Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison/ control (C)

Follow-up

Outcome measures and effect size

Comment

Aksoy, 2018

Type of study:

RCT

Setting and country:

Single centre, Turkey

Funding and conflicts of interest:

none

Inclusion criteria:

Patients with cervical radiculopathy confirmed through MRI, aged 18-65, neck pain on VAS ≥4

Exclusion criteria:

(1) previous surgical operation on the c-spine, (2) other systemic, neurological or psychiatric problems; (3) rheumatic and infectious disease; (4) current malignancy; (5) motor deficit in upper extremity; (6) previous treatment with cervical collar; (7) complaints >12 weeks

N total at baseline: 101

Intervention: 67 (34 group 1 and 33 group 2)

Control: 33

Important prognostic factors:

age:

I: 41 (G1), 40 (G2)

C: 46

Sex:

I: 23% M (G1), 46% M (G2)

C: 38% M

VAS score neck pain:

I: 8.26 (G1), 8.30 (G2)

C: 7.72

NDI score:

I: 66 (G1), 64 (G2)

C: 55

Groups comparable at baseline?

Yes

Intervention treatment:

Group 1 (G1): soft cervical collars (soft sponge) plus control treatment

Group 2 (G2): semi-rigid cervical collars (plastazote foam material) plus control treatment

• For week 1 and 2, wear collar 8 hours every day
• For week 3 and 4, reduce collar time every other day by one hour, quit collar at end of week 4.

Control treatment:

Home exercises comprising cervical isometric, cervical mobilisation, and shoulder protraction and retraction exercises for 6 weeks, twice a day, with 2x 10 repetitions at each session.

Advice to avoid holding the neck in prolonged flexion or extension during daily activities and to use a suitable pillow during sleep.

The use of NSAID was allowed when necessary.

Length of follow-up: 6 weeks

Loss-to-follow-up:

Intervention: 4 (G1) and 7 (G2) (10.9%)

Reasons: failed to complete, due to discomfort of collar (n = 7), no excuse given (n = 4)

Control: 5 (5%)

Reasons (describe): discontinuation with unknown reason

Incomplete outcome data: 14 (13.9%)

Same as above

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

Soft collar as add-on to exercise

-0.49 [-1.23 to 0.25]

Semi-rigid collar as add-on to exercise

0.26 [-0.59 to 1.11]

NDI (mean difference, 95%CI)

Soft collar as add-on to exercise

-5.5 [-9.9 to -1.1]

Semi-rigid collar as add-on to exercise

1.0 [-4.0 to 6.0]

Quality of Life (SF-36) (mean difference, 95%CI)

Soft collar as add-on to exercise

PCS: -3.80 [-6.54 to -1.06]

MCS: -0.60 [-3.03 to 1.83]

Semi-rigid collar as add-on to exercise

PCS: -1.60 [-4.62 to 1.42]

MCS: -2.50 [-5.40 to 0.40]

Kuijper, 2009

Type of study:

RCT

Setting and country: 3 hospitals, the Netherlands

Funding and conflicts of interest: non-commercial organisation for salary of research nurse, no conflicts of interest

Inclusion criteria:

Patients referred by GPs with <1 month symptoms and signs of cervical radiculopathy, aged 18-75 years, with arm pain VAS ≥40mm and radiation of pain distal to elbow. In addition, either provocation of arm pain by neck movements, or sensory changes in ≥1 dermatome, or diminished deep tendon reflexes in the arm, or muscle weakness had to be present.

Exclusion criteria:

(1) clinical signs of spinal cord compression, (2) previous treatment with physiotherapy or cervical collar

N total at baseline:205

Intervention: 139 (69 group 1 and 70 group 2)

Control: 66

Important prognostic factors²:

age ± SD:

I: 47.0 ± 9.1 (G1), 46.7 ± 10.9 (G2)

C: 47.7 ± 10.6

Sex:

I: 55% M (G1), 49% M (G2)

C: 48% M

VAS score neck pain

I: 57.4 ± 27.5 (G1), 61.7 ± 27.6 (G2)

C: 55.6 ± 31.0

NDI

I: 41.0 ± 17.6 (G1), 45.1 ± 17.4 (G2)

C: 39.8 ± 18.4

Groups comparable at baseline? Yes

Intervention treatment:

Group 1 (G1): semi-hard collar (one of six sizes with best fit)

• for week 1-3 wear collar during the day, and the advice to take as much rest as possible.
• For week 4-6, reduce collar time, quit collar at end of week 6

Group 2 (G2): Guided physiotherapy sessions consisting of hands off graded activity exercises to strengthen the superficial and deep neck muscles, twice a week for 6 weeks. In addition, physiotherapists educated the patients on home exercises, and practice those every day.

Written and oral reassurance about the usually benign course of the symptoms was given to patients in both G1 and G2.

Control procedure:

Continuation of daily activities without specific treatment other than written and oral reassurance about the usually benign course of the symptoms.

The use of painkillers was allowed when necessary.

Length of follow-up: 6 months

Loss-to-follow-up:

5 for primary outcome (6 weeks), 13 for entire follow-up (6 months)

Intervention: 6 (G1, 8.7%) and 2 (G2, 2.9%))

Reasons: not provided

Control: 5 (7.6%)

Reasons: not provided

Incomplete outcome data: 13 (6.3%)

Same as above

Outcome measures and effect size (95%CI):

Pain on VAS after 6 weeks (mean difference, 95%CI)

Collar compared to no intervention

-20.1 [-30.4 to -9.8]

Collar compared to exercise

-5.20 [-15.0 to 4.6]

Exercise compared to no intervention

-1.49 [-2.56 to -0.42]

NDI after 6 weeks (mean difference, 95%CI)

Collar compared to no intervention

-4.0 [-10.6 to 2.6]

Collar compared to exercise

-1.9 [-7.9 to 4.1]

Exercise compared to no intervention

-2.10 [-8.46 to 4.26]

Use of opiates after 6 weeks (relative risk, 95%CI)

Collar compared to no intervention

0.75 [0.39 to 1.44]

Collar compared to exercise

0.97 [0.49 to 1.94]

Exercise compared to no intervention

0.78 [0.41 to 1.48]

Working status after 6 weeks (relative risk, 95%CI)

Collar compared to no intervention

0.77 [0.48 to 1.25]

Collar compared to exercise

0.66 [0.42 to 1.03]

Exercise compared to no intervention

1.18 [0.78 to 1.77]

Patients who eventually underwent surgery: 5 in collar group, 3 in physiotherapy group, and 4 in control group.

2. Studies reporting on traction

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Bukhari, 2016

Type of study:

RCT

Setting and country: single hospital, Pakistan

Funding and conflicts of interest: no funding received, conflict of interest not disclosed.

Inclusion criteria:

Patients with evident radicular symptoms, cervical spine involvement, aged 20-70 years.

Exclusion criteria:

History of trauma, neck pain without radiculopathy

N total at baseline: 42

Intervention: 15 (mechanical traction)

Control: 21 (manual traction)

Important prognostic factors:

age: mean 45.8 years, not reported for both subgroups

Sex: 66% M (not reported for both subgroups)

NRS score:

I: 6.26 (± 1.20)

C: 6.80 (± 1.20)

NDI score:

I: 24.43 (± 8.64)

C: 21.92 (± 8.89)

Groups comparable at baseline? Unclear

Intervention treatment:

Mechanical traction, applied on patient in supine position, with 10 second pull and 5 second rest for 10 minutes, with a traction force of 10-15% of body weight.

In addition, segmental mobilisation of C3 to C7 by central posterior-anterior glide, 10 repetitions for 5 seconds was applied, and patients were advised to do a home exercise program with active range of motion, stretching, end isometric strengthening exercises 3 days a week for 6 weeks.

Control treatment:

Manual traction, applied on patient in supine position at 25 degree neck flexion, with 10 second pull and 5 second rest for 10 times.

In addition, segmental mobilisation of C3 to C7 by central posterior-anterior glide, 10 repetitions for 5 seconds was applied, and patients were advised to do a home exercise program with active range of motion, stretching, end isometric strengthening exercises 3 days a week for 6 weeks.

