Human Models of the Headaches

Christina Kruuse

Shazia Khan Afridi

The unpredictable and episodic nature of many of the primary headaches introduces practical difficulties into research in this field, in imaging studies as well as in therapeutic intervention studies. The difficulties include getting patients to travel to the research center while suffering with severe pain or nausea and vomiting from migraine. This delays the onset of any therapy in interventional studies and introduces a selection bias in that those most debilitated do not participate in the study because they are unable to travel. It also means that the attack cannot be followed from the very onset. The pathophysiology and clinical manifestations of the headache may change throughout the attack; for example, in migraine there may be a premonitory phase prior to the headache or aura phase. These problems have fueled the search for good human headache models. Although animal models may take us far in understanding certain steps of the pathophysiology of primary headaches, human models are of great importance in studying the complete disease process from initiation of attack to treatment and pain relief.

The question then arises as to what makes a good headache model. One major issue in developing these models is to agree on the diagnoses of the patients studied. The introduction of the diagnostic criteria for headache disorders of the International Classification of Headache Disorders (ICHD-2) (1) has helped tremendously in the comparability and validation of the different models. Also, the systematic use of these criteria makes it possible to compare the distinct symptomatology of spontaneous headache to induced headaches to allow an objective diagnosis to be made. Unfortunately, not all of the previous headache models have used the ICHD criteria, which may limit comparison.

An effective headache model enables the headache to be followed in a controlled environment throughout the attack. The model needs to be reliable, reproducible, and validated. Safety is obviously an issue, and it is preferable to use a well-known compound that is easy to administer, has few reversible and tolerable side effects, and does not compromise any blinding of the studies. To perform studies in an outpatient clinic, it is essential to find a factor or compound which within short latency time induces the primary headache, in all its phases of premonitory symptoms, attack, and pain relief.

In this chapter we first consider human migraine models, an area that has been comparatively well studied. We then consider cluster headache (CH) models, some of which display similarities with certain migraine models. Finally, we describe the limited number of tension-type headache models currently available.

HUMAN MIGRAINE MODELS

Past experience has shown that besides the spontaneous attacks, migraine may be precipitated by a variety of factors, some of which are known to repeatedly induce a migraine attack. These have received interest as potential models of migraine (65). Induction of migraine in healthy subjects is only rarely reported and then often in subjects with a family history of migraine. Thus, so far developing a migraine model only seems feasible in migraine patients. The proposed mechanisms for migraine induction are given as well as other actions of the compounds tested.

Table 23-1 shows an overview of the effects of several compounds tested for use as migraine models in migraine patients.

Glyceryl Trinitrate

The nitric oxide (NO) donor glyceryl trinitrate (GTN), which is used to treat angina pectoris, has proved to be an excellent compound for use in a migraine model and the model is well validated. From its use in cardiology, it is known to have well-described mechanisms of actions; well-known, mild, reversible, and short-lasting side effects, headache and hypotension being the most predominant; is easy to administer; and has a short half-life (3).

TABLE 23-1 Compounds Tested as Potential Migraine-Inducing Agents in Migraine Patients

