There were 33 randomised controlled trials (RCTs) of beta-blockers, 3 RCTs of tricyclics, 18 RCTs of calcium-channel blockers, 5 RCTs of anti-convulsants, 14 RCTs of non-steroidal anti-inflammatory drugs, 12 RCTs of serotonin-active drugs, 2 RCTs of SSRIs, 3 RCTs of anti-serotonin agents, and 6 RCTs of clonidine.
Beta-blockers: the ss of drugs studied and the number of trials involved are as follows:
Propranolol (18 RCTs) ss = 1.44; Atenolol (3 RCTs) ss = 2.33; Timolol (3 RCTs) ss = 3.00; Nadolol (1 RCT) ss = 1.00; Acebutolol (1 RCT) ss = -1.00; Pindolol (2 RCT) ss = -1.00; Oxprenolol (1 RCT) ss = -1.00; Alprenolol (1 RCT) ss = -1.00, Metoprolol (3 RCTs) ss = 4.33. Tricyclics: Amitryptyline (3 RCTs) ss = 2.33.
Calcium channel blockers: Verapamil (4 RCTs) ss = 1.00; Nimodipine (6 RCTs) ss = -0.83; Nifedipine (2 RCTs) ss = -0.5; Flunarizine (6 RCTs) ss = 2.17.
Non-Steroidal Anti-Inflammatory Agents: Naproxen (6 RCTs) ss = 2.17; Aspirin (2 RCTs) ss = -0.67; Indobufen (1 RCT) ss = 5.00; Fenoprofen (1 RCT) ss = 3.00; Ketoprofen (1 RCT) ss = 1.00; Flurbiprofen (1 RCT) ss = 5; Tolfenamic Acid (2 RCTs) ss = 5.00.
Anti-convulsants: Divalproex (4 RCTs) ss = 3.75; Carbamazepine (1 RCT) ss = 1.00.
SSRIs: Fluoxetine (2 RCTs) ss = 0.00.
Anti-serotonin agents: Methylsergide (3 RCTs) ss = 1.33.
Serotonin Active Drugs: Pizotifen (10 RCTs) ss = 1.10; Dihydroergotamine (2 RCTs) ss = 2.50.
Others: Clonidine (6 RCTs ) ss = 0.00.
38% of neurologists reported using Beta-blockers as the first line with 21% using this class of agents as a second line. Primary care physicians were in line with neurologists. Propranolol was the most popular beta-blocker. Tricyclic antidepressants and calcium channel blockers ranked second and third.
Overall, Spearman's rank correlation between neurologists' choice of first or second line of therapy and scientific proof of efficacy was strong (c) = 0.644, P = 0.018). Good correlation was found between primary care physicians' choice and scientific proof of efficacy (c) = 0.576, P = 0.050).
No correlation was found between physicians preference and cost of drugs (neurologist: r = -0.254, P = 0.450; primary care physician: r = -0.255, P = 0.455).