β-Adrenoceptor Blocking Drugs in Migraine Prophylaxis

Peer Tfelt-Hansen

Paul Rolan

β-Blockers (β-adrenoceptor antagonists) were introduced into medicine for the treatment of angina pectoris and cardiac arrhythmias, but they have subsequently proved valuable in many other conditions, including hypertension and migraine. The hypotensive action of these drugs was not predicted from their pharmacologic properties and was observed initially in patients receiving treatment for angina (43). Similarly, the value of propranolol in migraine was discovered in 1966 by Rabkin (45), who reported that migraine improved in a patient receiving propranolol for angina pectoris. Many controlled trials have since confirmed that propranolol is effective in the prophylaxis of migraine, and other β-blocking drugs—namely nadolol, metoprolol, atenolol, timolol, and bisoprolol—also have been demonstrated to be effective in the prophylaxis of migraine. In contrast, several β-blocking drugs with partial agonist activity—alprenolol, oxprenolol, pindolol, and acebutolol—have not been demonstrated to be effective in migraine prophylaxis. Several reviews of the prophylactic use of β-blocking drugs in the treatment of migraine have been published (3,31,32,51,58, 60,64).

The mode of action of β-blocking drugs in migraine remains to be elucidated, but the effective drugs are at present the drugs of first choice in migraine prophylaxis.

PHARMACOLOGIC BACKGROUND

The adrenergic receptors on which noradrenaline, the principal neurotransmitter at the peripheral sympathetic synapses, evokes responses have been classified as α and β types (1). α-Receptors are abundant in the resistance vessels of the skin, mucosa, and kidney and result in vasoconstriction when stimulated. β-Receptors have been subdivided into β₁-receptors, which predominate in the heart subserving myocardial excitation, and β₂-receptors in the arteries of the skeletal muscle and bronchi, subserving vasodilatation and bronchodilatation (30).

A β-blocking drug, for example, propranolol, which has equal affinity for β₁– and β₂-receptors, is described as being a nonselective drug, whereas agents such as metoprolol and atenolol have greater affinity for β₁– than for β₂-receptors and are examples of β₁-selective blocking drugs, even though the selectivity is not absolute. Propranolol is a pure antagonist and has no capacity to activate β-adrenoceptors. Several β-blocking drugs, for example, pindolol and acebutolol, activate β-receptors, but the intrinsic activities of these drugs are less than the full agonist, such as isoprenaline. These partial agonists are said to have intrinsic sympathomimetic activity (34). Some β-blocking drugs possess properties in addition to their effect in blocking β-receptors. Membrane-stabilizing activity has a direct effect on nervous tissue in the heart and is similar to the effect of local anesthetics. Membrane-stabilizing activity is present for some of these agents (Table 54-1).

There are considerable differences in the pharmacokinetics of the β-blocking drugs. The difference that could be important in the treatment of migraine might be expected to relate to the penetration of the drugs into the central nervous system. The entry of β-blocking drugs into the central nervous system depends on protein binding, ionization, and lipid solubility. Of these, the most important factor determining entry into the brain is lipid solubility, for which there are great differences in β-blocking drugs. Propranolol, alprenolol, oxprenolol, and metoprolol are extremely lipophilic and readily pass into the central nervous system. In contrast, atenolol is much more hydrophilic and passes into the central nervous system poorly (11).

For many years it has been suggested that 5-hydroxytryptamine (5-HT) may be involved in the development of a migraine attack, and several β-blocking drugs have substantial affinity for the 5-HT binding site in the
brain (see Table 54-1). Thus, alprenolol, oxprenolol, propranolol, and pindolol have high affinity for these binding sites, whereas atenolol has low affinity.

TABLE 54-1 Efficacy of β-Adrenoceptor Blockers in Migraine Prophylaxis and Their Properties

	*Efficacy in Migraine*	*Penetration into CNS*	*MSA*	*Cardioselective*	*PAA*	*Affinity for 5-HT in CNS*^a
Alprenolol	No	Yes	Yes	No	Yes	High
Oxprenolol	No	Yes	Yes	No	Yes	High
Propranolol	Yes	Yes	Yes	No	No	High
Pindolol	No	Yes	Yes	No	Yes	High
Nadolol	Yes	—	No	No	No	—
Timolol	Yes	Yes	No	No	No	—
Acebutolol	No	Yes	Yes	Yes	Yes	—
Atenolol	Yes	Poorly	No	Yes	No	Low
Metoprolol	Yes	Yes	No	Yes	No	—
Practolol	?^b	Poorly	No	Yes	Yes	—
Bisoprolol	Yes	—	—	Yes	No	—
^aAs judged from inhibition of specific ³H-5-HT binding to crude synaptic membrane from rats (35). ^bOnly an open trial reported efficacy of practolol before it was withdrawn due to side effects.
Abbreviations: MSA, membrane-stabilizing activity; PAA, partial agonist activity.

