Seventeen behavioural studies:
Thermal biofeedback (BFB)n=5
Progressive muscle relaxation (PMR)n=5
Progressive muscle relaxation in combination with thermal biofeedback (PMR and BFB) n=4.
Multi-component packages n=3
Hypnosis n=1
Cognitive therapy, Autogenic therapy n=1
Autogenic training combined with progressive muscle relaxation (Augenic training and PMR) n=1
Psychological placebo and wait-list control ES n=8
24 studies of drug treatments were identified:
Propranolol n=2
Flunarizine n=8
Nimodipine n=3
Serotonergic drugs (Amitriptyline Trazodone, Cyproheptadine, Pizotifen, 5-HTP Dopaminergic drugs, Clonidine, Papaverine, Acetylsalicyclic Acid)n=29,
Drug Vs Placebo n=6
Behavioural intervention: All the non-pharmacological treatment modalities were shown to be superior to placebo or wait-list controls. Thermal BFB, and PMR in combination with BFB, were found to lead to better outcomes than the other modalities.
Pharmacological interventions: Contrasts between the frequently-evaluated drugs (calcium-channel blockers, serotonergic drugs, placebo) demonstrated better treatment outcome for all drugs than for the placebo, but no evidence for differential effectiveness of calcium-channel blocking drugs. There is some (limited) evidence to suggest that propanolol has superior effectiveness than placebo and other active prophylactic drugs.
Comparison of behavioural and drug treatment studies: Pharmological and non-pharmological control conditions were not associated with different outcomes. Thermal BFB, as well as PMR + BFB, led to more headache reduction in comparison to calcium-channel blockers and serotonergic drugs. PMR alone, multi component treatment programmes and both calcium-channel blockers as sertonergic drugs did not yield different treatment effects. Propanolol was not found to be significantly more efficacious than any other behavioural and drug intervention.
Between-group meta-analysis: In the second meta-analysis which excluded studies without a comparison group consisting of placebo or no-treatment control, the findings from the first review were only partially replicated. Individual a priori contrasts indicated that only BFB was associated with a significantly higher treatment success than serotonergic drugs (Z2 = 6.5, p<0.05) and PMR displayed a tendency to be more effective than calcium- channel blockers (Z2 = 2.0 p<0.09). There were no significant differences between multi-component behavioural treatment and all three types of prophylactic medications.