Length of follow-up: 6 weeks

Loss-to-follow-up: 6 patients (unclear which group, no reasons provided)

Incomplete outcome data: 6 (14.3%)

Same as above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-2.42 [-3.31 to -1.53]

NDI (mean difference, 95%CI)

-9.98 [-17.28 to -2.68]

Methodological weak study with single intervention session.

Ojoawo, 2019

Type of study:

RCT

Setting and country:

Academic hospital, Nigeria

Funding and conflicts of interest:

none

Inclusion criteria:

Neck pain that radiated distal from the elbow.

Exclusion criteria:

Main complaint of headache or facial pain, and having received manual therapy of the cervical region in the past 3 months.

N total at baseline: 75

Intervention: 50 (25 group 1 and 25 group 2)

Control: 25

Important prognostic factors:

Age (mean ± SD):

I: 51.4 ± 6.5 (G1), 55.7 ± 5.4 (G2)

C: 59.5 ± 2.6

Sex:

I: 56% M (G1), 60% M (G2)

C: 44% M

VAS score (mean ± SD):

I: 6.87 ± 1.0 (G1, 7.63 ± 2.30 (G2)

C: 7.00 ± 0.8

NDI score (mean ± SD)::

I: 42.1 ± 16.9 (G1), 58.7 ± 8.9 (G2)

C: 55.3 ± 11.3

Groups comparable at baseline? NDI score varies relatively much at baseline, yet in general, yes.

Intervention treatment:

Group 1 (G1): cervical traction “over the door” with a water bat loaded to 10% of patient’s body weight, for 15 minutes. This was administered twice a week for 6 weeks.

Group 2 (G2): transverse oscillatory pressure (TOP) administered by the therapist manually, to the patient lying on his/her belly, on the side of the location of the pain. Administered 3 times for 20 seconds, with 2 min rest in between. TOP was given twice a week for 6 weeks.

In addition, G1 and G2 patients performed active exercises: cervical spine retraction, rotation, extension, and side-bending stretching. Ice packs were applied to the cervical region for 7 minutes, and massage. Both 2 times per week for 6 weeks.

Control treatment:

Patients performed active exercises: cervical spine retraction, rotation, extension, and side-bending stretching. In addition, ice packs were applied to the cervical region for 7 minutes, and massage. Both 2 times per week for 6 weeks.

Length of follow-up: 6 weeks

Loss-to-follow-up: 3

Intervention: 1 (G1) and 0 (G2) (4%)

Reasons: discontinued intervention

Control: 2 (8%)

Reasons: discontinued intervention

Incomplete outcome data: 14 (13.9%)

Same as above

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

Cervical traction compared to control

-1.25 [-1.55 to -0.95]

TOP compared to control

-1.09 [-1.47 to -0.71]

Cervical traction compared to TOP

-0.16 [-0.54 to 0.22]

NDI (mean difference, 95%CI)

Cervical traction compared to control

-1.50 [-8.76 to 5.76]

TOP compared to control

-5.17 [-9.47 to -0.87]

Cervical traction compared to TOP

3.67 [-4.29 to 11.63]

Risk of selection bias.

Unclear whether ITT (with unclear imputation methods) or PP was used for analysis.

Fritz, 2014

Type of study:

RCT

Setting and country:

Several physical therapy offices, USA

Funding and conflicts of interest:

Non-commercial grant. Conflict of interests not disclosed.

Inclusion criteria:

(1) Chief complaint of neck pain, with symptoms extending distal to acromioclavicular joingt or causal to the superior border of the scapula, (2) age 18-70 years, (3) NDI score ≥10.

Exclusion criteria:

(1) history of surgery to the neck or thoracic spine, (2) motor vehicle accident in past 2 weeks, (3) red flags indicative of possible nonmusculoskeletal condition, (4) diagnosis of cervical spinal stenosis on MRI or CT, (5) evidence of cervical myelopathy or CNS involvement

N total at baseline: 86

Intervention: 58 (31 group 1 and 27 group 2)

Control: 28

Important prognostic factors:

age:

I: 48.1 ± 10.0 (G1), 47.6 ± 10.9 (G2)

C: 44.9 ± 11.3

Sex:

I: 58% M (G1), 44% M (G2)

C: 36% M

NRS score:

I: 3.8 ± 2.1 (G1), 4.5 ± 2.1 (G2)

C: 4.4 ± 2.0

NDI score:

I: 31 ± 15 (G1), 33 ± 14 (G2)

C: 35 ± 14

Groups comparable at baseline? Yes

Intervention treatment:

Group 1 (G1): Mechanical cervical traction, with angle of pull of 15 degrees. Intermittent traction was performed with 60 seconds pull force and 20 seconds relaxation, for 15 minutes.

Group 2 (G2): Cervical traction using an over-the-door traction device for 15 minutes, used at home daily on days of no physiotherapy sessions.

In addition, patients received 10 individual physiotherapy sessions for 4 weeks (3x/week in first 2 weeks, 2x/week in final 2 weeks), plus an active exercise program, with cervical strengthening and scapula strengthening exercises, to be performed daily on the days between therapy sessions.

Control treatment:

10 individual physiotherapy sessions for 4 weeks (3x/week in first 2 weeks, 2x/week in final 2 weeks), plus an active exercise program, with cervical strengthening and scapula strengthening exercises, to be performed daily on the days between therapy sessions.

Length of follow-up: 12 months

Loss-to-follow-up: 32

Intervention: 10 (G1) and 10 (G2) (34.5%)

Reasons: various (unclear flow-diagram)

Control: 12 (44.4%)

Reasons: Various (unclear flow diagram)

Incomplete outcome data: 32 (37.2%)

Same as above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

Cervical traction as add-on to exercise

-0.92 [-1.80 to -0.04]

NDI (mean difference, 95%CI)

Cervical traction as add-on to exercise

-1.67 [-4.86 to 1.52]

Adverse events

no differences among treatment groups in number, type, duration or severity of adverse reactions

Moustafa, 2014

Type of study:

RCT

Setting and country:

Single centre, Egypt

Funding and conflicts of interest:

none

Inclusion criteria:

(1) Unilateral C5-6 or C6-7 disc herniation confirmed by CT or MRI, (2) dermatomal numbness C6 or C7, (3) current pain or discomfort for > 3 months, (4) radiation of pain in arm with diminished deep tendon reflexes, (5) increase in symptoms with cervical flexion or protrusion and decrease with retraction or side bending and rotation,(6) presence of 4 positive examination findings in provocation tests; all in patients recruited from outpatient physiotherapy department.

Exclusion criteria:

(1) presence of medical “red flags” (tumor, fracture, RA, osteoporosis, prolonged steroid use), (2) history of c- or t-spine surgery, (3) signs of upper motor neuron disease, (4) vestibulobasilar insufficiency, (5) amyothropic lateral sclerosis, (6) bilateral symptoms, (7) pregnancy, (8) complete loss of sensation along involved nerve root, (9) severe myelopathy from history taking of motor loss > 3 on MRC scale

N total at baseline: 216

Intervention: 144 (72 group 1 and 72 group 2)

Control: 72

Important prognostic factors:

age:

I: 40.2 ± 4.9 (G1), 41.5 ± 6.1 (G2)

C: 41.7 ± 5.5

Sex:

I: 43% M (G1), 61% M (G2)

C: 56% M

Groups comparable at baseline? Yes

Intervention treatment:

Group 1 (G1): intermittent ventroflexion traction with increasing traction force, and an on/off cycle of 50/10 for 20 minutes, 3 times per week for 4 weeks.

Group 2 (G2): FCR H-reflex-based traction, same as group 1 yet with different head posture. For 20 minutes, 3 times per week for 4 weeks.

Both group 1 and 2 received in addition the multimodal treatment as described under the control treatment.

Control treatment:

Multimodal program with

• pain relief methods (infrared radiation for 15 minutes, interferential therapy for 20 minutes at 100Hz, and massage),
• muscle strengthening via isometric contraction exercises increasing in time or resistance, twice daily
• Thoracic spine manipulation through thrust manipulation.