As clearly illustrated in Table 54-1 the fact that a β-blocking drug is effective in migraine prophylaxis does not depend on whether it penetrates easily into the central nervous system, is cardioselective, has membrane-stabilizing activity or binds to 5-HT sites in the brain. The only common property that the active β-blocking drugs have in common is the lack of partial agonist activity. It does seem that partial agonist activity prevents β-blocking drugs from exerting a beneficial effect in migraine prophylaxis, but the mechanism behind this remains obscure.

On the basis of one trial, it has even been questioned whether the efficacy of β-blocking drugs in migraine is related to blockade of β-adrenoceptors (55). It was reported that propranolol, in the clinically used racemic form d,l-propranolol, and d-propranolol, which has only a slight β-blocking effect, were significantly superior to placebo with no differences between the two forms of propranolol, indicating that an effect not related to β-blockade was an important factor in the action of propranolol in the prophylaxis of migraine. However, when the results were reanalyzed using conventional statistical methods, there was a significant effect for d,l-propranolol on the headache index, but no significant effect for d-propranolol compared with placebo (58). For headache days there were no difference between the three treatments. The trial thus did indicate that the β-blocking effect per se is important in migraine prophylaxis.

WHAT IS THE MODE OF ACTION OF β-BLOCKING DRUGS IN MIGRAINE?

It was originally hypothesized that β-blocking drugs were effective in migraine prophylaxis because they inhibit the vasodilatory phase of migraine. How this should be reconciled with the effectiveness of β-blocking drugs in migraine with aura, where a decrease in cerebral blood flow is present, is not clear. In a study on migraine with aura there was no increase in aura without headache (27), and one might presume that the preventive effect of β-blocking drugs must occur on the first phase of the attack, which clearly is initiated in the central nervous system.

There are some indications that β-blocking drugs exert their effect on the central catecholaminergic system. The contingent negative variation (CNV)—an event-related, slow, negative cerebral potential recorded over the scalp in simple reaction time tasks with warning stimulants—is significantly increased and its habituation reduced in untreated migraine patients in comparison with controls and tension-type headache sufferers. The CNV returned to normal values after migraine prophylaxis with β-blocking drugs (50). Furthermore, after 3 months of treatment with metoprolol or propranolol, it was shown that there was a significant correlation between CNV before treatment and the clinical response of β-blocking drugs: patients with higher CNVs tended to respond better to therapy. This indicates that in patients with central catecholaminergic hyperactivity, the chance of a positive response to β-blocking drugs in migraine prophylaxis is better and indirectly points to an effect in the central nervous system being responsible for the migraine prophylactic effect. In a recent study similar effects of β-blockers was found on intensity dependence of auditory evoked cortical potentials in migraine patients (49).

PHARMACOKINETICS OF β-BLOCKING DRUGS EFFECTIVE IN MIGRAINE

Propranolol is highly lipophilic and is well absorbed. Much of the drug is metabolized by the liver during its first
passage through the portal circulation, resulting in 25% bioavailability. There is great interindividual variation in the presystemic clearance of propranolol, resulting in enormous variability in plasma concentration after oral administration of the drug (approximately 20-fold). Propranolol is extensively metabolized, and one of the products of hepatic metabolism is 4-hydroxypropranolol, which possess some β-blocking effect. The half-life in plasma is about 4 hours but, as in hypertension, the drug is effective when administered twice daily. A sustained-release formulation of propranolol has been developed to maintain the concentration of propranolol in plasma over a 24-hour period (21).

Nadolol is hydrophilic and incompletely absorbed. The bioavailability is 35%, and interindividual variability in bioavailability is less than with propranolol. Nadolol is not extensively metabolized and is largely excreted intact in the urine. The half-life of nadolol in plasma is in the range of 12 to 20 hours, and it can consequently be administered once daily. Nadolol may accumulate in patients with renal failure.

Timolol is well absorbed and is subject to moderate first-pass metabolism, resulting in a bioavailability of 50%. It is metabolized by the liver, and the half-life in plasma is about 4 hours. It can be administered twice daily.

Metoprolol is well absorbed, but there is considerable first-pass metabolism, resulting in about 40% bioavailability. Its metabolism is subject to a genetic polymorphism with about 6% of the Caucasian population being poor metabolizers (26). As a result, plasma concentrations of the drug vary widely (up to 17-fold) and there is recent retrospective evidence that poor metabolizers may have a significantly higher adverse effect rate (67). The plasma half-life of metoprolol is 3 to 4 hours. In standard formulation it can be given twice daily. A sustained-released formulation of metoprolol has been developed to maintain the concentration in plasma over a 24-hour period; this formulation can be given once daily.

