Skip navigation

No evidence that complementary therapies work for tension-type and cervicogenic headache

Clinical bottom line

There is a lack of high quality trials providing evidence that complementary therapies including acupuncture, spinal manipulation, physiotherapy, electrostimulation, homeopathy, massage and therapeutic touch are of benefit for the treatment or prevention of tension-type or cervicogenic headache. The existing trials tended to be of poor methodological quality, and are therefore subject to bias.

Tension-type headache (TTH) is a common problem in Western societies with prevalence estimates ranging from 35 to 40% in adults. The International Headache Society (IHS) classification includes two types: headaches occurring less than 15 times per month are classified as episodic TTH, and headaches occurring more frequently that 15 per month are classified as chronic TTH. Cervicogenic headache (CH) is less common and is defined as pain referred to the head from the cervical spine. Increasingly, patients are seeking alternative/complementary therapies for the treatment of many painful conditions including headache. But how well do they work?

Systematic review:

Vernon H, McDermaid CS, Hagino C. Systematic review of randomised clinical trials of complementary/alternative therapies in the treatment of tension-type and cervicogenic headache. Complementary Therapies in Medicine. 1999; 7 142-155

  • Date review completed: 1998
  • Number of trials included: 24 randomised controlled trials
  • Number of patients:
  • Control groups: active controls including physiotherapy and acupuncture, and sham treatments (placebo)
  • Main outcomes: headache frequency and severity, headache index, medication use

Inclusion criteria were randomised controlled trials of complementary therapy treatment; non-migraneous headaches in adults; reporting on clinical outcomes; English language only.

Reviewers conducted searches of the main databases (MEDLINE) and references of retrieved reports though limited themselves to the English language only. Methodological quality of trials was assessed using an 18 point scale (van Tulder et al, 1997). The quality score of each trial was converted to a percentage where a rating of 0-40% indicated poor quality, 40-60% moderate quality and over 60% high quality. Reviewers provided a descriptive summary of all included reports and vote count of clinical outcome by type of intervention. Due to heterogeneity of trials, statistical pooling was not attempted.


Of the 24 included RCTs, 22 were of patients with TTH and two of CH. The trials were of the following complementary therapies: acupuncture, spinal manipulation, homeopathy, physiotherapy, electrotherapy, 'Tiger Balm' and therapeutic touch. The quality of the trials varied, but were generally of moderate quality. It was not possible to determine which aspects of methodological quality were high or low as reviewers only reported the total score. The reviewers presented very little information on the individual trials e.g. sensitivity of the trials, baseline values, adequacy of statistical testing, blinding status, and how and when outcomes were assessed. Therefore, it was difficult to verify their conclusions.


There were four studies with quality scores of 61 to 69% reporting on 99 patients. Treatment duration ranged from six to 12 weeks, three studies were sham controlled and one compared acupuncture with physiotherapy. Two trials reported a significant difference favouring acupuncture over placebo for headache frequency. However, these two trials were very small with a combined total of only 39 patients. One sham controlled study reported no significant difference between the two groups.

One trial reported 'significant improvement' in the acupuncture and physiotherapy groups with no difference between the two groups. There were four lower quality studies (44 to 50%) reporting on 173 patients. Average duration of treatment was three months. Three of the studies were positive reporting acupuncture was better than sham control for headache frequency, better than no treatment for headache severity and better than medication (type unspecified) for level of improvement. One study was negative and showed physiotherapy was better than acupuncture for improvement of symptoms.

Spinal manipulation

Reviewers reported on six RCTs, two of these had at least one group size less than 10 and are not discussed further here. Of the remaining four RCTs, two were of TTH and two of CH. The trials had quality scores of 64 to 80%. All four trials used active control groups only, no sham/placebo controls. These included: soft tissue mobilisation, brief rest, ice pack, amitriptyline and soft tissue therapy. Three of the four studies were positive showing spinal manipulation to be more effective than the active comparators. Since none of these therapies are of proven efficacy, this does not rule out a placebo effect and does not provide convincing evidence for the efficacy of spinal manipulation for TTH and CH. The results were of dubious validity due to small group sizes in two of the trials and unclear diagnostic criteria for CH in two trials.


Four RCTs investigated electrotherapy, three were of Transcutaneous Electrical Nerve Stimulation (TENS) and one of cranial electrotherapy. Three of the trials were sham/placebo controlled and one compared TENS with relaxation, biofeedback or both combined. Quality scores ranged from 39 to 61%. Overall three trials were positive and one negative, however, the methodological quality of the trials was low including mixed headache diagnoses, lack of blinding and small group sizes. Physiotherapy Three RCTs with quality scores of 33 to 58, reporting on 147 patients. Treatment duration ranged from two to three months.


treatment included relaxation, stretching, TENS, ice therapy and teaching control of muscle tension. One trial compared acupuncture with physiotherapy the result for this trial is presented in acupuncture section above. One study compared amitriptyline plus physiotherapy versus physiotherapy alone, and one compared physiotherapy plus biofeedback with attention control. Both studies claim positive benefit of physiotherapy treatment. It is difficult to make any sense of these studies as the treatment modalities were mixed in each trial thus is impossible to determine the effect of each individual treatment. The trials are methodologically flawed producing biased results.


One high quality (86%) trial that was both randomised and double blind compared individualised homeopathic treatment with placebo in 98 chronic TTH patients. After 12 weeks of treatment there was no difference between homeopathy and placebo on any outcome measure.

Tiger Balm

One RCT (quality 72%) compared 'Tiger Balm' topical ointment (analgesic/counter-irritant) with placebo or paracetamol 1000 mg in 57 patients with TTH. Both Tiger Balm and paracetamol were significantly better than placebo for pain relief up to three hours after a single treatment.

Therapeutic Touch

One RCT (quality 47%) compared therapeutic touch with sham touch in 60 patients. Twice as much pain relief was reported by real therapeutic touch patients as by the sham group immediately and four hours after the treatment.

Adverse effects

Only three of the included 24 RCTs made a statement about adverse effects. One patient had aggravation of pain from acupuncture needling, three patients complained of neck stiffness following spinal manipulation, 46 patients complained of dry mouth, drowsiness or weight gain following amitriptyline and 11 patients complained of irritation around the electrode site.



What this demonstrates is that there is a paucity of evidence from randomised trials that any of these interventions is truly effective.

Further reading

Van Tulder MW, Koes BW, Bouter LM. Conservative treatment of acute and chronic nonspecific low back pain. Spine. 1997; 22: 2128-2156.

Related topics

Acupuncture for headache

Homeopathy for headache

Cervical spine manipulation and mobilisation for neck pain and headache Therapeutic touch

Massage Physiotherapy