Duration of 4 weeks, 3 times per week.

Length of follow-up: 4 weeks, 12 months

Loss-to-follow-up:

Intervention: 7 (G1) and 13 (G2) (13.9%)

Reasons: loss to follow-up for non-medical causes

Control: 7 (9.7%)

Reasons: loss to follow-up for non-medical causes

Incomplete outcome data: 27 (12.5%)

Same as above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

Traction as add-on to manual therapy

-0.95 [-1.37 to -0.53]

NDI (mean difference, 95%CI)

Traction as add-on to manual therapy

-19.00 [-21.05 to -16.95]

Young, 2009

Type of study:

RCT

Setting and country:

Multi-centre, USA

Funding and conflicts of interest:

A grant from commercial company Saunders. Conflicts of interest not disclosed.

Inclusion criteria:

Consecutive patients with unilateral upper-extremity pain, paraesthesia or numbness, aged 18-70 years, with at least 3 out of 4 provocation tests positive.

Exclusion criteria:

(1) History of c-spine or t-spine surgery, (2) bilateral upper-extremity symptoms, (3) signs of upper motor neuron disease, (4) medical “red flags”(tumor, fracture, RA, osteoporosis, prolonged steroid use), (5) cervical spine injections (steroidal) in past 2 weeks, (6) current use of steroidal medication for radiculopathy symptoms

N total at baseline: 81

Intervention: 45

Control: 36

Important prognostic factors:

Age (mean ± SD):

I: 47.8 ± 9.9

C: 46.2 ± 9.4

Sex:

I: 31% M

C: 33% M

Duration of symptoms (> 3 months):

I: 40%

C: 58%

NRS score (mean ± SD):

I: 6.3 ± 1.9

C: 6.5 ± 1.7

NDI score (mean ± SD):

I: 19.8 ± 8.7

C: 17.1 ± 7.4

Groups comparable at baseline? Yes

Intervention treatment:

Sessions including:

• posture education (of the spine)
• Manual therapy of upper and mid thoracic spine through thrust or nonthrust (posterior-anterior glides or other) manipulation. Thereafter, nonthrust manipulation of cervical spine (30 seconds or 15-20 repetitions) were performed.
• Isometric and strengthening exercises of neck and shoulders
• Mechanical intermittent traction for 15 minutes with an on/off cycle 50/10, with c-spine at 15 degrees flexion angle. Increasing traction force was applied per visit

In addition, patients received a home exercise program.

Control treatment:

Sessions including:

• posture education (of the spine)
• Manual therapy of upper and mid thoracic spine through thrust or nonthrust (posterior-anterior glides or other) manipulation. Thereafter, nonthrust manipulation of cervical spine (30 seconds or 15-20 repetitions) were performed.
• Isometric and strengthening exercises of neck and shoulders
• Sham intermittent traction for 15 minutes with an on/off cycle 50/10, with c-spine at 15 degrees flexion angle. Max. 2.3 kg of force was applied.

In addition, patients received a home exercise program.

Length of follow-up: 4 weeks

Loss-to-follow-up:

Intervention: 6 (13.3%)

Reasons: -

Control: 6 (16.7%)

Reasons: -

Incomplete outcome data: 12 (14.8%)

Same as above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

Traction (as add-on to manual therapy and exercise) to control

0.20 [-1.23 to 1.63]

NDI (mean difference, 95%CI)

Traction (as add-on to manual therapy and exercise) to control

3.00 [-8.68 to 14.68]

Psychosocial outcomes on FABQ (mean difference, 95%CI)

Traction (as add-on to manual therapy and exercise) to control

Physical activity subscale:

-1.8 [-6.6 to 3.0]

Work subscale:

2.9 [95% -8.1 to 13.9]

3. Studies reporting on exercise*

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Dedering, 2018

Type of study:

RCT

Setting and country:

Academic hospital, Sweden

Funding and conflicts of interest:

Non-commercial support and grant, no conflict of interests

Inclusion criteria:

Patients with cervical radiculopathy were recruited from October 2010 to November 2012, with (1) verified cervical disc disease by MRI showing cervical nerve root compression, (2) neck and/or arm pain verified by neck extension test or neurodynamic provocation test.

Exclusion criteria:

Patients with (1) a previous cervical fracture, subluxation or surgery; (2) diagnosed psychiatric disorders; (3) spinal infection and malignancy; (4) other diseases or disorders contraindicating participation

N total at baseline: 144

Intervention: 72

Control: 72

Important prognostic factors:

age± SD:

I: 46.8 ± 9.6

C: 49.7 ± 9.5

Sex:

I: 47% M

C: 35% M

Neck pain frequency (daily):

I: 72%

C: 72%

Groups comparable at baseline? Yes

Intervention treatment:

Neck-specific training 3 sessions per week, for 3 months. This starts with gentle isometric neck movement, gradually progressing to low-load endurance training, individually tailored based on the patient’s response. A continuous cognitive behavioural approach was adopted during the sessions.

Additionally, patients received information folders with the elements of the intervention: pain physiology, consequences of stress and exercise, relaxation techniques, coping strategies, and ergonomic advice, plus a manual on the standardized neck-specific training program including instructions for progression.

Control treatment:

Prescribed physical activity of 30 minutes, 3 times per week, for 3 month. This starts with one individual counselling session with a cognitive behavioural approach, after which patients receive written recommendations on aerobic and/or muscular physical activity (not neck-specific).

Additionally, patients received information folders with the elements of the intervention: pain physiology, consequences of stress and exercise, relaxation techniques, coping strategies, and ergonomic advice

Length of follow-up: 24 months

Loss-to-follow-up:

Intervention: 0

Control: 4 (5.6%)

Reasons: time restriction

Incomplete outcome data:

Intervention: 31 (43%)

Reasons: discontinued intervention

Control: 36 (50%)

Reasons : discontinued intervention

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

Neck-specific training to prescribed activity

-0.30 [-1.42 to 0.82]

NDI (mean difference, 95%CI)

Neck-specific training to prescribed activity

-1.00 [-9.99 to 7.99]

Quality of Life (EQ-5D) (mean difference, 95%CI)

Neck-specific training to prescribed activity

Index: -0.03 [-0.15 to 0.09]

VAS health state: -4 [-13 to 5]

Psychosocial outcomes (FABQ) (mean difference, 95%CI)

Neck-specific training to prescribed activity

6 [0 to 12]

Diab, 2012

Type of study:

RCT

Setting and country:

Academic centre, Egypt

Funding and conflicts of interest:

None

Inclusion criteria:

Patients with (1) a craniovertebral angle <50°, (2) unilateral radiculopathy due to spondylotic changes of C5-C6 or C6-C7, (3) side-to-side amplitude differences of ≥50% in dermatomal sensory-evoked potentials, (4) duration of symptoms >3 months. Inclusion was from September 2009 to July 2010.

Exclusion criteria:

(1) Spinal canal stenosis, (2) rheumatoid arthritis, (3) vestibulobasilar insufficiency

N total at baseline: 96

Intervention: 48

Control: 48

Important prognostic factors:

Age ± SD:

I: 46.3 ± 2.05

C: 45.9 ± 2.1

Sex:

I: 56% M

C: 48% M

Groups comparable at baseline? Yes

Intervention treatment:

Posture corrective exercise programme of 10 weeks: 3 sets of 12 repetitions of 2 strengthening exercises, and 3 stretching exercises held for 30 seconds each, four times a week.

Plus 10 weeks, 3 times a week infrared radiation on the neck for 10 minutes, followed by continuous ultrasound application on upper trapezius for 10 minutes (1.5 W/cm²).

Control treatment:

10 weeks, 3 times a week infrared radiation on the neck for 10 minutes, followed by continuous ultrasound application on upper trapezius for 10 minutes (1.5 W/cm²)

Length of follow-up: 6 months

Loss-to-follow-up:

Not reported

Incomplete outcome data: not reported

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

Exercise to control

-0.70 [-1.24 to -0.16]

Selection bias through “conveniently selecting patients from our institution’s outpatient clinic”

Further risk of bias because no loss-of-follow up was reported: there might have been selective drop-out, or a per-protocol analysis performed.

