Skip navigation

Cervical spine manipulation and mobilisation for neck pain and headache

Clinical bottom line: Based on poor quality trials and diverse treatments studied, it is difficult to clearly assess benefits of manipulation and mobilisation. Based on poor quality trials, there does not appear to be any consistent benefit of manipulation for chronic neck pain when given alone or in conjunction with other treatments. There is a lack of evidence for manipulation for acute neck pain and for mobilisation for acute and chronic neck pain. There is a lack of evidence for manipulation and mobilisation for tension headache and migraine. In all cases, where benefit was seen, it was sporadic and of short duration, and therefore of questionable clinical relevance.

A number of different therapists carry out spinal manipulation and mobilisation for pain relief, including osteopaths, chiropractors and physiotherapists. In general, manipulation involves a high velocity thrust to a joint beyond its restricted range of movement, and mobilisation uses low velocity passive movements within or at the limit of the joint range.

Systematic review:

Hurwitz EL, Aker PD, Adams AH, Meeker WC, Shekelle PG, Barr JS Jr. Manipulation and mobilization of the cervical spine: A systematic review of the literature. Spine. 1996; 21(15): 1746-1760.

Date review completed: 1994

Number of trials included: 13 randomised controlled trials

Number of patients:

Control groups: active and inactive

Main outcomes: pain

Inclusion criteria were blinded and unblind studies of manipulation and mobilisation for cervical spine pain (e.g. neck pain, headache, whiplash, face pain, cervical angina); trials of complications of cervical spine manual therapy; acute (less than 3 weeks), subacute (3 to 13 weeks) and chronic (greater than 13 weeks ) pain; any language.

Reviewers pooled information from randomised controlled trials where appropriate, and provided descriptive summary of included studies. Pooled data were used to calculate effect size using a random effects model. This was assessed for clinical relevance by estimating what the pooled effect size represented on a 100 mm visual analogue scale. Trials involving manipulation and mobilisation were classified as manipulation trials.

Because pain is a subjective outcome, we have only considered randomised controlled trials in this summary. It was difficult to assess from the review the nature of blinding in trials.


Quality of included randomised controlled trials varied, but was generally low. Trial methodologies, included patients and control groups varied greatly. Interventions varied from one session to three months of up to twice weekly sessions. No statement was made about the appropriateness of statistical testing, which is of particular importance in these trials where many outcomes are measured over long periods of time - risk of false positives is therefore high. No statement was made about the levels of baseline pain or the power of these trials to detect a change - risk of false negatives is therefore high. Only five of the 14 trials made any attempt to blind patients/evaluators.

Neck pain: acute


No trials looked at manipulation.

Three trials were considered. One trial of 30 patients showed no statistical difference between mobilisation plus collar, transcutaneous electrical nerve stimulation plus collar and collar alone at 1, 3 and 12 weeks (cervical mobility and pain). However, all patients were allowed to consume analgesics. Two trials of quite intensive Maitland mobilisation for acute flexion-extension sprains showed better pain relief and range of movement compared with control. All patients consumed analgesics as needed. The first of these trials (61 patients) showed benefit at four and eight weeks, and also showed a similar benefit with patients receiving a neck collar and advice (control group, rest and analgesics). The second (170 patients) showed benefit at four and eight weeks compared with collar. No trials were found for mobilisation.

Neck pain: subacute and chronic

Five trials looked at manipulation. Data were pooled for 3 week pain outcomes for 155 of these patients. Effect size showed no benefit (0.42, 95% confidence interval -0.005 to 0.85), and the improvement seen represented approximately 10 mm on a 100 mm visual analogue scale. Of the included trials, the highest quality trial (100 patients) showed no benefit on a number of outcomes immediately after treatment, and a second, very small (9 patients), low quality trial showed benefit. Both compared a single treatment session of manipulation with mobilisation. Looking at longer term benefit, one trial with some blinding in 39 patients showed no benefit after a single session of manual therapy plus diazepam compared with diazepam alone (although a patient rating of 'helpfulness of treatment' showed preference for manipulation, this is not valid give the lack of adequate control). One trial of 52 patients showed no benefit of one to three sessions of manipulation plus azapropazone versus azapropazone only. The fifth trial in 64 patients compared up to three months of manipulation with exercise/massage over a 12 month period. There was a very modest benefit based on a measure of functional status 12 weeks only, and no differences in pain.

One trial of mobilisation in 63 patients was identified. Significant benefit of mobilisation plus salicylate was seen at one week post-treatment compared with salicylate plus massage and electrical stimulation and traction and compared with salicylate only. At three weeks only, there was greater cervical mobility.

Muscle tension headache

Three trials looked at manipulation In chronic headache sufferers. One trial of 126 patients compared twice weekly manipulation plus heat and light massage with daily amitriptyline, both given for six weeks (analgesics allowed as required). No differences were seen between groups immediately after treatment, but four weeks later, treatment benefits were more likely to persist with manipulation compared with amitriptyline. One trial of 19 patients compared two sessions of manipulation with cold packs for posttraumatic headache found significantly better pain relief at two weeks with manipulation, but not at five weeks. The third trial in 22 patients compared one session of manipulation with palpation and with supine rest. Benefit was seen at 5 minutes. No long term assessments were made.


One trial in 85 patients looked at two months of twice-weekly manipulation compared with mobilisation. At two months, manipulation given by chiropractors showed significantly less pain, greater reduction in frequency compared with mobilisation. However, physiotherapy manipulation was not better than mobilisation, and neither type of manipulation reduced mean frequency of attacks, mean duration of attacks or mean disability.

Adverse effects

Reviewers specifically included reports containing information on complications of manipulation and mobilisation. This information did not come from randomised controlled trials, and risk cannot therefore be quantified. Based on 118 case reports, reviewers note that most complications from manipulation involve vertebrobasilar accidents, including brain stem or cerebellar infarction, obstruction of the posterior inferior cerebellar artery, occlusion of basilar artery, spinal cord compression, vertebral fracture, tracheal rupture, diaphragm paralysis, internal carotid hematoma and cardiac arrest. Of these 118 cases, 21 patients died, 52 survived with serious neurological deficit, paralysis or permanent functional impairment. Work of others estimates risk at 1 in 40,000 manipulations for mild complication and 1 in 4000,000 to over 1 million for serious complications.



A bit of a mixed bag. A remarkable lack of evidence for interventions carried out so commonly.

Further reading

Related topics

Spinal manipulation and mobilisation for back pain