	*GTN*	*Histamine*	*Sildenafil*	*CGRP*	*Somatostatin Withdrawal*	*m-CPP*	*Fenfluramine*	*Reserpine*
Migraine-inducing mechanisms	NO donor	H₁ receptor	PDE5 inhibition	CGRP receptor	Pronociceptive	5-HT_1A,2 agonist	Decrease brain serotonin	Serotonin release
Main second messenger mechanisms	↑ cGMP	↑ cGMP by NOS activation	↑ cGMP	↑ cAMP, (↑ cGMP)^c	Not clarified	Not clarified	Not clarified	Not clarified
Other mechanisms of action	May release CGRP^d	H₂ receptor ↑ cAMP	PDE1 and PDE6 inhibition	May activate NOS	Changes in growth hormone and insulin	Interacts with alpha₂-adrenergic receptor	Reduce catecholamines	Monoamine release in general
Placebo control	Yes	No	Yes	Yes	No	Yes	Yes	Yes
Tolerability	Good	Good	Good	Relative	Good	Relative	Relative	Poor
Reproducibility	Good	ND	ND	ND	ND	Not good	ND	ND
Limitations		Flushing		Hypotension Flushing	Not effective	Variable oral bioavailability	Not to be used. Serious side effects.	Long-lasting side effect Flushing
Effects on:
CBF	ND	No change	No change	No change	ND	ND	ND	ND
Cerebral artery diameter	↑	↑	No change	↑	ND	ND	ND	ND
Extra cerebral artery diameter	↑	↑	No change	ND	ND	ND	ND	ND
Flushing	Minor	Major	Minor	Major	No reports	No reports	No reports	Major
Blood pressure	↓	↓	No change	↓	ND	↑	ND	No change
Heart rate	↑	↑	↑	↑	ND	↑	ND	ND
Rate of migraine induction^a	˜80%	˜70%	˜80%	˜33%	none	53-90%^h	˜75%	60-90%
Time to attack^b (hours)	˜5.5^f	2.5-5^f	˜4.5^f	˜5^f	—	˜8-12^f	˜4.5^g	˜6^f
	˜2.5-4.7^e		˜3.8^e			˜6^g		˜2^g
References	(2,9,40,77,79)	(44,50,52,75,83)	(46)	(19,41,48,58)	(56)	(10,28,29,55,69)	(68,69,72,74)	(15,24,60,70)
^aThe ability to induce migraine in migraine patients is shown as rate of migraine induction; however, some studies did not use ICHD criteria when evaluating the induced migraine. ^bWhen evaluating time to attack, several endpoints are given in the various papers. A consensus should be made and perhaps time to fulfillment of ICHD-2 criteria or patient assessed migraine when rescue medication is used, would be the best endpoint to be recommended for use in the future. ^cIt is not clarified if and how CGRP may increase cGMP levels. ^dOnly shown in animals, no increase in plasma CGRP was found in healthy subjects after GTN (33). ^eTime to fulfillment of IHS criteria for migraine without aura. ^fTime to peak headache score; IHS’s criteria may have been fulfilled at an earlier time point; however, this is not given in the paper. ^gTime to onset of headache ^hIn predisposed individuals or migraine patients.
The effect of the compounds are given with the limitation that some effects have not yet been investigated (ND= not done). Abbreviations: ↓, decrease; ↑, increase; NOS, nitric oxide synthase; PDE, phosphodiesterase.

GTN has long been known to induce headache as a side effect, and has over the years been proposed for use as a diagnostic test in migraine (18,71). The headache-initiating effects seem to be mediated through liberation of NO and not other metabolites (38) (see Chapters 15 and 30 for more details). In 1987, Sicuteri et al. described the biphasic headache response in migraine patients and in healthy subjects with a family history of migraine (73). This study was followed up in several studies by the Copenhagen Group introducing GTN administration as a headache model using intravenous infusion of GTN in both healthy subjects, tension-type headache (TTH) and patients with migraine without aura (MO) (36,40,63,77).

The GTN-induced headache consists of an immediate headache dissimilar to usual migraine. This develops in almost all patients and in a large group of healthy subjects with more intensive pain in MO patients than in both TTH and healthy controls (Fig. 23-1). In contrast to healthy subjects approximately 8 of 10 MO patients develop delayed headache resembling their usual MO attack with usual lateralization and accompanying symptoms fulfilling ICHD-1 criteria for MO, with peak headache score approximately 5.5 hours after infusion (Fig. 23-2) (77). Only very few healthy subjects experience a delayed headache. The basic studies showed good reproducibility and a dose-dependent headache response with a ceiling effect and maximum headache score at 0.5 μg · kg⁻¹ · min⁻¹ GTN; thus, this dose was applied in subsequent studies (see Fig. 23-1) (63).

FIGURE 23-1. Immediate headache responses over time during infusion of four different GTN doses and postdose in MO patients (n = 17), TTH patients (n = 9), and healthy controls (n = 17). Headache was scored on a 10-point verbal scale, 1 being mild headache including a feeling of pressing or throbbing, 5 moderate headache, and 10 the worst headache imaginable. The most severe headache was reported in migraine patients and a ceiling effects on headache pain was found with infusion of 0.5 μg · kg⁻¹ · min⁻¹. (From Olesen et al. (63), with permission.)

FIGURE 23-2. Median headache intensity is shown over time during (filled circles) and after infusion (open circles) of 0.5 μg · kg⁻¹ · min⁻¹ GTN or placebo in 10 patients with MO. Headache was scored on a verbal scale from 0 to 10 (10 = worst). The cumulative number of patients fulfilling the ICHD-1 criteria’s is given above the curve. Eight patients fulfilled the criteria for MO with a peak headache score 5.5 hours after termination of GTN infusion. (From Thomsen et al. (77), with permission)

Only gold members can continue reading. Log In or Register to continue