4. Studies reporting on neurodynamic mobilisation

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Ayub, 2019

Type of study:

RCT

Setting and country:

Single centre, Pakistan

Funding and conflicts of interest:

No conflict of interest, information on funding not disclosed.

Inclusion criteria:

Female patients aged 30-50 years with chronic cervical radiculopathy and neck pain for ≥6 months, and positive provocation tests.

Exclusion criteria:

(1) Recent neck trauma, (2) positive vertebrobasilar insufficiency signs, (3) receiving any form of physiotherapy or medical treatment for the last 6 weeks

N total at baseline: 44

Intervention: 22

Control: 22

Important prognostic factors:

Age (median ± IQ):

I: 41 ± 12

C: 40 ± 11

Sex: not reported

Neck pain (median ± IQ):

I: 6 ± 2

C: 6 ± 2

NDI (median ± IQ):

I: 10 ± 5

C: 10 ± 10

Groups comparable at baseline? Yes

Intervention treatment:

Moist heat packs for 10 minutes followed by mechanical traction of the cervical spine for 15 minutes. Also 3 sets of slow gentle segmental mobilisation (unilateral posterior anterior glide) with at least 15 to 20 repetitions on the first session, thereafter modified based on patient response.

Then active upper extremity neurodynamic mobilisation, for 6 to 8 repetitions.

3 sessions per week, for 4 weeks.

Control treatment:

Moist heat packs for 10 minutes followed by mechanical traction of the cervical spine for 15 minutes. Also 3 sets of slow gentle segmental mobilisation (unilateral posterior anterior glide) with at least 15 to 20 repetitions on the first session, thereafter modified based on patient response.

Then passive upper extremity neurodynamic mobilisation.

3 sessions per week, for 4 weeks.

Length of follow-up: 4 weeks

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

0.0 [-1.86to 1.86]

NDI (mean difference, 95%CI)

No reporting of scale used

Cervical ROM (mean difference, 95%CI)

Reported in medians

Selection bias (convenience sampling)

Basson, 2020

Type of study:

RCT

Setting and country:

Multicentre, South-Africa

Funding and conflicts of interest:

Funding received by orthopaedic research investment fund of South African society of physiotherapy and the faculty research committee of the University of Witwatersrand

Inclusion criteria:

(1) Aged > 18 years, (2) nerve related neck and arm pain by physical examination, (3) recent onset of pain (≤12 weeks), and (4) positive upper limb neurodynamic test

Exclusion criteria:

(1) Surgery or recent fractures of the cervical spine, (2) serious neurological signs, (3) RA, neurological disease, stroke, cerebral palsy, carcinoma, or any other red flags

N total at baseline: 86

Intervention: 60

Control: 26

Important prognostic factors:

Age (mean ± SD):

I: 46.5 ± 14.1

C: 48.6 ± 13.6

Sex: not reported

Duration of pain (mean days ± SD):

I: 30.2 ± 27.4

C: 23.5 ± 22.9

Groups comparable at baseline? Yes

Intervention treatment:

neurodynamic mobilisation along the tract of the nerve, directly and indirectly, concentrating on areas where the nerve is mechano-sensitive to palpation. From hand or elbow up along the arm, first rib, scalene and into the neck, first in a non-tensioned position, progressing into a more tensioned position as pain and irritability improved.

In addition, patient received usual care similar to patients in the usual care group.

The number of treatments was determined by the treating physiotherapist.

Control treatment:

Usual care with (unilateral) posterior-anterior mobilisation of the cervical and thoracic spine, exercises, and advice to stay active.

The number of treatments was determined by the treating physiotherapist.

Length of follow-up: 6 weeks, 6 months, 12 months

Loss-to-follow-up:

Intervention: 7 (11.7%)

Reasons: could not be reached (of whom 2 had relocated)

Control: 1 (3.8%)

Reasons: could not be reached

Incomplete outcome data: 0

Multiple imputation applied.

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-0.60 [-1.77 to -0.51]

Function on PSFS (mean difference, 95%CI) (scale 0 to 30, high score is better)

0.05 [0-0.41 to 0.51]

Quality of Life (EQ-5D) (mean difference, 95%CI)

1.1 [4.3 to 6.3]

Ibrahim, 2021

Type of study:

RCT

Setting and country:

Single centre, Egypt

Funding and conflicts of interest:

No conflicts of interest, funding information not explicitly reported.

Inclusion criteria:

Patients (1) 20-40 years of age, (2) history of pain >3 months, (3) radiating pain in one upper limb, (4) met at least 3 of Wainner criteria

Exclusion criteria:

(1) history of high level spinal cord injury, (2) malignancy, (3) any medical red flag (tumor, fracture, RA, osteoporosis, prolonged steroid use), (4) circulatory disturbances of upper extremity, (5) traumatic injuries of upper limb and cervical spine, (6) dizziness

N total at baseline: 40

Intervention: 20

Control: 20

Important prognostic factors:

Pain (median ± range):

I: 3.75 ± 7

C: 3.75 ± 7

Further no baseline characteristics reported

Groups comparable at baseline? No information

Intervention treatment:

Tensioning neurodynamic mobilisation of brachial plexus, with the arm in neurodynamic testing position, 10 cycles of elbow extension and flexion (each 3 seconds) were administered

Plus tradition physiotherapy (consisting of infrared radiation and manual traction), for 3 sessions per week, over the course of 3 weeks.

Control treatment:

Traditional physiotherapy:

• Infrared radiation 50/60 Hz for 20 minutes
• Manual traction 15 seconds pull with 30 seconds rest, for 3 sets of 10 repetitions, with 60 seconds rest between sets.

For 3 sessions per week, over the course of 3 weeks.

Length of follow-up: 3 weeks

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

0.0 [-1.17 to 1.17]

Selection bias

Kayiran, 2021

Type of study:

RCT

Setting and country:

Single centre, Turkey

Funding and conflicts of interest:

None

Inclusion criteria:

Patients for whom surgery was not recommended by the neurosurgeon, with (1) cervical disc herniation at C5/C6/C7/C8 confirmed by MRI, (2) aged 20-50 years, (3) cervicobrachial radicular pain for ≥6 weeks (4) pain severity ≥5 on VAS, (5) sensitivity and numbness in radial, median and/or ulnar nerve dynamic tests, (6) not receiving any other treatment or pharmacological agents.

Exclusion criteria:

(1) Spinal stenosis, (2) RA, (3) previous c-spine surgery, (4) severe neurological loss, (5) upper extremity vascular problems, (6) severe osteoporosis, (7) diabetes mellitus, (8) pregnant women.

N total at baseline: 71

Intervention: 36

Control: 35

Important prognostic factors:

Age (mean ± SD):

I: 47.2 ± 12.5

C: 43.3 ± 12.0

Sex: not reported

Neck pain (mean ± SD):

I: 4.83 ± 1.12

C: 4.83 ± 1.23

NDI (mean ± SD)

I: 18.93 ± 6.43

C: 18.93 ± 5.10

Groups comparable at baseline? Yes

Intervention treatment:

neurodynamic mobilisation on radial, median and ulnar nerves, 10 times 10 seconds for each nerve.

For 10 sessions over 3 weeks.

Plus conservative physiotherapy (hotpacks, TENS, ultrasound and exercises) for 3 weeks, 5 sessions per week.

Control treatment:

Conservative physiotherapy:

• Hotpack for 20 minutes
• Transcutaneous electric nerve stimulation (TENS) for 20 minutes at 100Hz
• Continuous mode ultrasound for deep heating for 5 minutes at 1 MHz
• Exercises for stretching and strengthening for 5 seconds with 10 repetitions

For 3 weeks, 5 sessions per week

Length of follow-up: 3 weeks

Loss-to-follow-up:

Intervention: 6 (16.7%)

Reasons: no belief in treatment (n = 3), no adaptation to treatment time (n = 2), relocation (n = 1)

Control: 5 (14.3%)

Reasons: no belief in treatment (n = 2), no adaptation to treatment time (n = 2), quit treatment because of poor health (n = 1)

Incomplete outcome data:

Intervention: 11 (15.5%)

Reasons: as mentioned above.