Atenolol is incompletely absorbed, but most of the absorbed drug reaches the systemic circulation, resulting in a bioavailability of 50%. There is relatively little variation in the plasma concentration of atenolol, with a four-fold range between patients. The drug is excreted largely unchanged in the urine, and the half-life in plasma is 5 to 8 hours. It can be given once daily and may accumulate in patients with renal failure.

Bisoprolol is 90% bioavailable and is eliminated by both renal and hepatic mechanisms. Its long half-life makes it suitable for administration once daily.

RESULTS OF CLINICAL TRIALS WITH β-BLOCKING DRUGS IN MIGRAINE

On the basis of controlled clinical trials in which a β-blocking drug was compared with a placebo, it can be concluded that propranolol, metoprolol, timolol, nadolol, atenolol, and bisoprolol have documented efficacy in migraine prophylaxis (3).

The main effect has been to reduce the frequency of attacks in patients with migraine with aura and without aura. In most trials, a mixed population of patients with migraine with aura and without aura have been included, but some trials (27,42,59) have studied these two forms of migraine separately and found similar results, as in the mixed patient populations. There is thus no reason to believe that the two forms of migraine respond differently to prophylaxis with β-blocking drugs.

In this chapter, only trials in which β-blocking drugs have been compared with placebo or in which two β-blocking drugs have been compared with each other are reviewed. In addition, for many years, propranolol has been the standard comparative drug for migraine prophylaxis and has been compared with several agents that are not β-blocking drugs. These trials are mentioned in the chapters concerning these agents, but generally the results have shown similar efficacy for the new drug and propranolol.

The crossover design comparing active drug against placebo has been used in most trials (Table 54-2). The blinding of the patients in these trials may be open to question because patients can often determine that they are on a β-blocking drug because of the pulse-slowing effect, particularly during effort. There are also negative trials with β-blocking drugs, however, suggesting that the blinding problem is not that great. In addition, it is reassuring that β-blocking drugs also have been found effective in the parallel group design, where the problem with blinding is less important.

β-Blocking Drugs Compared With Placebo

A number of double-blind controlled clinical trials comparing one β-blocking drug against placebo are summarized in Table 54-2. A total of 1,535 patients were recruited for the trials, and 83% of the patients completing the treatments were evaluable. By modern standards, many of the early studies can be criticized from a methodologic point of view. In many studies there are too few patients and the treatment periods are short (e.g., 4 to 6 weeks). However, the conclusion that propranolol is effective in migraine prophylaxis has been confirmed in more recent trials with better methodology. In addition, nadolol, timolol, metoprolol, atenolol, and bisoprolol have shown better efficacy than placebo in double-blind controlled clinical trials (see Table 54-2). However, in some trials propranolol failed to show a significant difference from placebo (2,22,56). It is most likely that the apparent lack of effect in these trials may be a statistical type 2 error (lack of power to detect the difference).

β-Blocking drugs possessing partial agonist activity, pindolol, alprenolol, oxprenolol, and acebutolol showed no significant difference from placebo in several early trials (see Table 54-2). It was suggested by an open study with practolol that partial agonist activity does not exclude efficacy in migraine prophylaxis because good results were obtained in 39 and 43 patients treated (48), but this result has never been confirmed in a double-blind trial. The fact that the β-blocking drugs with partial agonist activity were only studied in relatively small trials some time ago has led reviewers to state that the beneficial effect may have escaped detection (3). Because propranolol was found to be effective in similar small trials, the β-blocking drugs with partial agonist activity probably are not effective in migraine.

TABLE 54-2 Controlled Double-Blind Clinical Trials Comparing β-Blocking Drugs With Placebo in the Prophylaxis of Migraine

