little finger.

Classification of injury:

The trigger finger was graded according to Quinnell classification (graded 0 to IV) (Quinnell 1980).

Interventions

Timing of intervention: not reported.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous surgery: after 1m of lidocaine 1% was infiltrated around A1 pulley, a 20‐gauge needle was inserted with the sharp bevel parallel to the tendon, in position correlating with the middle of the A1 pulley. The needle was stroked longitudinally, proximally and distally to transect the A1 pulley.

• Steroid injection: after 1m of lidocaine 1% was infiltrated around A1 pulley, was injected 1 mL of triamcinolone mixed with 1 mL of 1% lidocaine into the tendon sheath and around the nodule.

Rehabilitation: the authors did not clearly describe if all study participants did physiotherapy. They reported that in percutaneous surgery group 2 participants developed stiffness of digit which responded to aggressive physiotherapy.

Any co‐interventions: analgesia was given for 3 days in steroid injection group.

Outcomes

Length of follow‐up:

• Follow‐up was 1 year.

• Participants were evaluated weekly for 1 month and 3‐monthly for a year.

Loss of follow‐up: not reported.

Primary outcomes:

Pain.

Patient satisfaction.

Recurrence of triggering: the authors did not clearly define what they considered recurrence.

Adverse events:

• Partial loss of movement.

• Dysaesthesia.

• Pulley or tendon injury.

Outcomes included in this review:

Pain.

Patient satisfaction.

Recurrence of triggering

Adverse events:

• Partial loss of movement.

• Dysaesthesia.

• Pulley or tendon injury.

Notes

• Pain and patient satisfaction were described incompletely (no numerical data was reported), and we were unable to include these data in the results.

• We tried unsuccessfully to contact the authors to obtain further information on pain.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Quasi‐randomised controlled trial.

Patients were randomised to either steroid injection or percutaneous surgery using their birth year. Those with even numbers were allocated to the steroid group and uneven numbers to the percutaneous group.

Allocation concealment (selection bias)

High risk

Quasi‐randomised controlled trial.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

High risk

The self‐reported subjective outcomes assessors were not blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

High risk

The objective outcome assessors were not blinded.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Not reported.

Selective reporting (reporting bias)

High risk

No protocol was published.

Outcomes of interest in the review (resolution of trigger finger and functional status of the hand) were not reported. Pain was reported incompletely, without numerical values.

Other bias

Unclear risk

The authors did not report data about baseline balance, and they did not clearly report about care providers and rehabilitation.

Methods

Study design: randomised controlled trial.

Duration of the study: January 2008 to May 2009.

Protocol was published before recruitment of patients: not reported.

Details of trial registration: not reported.

Funding sources: none declared.

Participants

Place of study: Pomeranian Medical University in Szczecin, Poland.

Number of participants assigned: 115 participants.

Number of participants assessed: 95 participants (105 digits); 43 participants (46 digits) percutaneous surgery and 52 participants (59 digits) steroid injection.

Inclusion criteria: participants with trigger digits.

Exclusion criteria: not reported.

Age:

• Percutaneous surgery (mean): 55 years.

• Steroid injection (mean): 58 years.

Gender:

• Percutaneous surgery: 27 female and 16 male.

• Steroid injection: 40 female and 12 male.

Side:

• Percutaneous surgery: 24 left and 19 right.

• Steroid injection: 34 left and 18 right.

Digits:

• Total of digits assessed: 39 thumb, 1 index, 22 long, 35 ring and 8 little.

Classification of injury:

Trigger fingers were graded according to Froimson’s modification of Quinnell’s classification (graded I to IV) (Froimson 1993).

Interventions

Timing of intervention:

• Percutaneous surgery (mean): 5 months.

• Steroid injection (mean): 6 months.

Duration of treatment: not reported.

Type of intervention:

• Percutaneous A1 pulley release was performed in the outpatient clinic, using a 19 gauge hypodermic needle, after preparation of the skin and injection of 1 ml 2% plain lidocaine.

• The steroid injection of 1 ml of betamethasone into the flexor tendon sheath was also performed in the outpatient clinic.

Rehabilitation: not reported.

Any co‐interventions: not reported.

Outcomes

Length of follow‐up:

6 months (at 1 and 6 months).

Loss of follow‐up:

20 participants (12 in the percutaneous surgery group and 8 in the steroid injection group).

Primary outcomes:

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Active range of motion (AROM) of the affected digit.

Total grip strength: expressed as a proportion of the strength of the contralateral, healthy hand.

Recurrence of triggering: considered as the return to the baseline grade, after a period of total or partial improvement of the trigger finger.

Adverse event:

• Partial loss of movement.

• Neurovascular injury.

Outcomes included in this review:

Pain: measured by Visual Analog Scale (VAS: 0 to 10 scale).

Recurrence of triggering.

Adverse event:

• Partial loss of movement.

Neurovascular injury.

Notes

• The authors reported pain, adverse events and neurovascular injury incompletely. They did not report standard deviations on VAS score, and they reported that infection, algodystrophic syndrome and neurovascular injury did not occur in the steroid injection group, but did not mention whether these adverse events occurred in the percutaneous surgery group.

• Twenty of 115 patients (17%) who were recruited did not attend follow‐up (12 in the group treated operatively and 8 in the group treated by injection).

• We tried unsuccessfully to contact the authors to obtain further information on standard deviations for pain (VAS‐score), and missing data on adverse events and neurovascular injury; asked also for further information on total patients' fingers lost to follow‐up. In analysis we used the same standard deviation reported by Chao 2009 for pain (VAS‐score), as foreseen in our protocol Ventin 2014.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Patients were randomly assigned to the groups by drawing slips of paper marked 1 (percutaneous release) or 2 (steroid injection) from a sealed envelope in the presence of a witness.

Allocation concealment (selection bias)

Low risk

Sealed envelopes were used.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Participants and personnel were not blinded.

Blinding of outcome assessment (detection bias) ‐ subjective outcomes (e.g. resolution, recurrence, pain, function, satisfaction)
All outcomes

Low risk

The self‐reported subjective outcomes assessors were blinded.

Blinding of outcome assessment (detection bias) ‐ objective outcomes (e.g. adverse events, neurovascular injury)
All outcomes

Low risk

The objective outcome assessors were blinded.

Incomplete outcome data (attrition bias)
All outcomes

High risk

20 of 115 patients (17%) who were recruited did not attend follow‐up and were excluded (12 participants in the percutaneous surgery group and 8 in the steroid injection group). The authors did not describe how many fingers were affected in these 20 patients.

Selective reporting (reporting bias)

High risk

No protocol was published.

Outcomes of interest in the review (resolution of trigger finger and functional status of the hand) were not reported. Pain was reported incompletely, without any measure of variance for the validated instruments (VAS).

Other bias

Unclear risk

There was no baseline imbalance and no risk of bias was associated with care providers, but the author did not report about rehabilitation.