Outcome measures and effect size (95%CI):

Pain on VAS (mean difference, 95%CI)

-1.04 [-1.57 to -0.51]

NDI (mean difference, 95%CI)

No reporting of scale used

Cervical ROM (mean difference, 95%CI)

Flexion: 3.9 [1.1 to 6.7]

Extension: 6.2 [3.0 to 9.5]

Side bending right: 4.4 [1.5 to 7.3]

Side bending left: 5.6 [2.6 to 8.6]

Rotation right: -0.7 [-5.3 to 3.8]

Rotation left: -0.3 [-5.4 to 4.8]

Kim, 2017

Type of study:

RCT

Setting and country:

Single centre, Korea

Funding and conflicts of interest:

No conflict of interest. Funding information not disclosed.

Inclusion criteria:

Patients with (1) a diagnosis of cervical radiculopathy ≥3 months, (2) aged 26 to 60 years, (3) unilateral pain, (4) ≥3 out of 4 positive provocation tests

Exclusion criteria:

None reported

N total at baseline: 30

Intervention: 15

Control: 15

Important prognostic factors:

Age (mean ± SD):

I: 29.3 ± 3.3

C: 29.3 ± 3.1

Sex:

I: 40% M

C: 33% M

NRS (mean ± SD):

I: 7.0 ± 0.85

C: 7.1 ± 0.80

NDI (mean ± SD):

I: 21.7 ± 4.1

C: 22.1 ± 3.0

Groups comparable at baseline? Yes

Intervention treatment:

Similar treatment as control treatment group, yet during the manual cervical traction, another physiotherapist applied neurodynamic mobilisation using a slider technique for the median nerve in a smooth and rhythmic manner (with elbow extension/flexion and wrist flexion/extension), for 6 times 1 minute with 30 seconds rest inbetween.

For 8 weeks, 3 times per week.

Control treatment:

Conservative physiotherapy (total 35 minutes):

• Hot pack for 20 minutes
• TENS at 60Hz for 15 minutes

Plus manual cervical traction for 6 repetitions of 1 minute pull with 30 seconds rest (total 10 minutes)

For 8 weeks, 3 times per week.

Length of follow-up: 4 and 8 weeks

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-1.0 [-1.69 to -0.31]

NDI (mean difference, 95%CI)

-6.92 [-11.41 to -2.43]

Cervical ROM (mean difference, 95%CI)

Flexion: 3.3 [0.3 to 6.4]

Extension: 5.1 [2.2 to 8.0]

Side bending right: 2.6 [0.7 to 4.5]

Side bending left: 2.4 [0.9 to 3.9]

Rotation right: 2.4 [0.3 to 4.5]

Rotation left: 3.6 [1.8 to 5.4]

Rodriguez-sans, 2017

Type of study:

RCT

Setting and country:

Single centre, Venezuela

Funding and conflicts of interest:

none

Inclusion criteria:

Consecutive patients seeking treatment for cervicobrachial pain, who (1) had clinical signs of cervicobrachial pain (arm pain, paresthesia, numbness in upper limb), confirmed through MRI, (2) aged 18 to 45, (3) unilateral symptoms for at least 3 months, (4) 3 positive provocation tests

Exclusion criteria:

(1) Use of any type of treatment to relieve pain (therapy, procedures or drugs), (2) using anticonvulsant, antidepressant, or psychotropic medication, (3) vertebral instability, osteoporosis or spine infection, (4) neurologic diseases, (5) cervical stenosis myelopathy, (6) kinesiophobia, (7) pregnancy, (8) endocrine disorders and menopause, (9) history of spine surgery, (10) severe mental illness, (11) intoxication, (12) intellectual disability, (13) severe sleep deprivation, (14) Alzheimer’s disease

N total at baseline: 58

Intervention: 29

Control: 29

Important prognostic factors:

Age(mean ± SD):

I: 33.3 ± 5.0

C: 32.5 ± 4.6

Sex:

I: 56% M

C: 44% M

NRS (mean ± SD):

I: 6.08 ± 0.99

C:6.44 ± 0.93

Groups comparable at baseline? As far as information is provided, yes

Intervention treatment:

Cervical lateral glide (CLG) neural mobilisation administered by a physiotherapist, to the contralateral side of pain in a slow oscillating manner. CLG was applied continuously for two minutes in 5 consecutive applications, with one minute rest in between. For 5 days per week, for 6 weeks.

Control treatment:

Waiting list for 6 weeks (did not receive an type of pain-modulating treatment)

Length of follow-up: 6 weeks (30 treatment days)

Loss-to-follow-up:

Intervention: 4 (13.8%)

Reasons: did not complete treatment as allocated

Control: 2 (6.9%)

Reasons: did not complete treatment as allocated

Incomplete outcome data: 6 (10.3%)

Reasons: see above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-2.96 [-3.54 to -2.38]

Function (Quick DASH)

-21.5 [-27.4 to -15.73]

Function (ipsilateral cervical rotation) (mean difference, 95%CI)

8.0 [4.9 to 11.1]

Savva, 2016

Type of study:

RCT

Setting and country:

Single centre, Cyprus

Funding and conflicts of interest:

None.

Inclusion criteria:

Consecutive patients with unilateral cervical radiculopathy diagnosed by MRI or CT referred to the physiotherapy department, with (1) unilateral sharp pain, muscle weakness and numbness in upper arm, (2) ≥3 out of 4 positive provocation tests.

Exclusion criteria:

(1) current cervical myelopathy or signs of upper motor neuron disease, (2) bilateral complaints, (3) other musculoskeletal conditions in the affected limb, (4) use of analgesia or anti-inflammatory medication in the prior 2 weeks.

N total at baseline: 42

Intervention: 21

Control: 21

Important prognostic factors:

Age (mean ± SD):

I: 45.2 ± 13.5

C: 49.2 ± 8.5

Sex:

I: 38% M

C: 62% M

NRS (mean ± SD):

I: 5.62 ± 2.52

C: 5.19 ± 2.11

NDI (mean ± SD):

I:33.3 ± 17.6

C: 30.2 ± 16.4

Groups comparable at baseline? Yes

Intervention treatment:

Intermittent (pain-free) cervical traction with 1 minute pull and 1 minute rest for 6 sets. During the cervical traction, slider neurodynamic mobilisation using a median nerve bias were performed in slow and oscillatory fashion, with the patients’ elbow, wrist and finger repositioning. 3 treatment sessions per week, for 4 weeks.

In addition, patients received advice to avoid prescription or over-the-counter analgesia or anti-inflammatory medication.

Control treatment:

Did not receive any type of treatment, with advice to avoid prescription or over-the-counter analgesia or anti-inflammatory medication, for the duration of 4 weeks.

Length of follow-up: 4 weeks

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-3.36 [-4.56 to -1.96]

NDI (mean difference, 95%CI)

-15.33 [-23.26 to -7.40]

Function on PSFS (mean difference, 95%CI) (scale 0 to 10)

1.62 [0.91 to 2.32]

Cervical ROM (mean difference, 95%CI)

Flexion: 5.7 [0.5 to 11.0]

Extension: 8.8[-0.5 to 18.1]

Side bending ipsilateral:

6.4 [1.6 to 11.1]

Side bending contralateral:

6.0 [1.5 to10.4]

Rotation ipsilateral:

7.9 [1.5 to 14.3]

Rotation contralateral:

10.6 [3.6 to 17.6]

Savva, 2021

Type of study:

RCT

Setting and country:

Single centre, Cyprus

Funding and conflicts of interest:

None.

Inclusion criteria:

Consecutive patients with unilateral cervical radiculopathy referred to the physiotherapy department, (1) aged 20 to 75 years, with (2) unilateral upper limb pain, sensory and/or motor symptoms, (3) ≥3 out of 4 positive provocation tests.