*Trial (Ref)*	*Drug, Dosage (mg)*	*Study Design*	*No. of Patients (No. Evaluated), Type of Migraine*	*Run in*	*Duration of Treatment*	*Factors Evaluated*	*Investigators’ Conclusion*
Weber and Reinmuth (63)	Prop 20 qid	CO	25 (19) MO, MA		3 mo × 2	“Symptomatic response”	Prop > Pl
Malvea et al. (35)	Prop ? mg	CO	31 (29) MO	30 days open	6 wk × 2	Preference, headache units per day, relief medication	Prop very effective in some patients
Widerø and Vigander (65)	Prop 40 qid	CO	45 (30) MO, MA (responders in pilot)	Open pilot study of Prop for 2-11 mo	3 mo × 2	Attack rate, preference	Prop > Pl
Børgesen et al. (7)	Prop 40 qid	CO	45 (30) MO, MA	2 wk no drug	12 wk × 2	Frequency, preference	Prop > Pl
Ludvigsson (33)	Prop 20/40 tid	CO	32 (28) children MO, MA		13 wk × 2	Frequency	Prop > Pl
Forssman et al. (16)	Prop 80 tid	CO	40 (32) MO, MA	10 wk no drug	12 wk × 2	Attack rate, headache days, “integrated headache,” relief medication	Prop > Pl
Diamond and Medina (12)	Prop 80-160/day	CO	83 (62) MO, MA		4-8 wk × 2	Preference, headache index, relief medication index	Prop > Pl
Holdorff et al. (22)	Prop 40 bid-tid	Pa	56 (36) MO, MA		12 wk	Migraine index, subjective rating	Prop = Pl
Nadelmann et al. (38)	Prop 20-80 qid	CO	64 (41) MO, MA	6 wk dose-finding	12 wk × 2	Headache unit, relief medication index	Prop > Pl
McDevitt (34)	Prop LA 160?	CO	38 (31) MO, MA		8 wk × 2	Frequency, severity, duration	PropLA > Pl
Pradalier et al. (44)	Prop LA 160	Pa	55 (41) MO	4 wk Pl	12 wk	Frequency	Prop LA > Pl
Al-Qassab and Findley (2)	Prop LA 80 Prop LA 160	CO	45 (30)	4 wk Pl	8 wk (1 wk wash-out)	Frequency, duration, severity	Prop LA 80 and Prop LA 160 vs. Pl NS
Sjaastad and Stenrud (52)	Pind 7.5-15 per day	CO	28 (24) MO, MA	3 wk no drug	3 wk × 2 (3 wk wash-out)	Headache index, headache days	Pind vs. Pl NS
Ekbom (13)	Alpren 200 bid	CO	33 (28) MO, MA		6 wk × 2 (1 wk wash-out)	Frequency, preference, headache index	Alpren vs. Pl NS
Ekbom and Zetterman (15)	Oxpren 80 tid	CO	34 (30) MO, MA		8 wk × 2 (1 wk wash-out)	Frequency, preference	Oxpren vs Pl NS
Nanda et al. (39)	Acebut	CO	43 (33) “migraine”	4 wk no drug	12 wk × 2 (4 wk wash-out)	Frequency	Acebut vs. Pl NS
Briggs and Millac (8)	Tim 10 bid	CO	14 MO, MA	4 wk no drug	6 wk × 4^a	Frequency, preference	Tim > Pl
Steiner et al. (53)	Tim 10 bid	CO	107 (94) MO, MA	4 wk Pl	8 wk × 2	Frequency, global preference	Tim > Pl
Ryan et al. (47)	Nad 80 od Nad 80 bid Nad 80 tid	Pa	80 (79) MO, MA	2 mo Pl	3 mo	Frequency, severity	Nad in all group > Pl
Andersson et al. (4)	Met LA 200 od	Pa	71 (62) MO, MA	4 wk no drug	8 wk	Frequency, migraine days, severity score, relief medication	Met LA > Pl
Kangasneimi et al. (27)	Met LA 200 od	CO	77 (74) MA?	4 wk no drug	8 wk × 2 (4 wk wash-out)	Frequency, migraine days, global duration, relief medication	Met LA > Pl
Steiner et al. (53)	Met 50-100 bid	Pa	59 (54) MO, MA	4 wk Pl	8 wk	Frequency, severity score, relief medication	Met vs Pl NS^b
Forssman et al. (17)	Aten 100 od	CO	24 (20) MO, MA	60 days no drug	90 days × 2 (2 wk wash-out)	Frequency, integrated headache, relief medication	Aten > Pl
Johannson et al. (25)	Aten 100 od	CO	72 (63) MO, MA	8 wk	12 wk × 2 (3 wk wash-out)	Integrated headache, migraine days	Aten > Pl
van de Ven et al. (61)	Bis 5 od Bis 10 od	Pa	226 (195) MO, MA	4 wk	12 wk	Frequency, attack duration	Bis 5 = Bis 10 > Pl
^aPatients crossed over twice, receiving timolol during two periods and placebo during two. ^bIn the initial double-blind 12-week treatment there was no difference between metoprolol and placebo; but in a further follow-up of 12 weeks, nonresponders to placebo or metoprolol switching to metoprolol 50 bid or metoprolol 100 mg bid, respectively, resulted in significant improvement.
Abbreviations: Acebut, acebutol; Alpren, alprenolol; Aten, atenolol; Met, metoprolol; Nad, nadolol; Oxpren, oxprenolol; Pind, pindolol; Prop, propranolol; Tim, timolol; Bis, bisoprolol; LA, long-acting, slow-release formulation; Pl, placebo; od, once daily; bid, twice daily; tid, three times daily; qid, four times daily; CO, crossover; Pa, parallel groups; MO, migraine without aura; MA, migraine with aura; NS, no statistically significant difference; >, more effective than.
Modified and extended from Andersson and Vinge (3).