Exclusion criteria:

(1) bilateral complaints, (2) other musculoskeletal conditions in the affected limb, (3) evidence of central nervous system involvement, (4) history of medical red flags (tumor, metabolic disease, RA, osteoporosis), (5) use of analgesia or anti-inflammatory medication in the prior 2 weeks.

N total at baseline: 66

Intervention: 22 (G1), 22 (G2)

Control: 22

Important prognostic factors:

Age (mean ± SD):

I: 47.7 ± 10.8 (G1), 48.1 ± 11.9 (G2)

C: 48.5 ± 12.3

Sex:

I: 50% M (G1), 59% M (G2)

C: 36% M

Duration of symptoms >3 months:

I: 54% (G1), 41% (G2)

C: 55%

NRS (mean ± SD):

I: 6.1 ± 2.2 (G1), 6.1 ± 2.6 (G2)

C: 5.4 ± 1.8

NDI (mean ± SD):

I: 33.3 ± 17.6 (G1), 34.5 ± 14.2 (G2)

C: 29.0 ± 15.9

Groups comparable at baseline? Gender seems to differ sligthly at baseline between intervention and control groups.

Intervention treatment:

Group 1 (G1): Cervical traction with neurodynamic mobilisation

Intermittent (pain-free) cervical traction with 1 minute pull and 30 seconds rest for 10 sets. During the cervical traction, slider neurodynamic mobilisation of the median nerve was performed in slow and oscillatory fashion, with repeated passive flexion and extension of the patients’ elbow, wrist and fingers. 3 treatment sessions per week, for 4 weeks.

Group 2 (G2): cervical traction with sham neurodynamic mobilisation

Intermittent (pain-free) cervical traction with 1 minute pull and 30 seconds rest for 10 sets. During the cervical traction, sham neurodynamic mobilisation with sustained position of elbow and wrist, and repeated passive flexion and extension of the patients’ fingers within the ROM. 3 treatment sessions per week, for 4 weeks.

Control treatment:

Waiting list without any type of treatment for 4 weeks.

Length of follow-up: 4 weeks

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

NM + cervical traction to waiting list

-3.30 [-4.51 to -2.09]

NM as add on to cervical traction to sham NM

-2.4 [-3.75 to -1.05]

NDI (mean difference, 95%CI)

-32.60 [-47.56 to -17.64]

Function on PSFS (mean difference, 95%)

1.39 [0.72 to 2.05]

Cervical ROM

Flexion: 5.9 [0.8 to 11.0]

Extension: 5.8[-3.1 to 14.7]

Side bending ipsilateral:

6.2 [1.4 to 11.1]

Side bending contralateral:

6.3 [2.2 to10.4]

Rotation ipsilateral:

8.7 [1.7 to 15.7]

Rotation contralateral:

13.5 [6.7 to 20.3]

5. Studies reporting on manual therapy**

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Hassan, 2020

Type of study:

RCT

Setting and country:

single centre, Pakistan

Funding and conflicts of interest:

No conflicts of interest, funding information not disclosed.

Inclusion criteria:

(1) Patients aged 30-70 with positive findings of cervical radiculopathy on X-ray, (2) decreased range of motion, (3) positive neurodynamic provocation tests, (4) neck pain <8 on NRS, and (5) numbness or paresthesia or pain in arm or hand

Exclusion criteria:

(1) Cervical myelopathy, (2) vertebrobasilar insufficiency, (3) recent history of trauma, (4) thoracic outlet syndrome, (5) carpal tunnel syndrome, (6) use of pain medications for cervical radiculopahty

N total at baseline: 46

Intervention: 23

Control: 23

Important prognostic factors:

age± SD:

Overall: 43.1 ± 8.2 years

Sex:

I: 65% M

C: 70% M

NRS score (median ± IQ ):

I: 8.0 ± 0.75

C: 8.0 ± 1.75

NDI score (median ± IQ):

I: 39.5 ± 4

C: 41.0 ± 10

Groups comparable at baseline? Insufficient information, yet seems so.

Intervention treatment:

Oscillatory mobilisation (Maitland), 3 sets of 15 repetitions of unilateral postero-anterior glide on the involved segment. For 7 treatment sessions over 2 weeks.

Plus a home exercise plan (stretching and strengthening exercises), heat therapy and TENS for 10 minutes

Control treatment:

Sustained stretch mobilisation (Kaltenborn): 3 sets of cervical traction and cervical segment flexion, coupled with side bending and rotation. For 7 treatment sessions over 2 weeks.

Plus a home exercise plan (stretching and strengthening exercises), heat therapy and TENS for 10 minutes

Length of follow-up: 2 weeks

Loss-to-follow-up: 6

Intervention: 3 (13%)

Reasons: loss to follow up (n=3)

Control: 3 (13%)

Reasons: Loss to follow-up (n = 2), discontinued intervention (n = 1)

Incomplete outcome data: 6 (13%)

See above.

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-1.0 [-2.0 to 0.0]

NDI (mean difference, 95%CI)

Not reported which scale used

Cervical range of motion (ROM, medians)

Flexion: 47 to 24

Extension: 59 to 45

Side bending right: 45 to 45

Side bending left: 45 to 45

Right rotation: 80 to 50

Left rotation: 80 to 53.5

Selection bias

Per protocol analysis

Shafique, 2019

Type of study:

RCT

Setting and country:

Single centre, Pakistan

Funding and conflicts of interest:

none

Inclusion criteria:

(1) Cervical radiculopathy patients aged 20 to 60 years, (2) ≥3 out of 4 positive provocation tests; (3) pain and paraesthesia in the unilateral upper extremity and limited cervical ROM

Exclusion criteria:

(1) bilateral upper extremity symptoms, (2) previous C-spine or T-spine injury, (3) recent gracture or surgery in and around the shoulder, (4) any systemic disease, (5) unsTable spine

N total at baseline: 38

Intervention: 19

Control: 19

Important prognostic factors:

age± SD:

I: 42.3 ± 10.4

C: 41.0 ± 9.3

Sex:

I: 33% M

C: 44% M

NRS score (mean ± SD ):

I: 6.87 ± 1.06

C: 6.94 ± 0.85

NDI score (mean ± SD):

I: 23.7 ± 7.9

C: 20.2 ± 6.8

Groups comparable at baseline? Yes

Intervention treatment:

Spinal mobilisation with arm movement, through maintaining transverse glide with 10 repetitions in the first sessions, increasing to 30 repetitions in sets of 3 in further sessions. For 2 sessions per week, during 3 weeks.

Plus treatment the control group received.

Control treatment:

Conventional treatment, consisting of:

• Hot pack applied for 10 minutes
• active range of motion exercises with 3 sets of 10 repetitions and isometric exercises repeated 20x with a hold of 6 to 10 seconds
• Sliding technique for 10 repetitions
• Manual traction for 10 minutes with 10 second pull and 5 second rest.

For 2 sessions per week, for 3 weeks.

Length of follow-up: 3 weeks

Loss-to-follow-up:

Intervention: 4 (21.1%)

Reasons: drop-out

Control: 3 (15.8%)

Reasons: drop-out

Incomplete outcome data: 7 (18.4%)

Reasons: same as above

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-1.41 [-2.10 to -0.72]

NDI (mean difference, 95%CI)

Not reported which scale used

Cervical ROM (means)

Flexion: 41.06 to 45.93

Extension: 51.6 to 45.31

Side bending right: 51.6 to 29.94

Side bending left: 39.6 to 31.25

Right rotation: 69.27 to 61.5

Left rotation: 67.67 to 62.81

z

Unclear eligibility and exclusion criteria (unsTable spine?)

Young, 2019

Type of study:

RCT

Setting and country:

Multicentre, United states of America

Funding and conflicts of interest:

none

Inclusion criteria:

Consecutive patients with unilateral upper extremity pain, paresthesia or numbness, (1) 18 to 65 years of age, (2) NDI score of ≥10/50, (3) ≥3 out of 4 positive provocation tests

Exclusion criteria:

(1) previous c- or t-spine surgery, (2) bilateral upper extremity symptoms, (3) signs of upper motor neuron disorder, (4) medical red flags (tumor, fracture, RA, osteoporosis, prolonged steroid use), (5) cervical steroidal injection or medication within past 2 weeks

N total at baseline: 43

Intervention: 22

Control: 21

Important prognostic factors:

age± SD:

I: 48.8 ± 11.5

C: 43.1 ± 10.8

Sex:

I: 23% M

C: 43% M

Neck pain onset > 6 months:

I: 36%

C: 52%

Groups comparable at baseline? Moderately comparable; differences in gender, onset time, employment status, previous treatment

Intervention treatment:

Thoracic spine manipulation: supine, high-velocity, low-amplitude thrust manipulation technique, at exhalation of patient, directed bilaterally to the upper and mid thoracic spine, with audible cavitation.

Control treatment:

Sham manipulation: same position as intervention group, yet the physiotherapist’s hand over the inferior vertebrae was open with fingers extended. No thrust manipulation delivered during exhalation.

Length of follow-up: 48 to 72 hours after treatment

Loss-to-follow-up:

Intervention: 0

Control: 0

Incomplete outcome data: 0

Outcome measures and effect size (95%CI):

Pain on NRS (mean difference, 95%CI)

-3.1 [-4.5 to -1.7]

NDI (mean difference, 95%CI) (scale 0 to 50)

-8.0 [95%CI -11.6 to -4.5]

Cervical ROM

Reported in mean differences in change of score

Adverse events

No increase in neck, arm, or hand symptoms were reported, and no participants reported soreness lasting more than 3 hours after treatment

Study reference

Was the allocation sequence adequately generated?

Was the allocation adequately concealed?

Blinding: Was knowledge of the allocated interventions adequately prevented?

• Were patients blinded?
• Were healthcare providers blinded?
• Were data collectors blinded?
• Were outcome assessors blinded?
• Were data analysts blinded?

Was loss to follow-up (missing outcome data) infrequent?

Are reports of the study free of selective outcome reporting?

Was the study apparently free of other problems that could put it at a risk of bias?

Overall risk of bias

If applicable/ necessary, per outcome measure

Definitely yes

Probably yes

Probably no

Definitely no

Definitely yes

Probably yes

Probably no

Definitely no

Definitely yes

Probably yes

Probably no

Definitely no

Definitely yes

Probably yes

Probably no

Definitely no

Definitely yes

Probably yes

Probably no

Definitely no

Definitely yes

Probably yes

Probably no

Definitely no

LOW

Some concerns

HIGH

1. Studies reporting on the use of a collar

Aksoy, 2018

Probably yes

Reason: use of a computer-generated randomization Table, yet randomization process not described.

Probably no

Reason: no allocation process described.

Probably no

Reason: Patients, care provider not blinded. Outcome assessor was blinded, yet blinding of data analysts not reported.

Probably no

Reason: Loss to follow-up was infrequent (max. 10%) in intervention and control group. However, no imputation methods were used.

Unclear

Reason: No protocol available or previously published.

Probably no

Reason: No ITT analysis. Possible selection bias. No sample size calculation, possibly underpowered.

HIGH

• Inadequate allocation concealment
• per protocol analysis
• Unclear whether deviation from protocol occurred
• possible selection bias

Kuijper, 2009

Definitely yes

Reason: computer generated sequence per hospital.

Definitely yes

Reason: sequentially numbered opaque, sealed envelopes.

Probably no

Reason: Patients, physiotherapists and investigators not blinded (blinding of data analysts not reported)

Probably no

Reason: Loss to follow-up was infrequent in intervention and control group. Inadequate imputation methods were used (last observation carried forward). Sample size not reached, possibly underpowered.

Definitely no

Reason: secondary outcomes in protocol included time to work, persistent symptoms and surgery; in study report these are sick leave, satisfaction, and use of analgesics

Probably no

Reason: patient selection might have occurred, as cervical radiculopathy is a clinical diagnosis, whereas all patients had to have MRI positive findings (which can be false negative)

Some concerns

• No investigator blinding
• Deviation from protocol in outcome reporting
• possible selection bias
• Underpowered

2. Studies reporting on cervical traction

Bukhari, 2016

Definitely no

Reason: described as “toss and trial method” without further explanation

Probably no

Reason: no information on allocation concealment provided

Probably no

Reason: Patients (self-reported outcome measures) and health care providers not blinded. Unclear whether data-analysts were blinded.

Probably no

Reason: more than 10% loss to follow-up (can have impact on the intervention effect estimate), unclear in which groups it occurred

Probably no

Reason: no protocol provided, scarce number of outcome measures.

Probably no

Reason: no limitations of study mentioned in discussion, yet effect of traction compared to other treatment both groups received seems negligible. No sample size calculation, possibly underpowered.

HIGH

• Inadequate randomization
• Risk of selection bias
• No blinding
• Frequent and possible selective loss to follow-up
• Underpowered

Ojoawo, 2019

Definitely yes

Reason: patient drew opaque envelope with allocation

Probably yes

Reason: opaque envelopes, yet unclear whether caregiver or data assessor was involved in process.

Probably no

Reason: Patients (self-reported outcome measures) and health care providers not blinded. Unclear whether data-analysts were blinded.

Probably yes

Reason: discontinuation of 3 persons has been reported. Unclear imputation method used (probably LOCF). Based on sample size calculation, study is underpowered.

Probably no

Reason: no protocol provided, scarce number of outcome measures.

Definitely no

Reason: Unclear whether ITT or PP analysis was done. Outcomes are reported in method section. Unclear inclusion and exclusion criteria

HIGH

• High risk of selection bias
• unclear allocation concealment
• no research protocol
• Unclear whether ITT or PP analysis has been performed
• Underpowered

Fritz, 2014

Definitely yes

Reason: opaque sealed envelopes prepared prior to enrolment, based on web-based randomization generator with allocation block sizes of 6, 8 or 10.

Probably yes

Reason: research assistant opened randomization envelope

Probably yes

Reason: patients and physiotherapists were not blinded, yet assessors and researchers were blinded.

Probably yes

Reason: loss-to-follow up was frequent (37%), yet adequate imputation method through linear mixed models with repeated measurements, using maximum-likelihood estimation. However, imputation is not advised for >10% missing values.

Probably no

Reason: as secondary outcome measure in protocol, “global rating of change” has been reported.

Probably no

Reason: calculated required sample size not reached (underpowered), several baseline differences in treatment groups (e.g. duration of symptoms). Sample size calculation is not reached, underpowered.

Some concerns

• Due to large loss to follow-up and crossover, results might be slightly biased
• Deviation from protocol in outcome reporting
• Underpowered

Moustafa, 2014

Definitely yes

Reason: random number generator restricted to permuted blocks of different sizes, transferred to a sequence of consecutively numbered, sealed opaque envelopes

Definitely yes

Reason: independent person picked next sealed opaque envelope

Definitely no

Reason: Patients (self-reported outcome measures) and health care providers not blinded. Unclear whether data-analysts were blinded.

Probably yes

Reason: loss to follow-up 12.5%, missing data imputed with linear mixed model with repeated measurements

Probably no

Reason: protocol retrospectively registered. Patients included from 35-48 years of age not specified in manuscript.

Probably no

Reason: highly selective inclusion criteria for study.

Some concerns:

• Patients not random sample of population
• Risk of further selection bias
• No protocol available prior to study start

Young, 2009

Probably yes

Reason: numbered, sequential, sealed envelopes containing group allocation, stratified by clinic. Unclear generation of randomization sequence.

Definitely yes

Reason: opaque sealed envelopes opened by evaluating therapist

Probably yes

Reason: Patients were -semi blinded (sham treatment. Health care providers were not blinded. Support staff unaware of group assignment administered self-report measures and testing. Data analysts not involved in treatment.

Probably yes

Reason: 15% loss to follow-up, yet repeated-measure linear mixed model analysis was performed for imputation.

Unclear

Reason: no protocol available.

Probably no

Reason: Possible selection bias due to patient selection through a clinical prediction rule. Home exercise adherence was not recorded.

Some concerns:

• possible selection bias
• possible selective outcome reporting

3. Studies reporting on exercise

Dedering, 2018

Definitely yes

Reason: computer-generated randomized Table, randomizing patients in blocks of 8 with allocation ratio 1:1

Definitely yes

Reason: treatment allocation put in opaque sealed envelope by statistician not involved in recruitment

Probably yes

Reason: patients and care providers were not blinded (but difficult with regard to the intervention), but data entering and analysis were performed by different blinded individuals.

Definitely no

Reason: large proportions in both groups (nearly 50%) discontinued intervention/control. Unclear imputation methods, imputation is not advised for >10% missing values.

Probably no

Reason: in clinical trial registry additional secondary outcome measures are described

Probably yes

Reason: ITT analysis performed (PP would be underpowered), no other problems mentioned.

Some concerns

• High dropout rate
• Selective outcome reporting

Diab, 2012

Probably yes

Reason: random permuted size 4 blocks with equal allocation ratio, by the roll of a dice (odd/even)

Probably no

Reason: end of block size prediction, resident blinded to research protocol operated the random assignment

Definitely no

Reason: patients, care providers, and outcome assessors not blinded.

Probably no

Reason: loss to follow-up not reported for outcomes after 6 months

Unclear

Reason: all prespecified outcomes in methods were reported. No protocol available.

Probably no

Reason: unclear whether ITT has taken place. Analysis was adjusted for baseline value.

HIGH

• inadequate blinding
• unclear dropout rate
• high risk of biased analysis

4. Studies reporting on neural mobilisation

Ayub, 2019

Probably no

Reason: random allocation sequence was generated by one of the researchers

Definitely no

Reason: inadequate randomization procedure

Probably yes

Reason: data collector and analyst blinded for group allocation. Patients, health care professionals and data assessors not blinded.

Definitely yes

Reason: No loss to follow-up, yet insufficient recruitment (underpowered)

Unclear

Reason: No protocol available or previously published.

Definitely no

Reason: selection bias through convenience sampling. Unclear interventions. Unclear choice for only female patients.

HIGH

• Selection bias
• inadequate randomization and allocation concealment
• Underpowered

Basson, 2020

Definitely yes

Reason: blocked randomization with a 2:1 ratio in blocks of 6 with a computer random numbet generator

Definitely yes

Reason: Administrative research assistant informed participating physiotherapist of allocation after baseline measurements

Probably yes

Reason: Patients and physiotherapists were not blinded. Follow-up measurements were taken by blinded physiotherapist

Definitely yes

Reason: multiple imputation applied for missing data (larger in intervention group).

Probably no

Reason: Global Rating of Change Scale reported in protocol, yet not in article.

Probably no

Reason: early stop of recruitment, reducing power to 85%. Possible underpowered for detecTable difference.

LOW

Ibrahim, 2021

Probably yes

Reason: opaque sealed envelopes containing the name of one of the groups

Definitely yes

Reason: envelopes were picked by investigator not participating in the study

Probably no

Reason: Patients and physiotherapists not blinded. No reporting of data collector, assessor, or analyst blinding.

Definitely yes

Reason: no loss to follow-up.

Probably no

Reason: no baseline data provided. No protocol available.

Probably no

Reason: high risk of selection bias. Sample size calculated for disability, yet no disability outcome measure was taken.

HIGH

• High risk of selection bias
• No blinding of data collectors or analysts
• unclear sample size calculation

Kayiran, 2021

Probably no

Reason: no information available

Probably no

Reason: no information available

Probably yes

Reason: patients and physiotherapists were not blinded. Pre- and post-treatment evaluation by blinded physiotherapists. Unclear whether data analysts were blinded.

Definitely no

Reason: >10% loss to follow-up without the use of adequate imputation methods (per protocol analysis).

Unclear

Reason: no protocol available

Definitely no

Reason: no sample size calculation available (underpowered?)

HIGH

• Possible selection bias
• high risk of inadequate randomization
• Unclear sample size calculation

Kim, 2017

Probably no

Reason: no information available

Probably no

Reason: no information available

Probably no

Reason: no information available.

Definitely yes

Reason: no loss to follow-up.

Unclear

Reason: no protocol available

Probably no

Reason: risk of selection bias.

HIGH

• possible selection bias
• Lacking relevant information for methodological assessment of study

Rodriguez-sans, 2017

Definitely yes

Reason: restricted block randomization with 1:1 allocation ratio through computerized randomization software

Definitely yes

Reason: printed cards in consecutively numbered opaque sealed envelopes

Definitely yes

Reason: subjects, physiotherapists and outcome assessors were blinded to the hypothesis tested and the existence of a second group (unethical?). Data analyst were blinded.

Probably no

Reason: loss-to-follow-up was >10%, has impact on intervention effect estimate. As basis for sample size calculation unclear, risk for being underpowered.

Unclear

Reason: protocol registered after recruitment of patients had started.

Probably yes

Reason: no other problems reported, though authors state themselves that generalization of results is difficult.

Some concerns

• Per protocol analysis (with risk of underpower)
• Possible selective outcome reporting (

Savva, 2016

Definitely yes

Reason: block randomization with blocks of 4 in sealed envelopes

Definitely yes

Reason: researcher responsible for participant allocation (not involved in data collection or treatment) was blinded to block size

Probably yes

Reason: Patients blinded to group assignment. Physiotherapists not blinded. Data collectors and assessors were blinded.

Definitely yes

Reason: no loss to follow-up. Sample size calculation powered on NDI.

Unclear

Reason: no protocol available

Probably yes

Reason: no other issues reported. Patients unaware of other treatment group seems unethical.

LOW

Savva, 2021

Probably yes

Reason: block randomization with block size of 4, yet sequence generation not described

Definitely yes

Reason: researcher responsible for participant allocation (not involved in data collection or treatment) was blinded to block size

Probably yes

Reason: Patients were blinded to group assignment. Data collectors were blinded to nature and purpose of study.

Definitely yes

Reason: no loss to follow-up. Sample size calculation powered on NDI.

Unclear

Reason: protocol registered after recruitment of patients had started.

Probably yes

Reason: no other issues reported.

LOW

5. Studies reporting on manual therapy

Hassan, 2020

Probably yes

Reason: lottery method

No information

Reason: specifics of lottery method not described

Probably yes

Reason: patient and data analyst were blinded. Caregiver not blinded (not possible).

Probably no

Reason: 13% loss to follow-up, has impact on intervention effect estimate (underpowered).

Probably yes

Reason: all prespecified outcomes in methods were reported. No protocol available.

Definitely no

Reason: selection bias through purposive sampling. Per protocol analysis.

HIGH:

• Selection bias
• Unclear allocation concealment
• Per protocol analysis
• Underpowered

Shafique, 2019

Probably yes

Reason: lottery method

No information

Reason: specifics of lottery method not described

Probably no

Reason: no information provided on blinding of subjects, caregivers, data collectors or data analysts.

Definitely no

Reason: Nearly 20% loss to follow up, current per protocol analysis has impact on effect estimate of intervention

Probably yes

Reason: all prespecified outcomes in methods were reported. No protocol available.

Definitely no

Reason: unclear eligibility and exclusion criteria. Unclear sample size calculation (likely underpowered). Unclear choice for statistical analysis.

HIGH

• risk of selection bias
• Unclear allocation concealment
• Possibly underpowered
• Per protocol analysis

Young, 2019

Definitely yes

Reason: numbered, sequential, sealed opaque envelopes, stratified in blocks of 2 and 4 by clinic.

Definitely yes

Reason: opaque sealed envelopes

Probably yes

Reason: patients were blinded. Caregiver collected baseline outcomes, while another clinician blinded to allocation collected follow-up outcomes

Definitely yes

Reason: due to short follow-up time (72 hours), no loss to follow-up occurred

Probably yes

Reason: no deviations from published protocol, yet protocol was registered after inclusion start

Probably no

Reason: trial stopped early due to exceeded effect sizes (larger than estimate)

LOW