Relationship Between Pain and Autonomic Phenomena in Headache and Other Pain Conditions

Abstract

Involvement of the (efferent) autonomic nervous system in the generation of pain is ongoing matter of debate. Based on clinical and experimental observations, there are good arguments that the sympathetic nervous system may be involved in pain following trauma, with and without nerve lesion, at an extremity, such as in complex regional pain syndrome type I and II. However, the mechanisms involved are in many cases still unclear. In various types of headache there is no convicing evidence that the sympathetic nervous system is involved in the generation of pain, although these pains may be accompanied by considerable autonomic reactions which are dependent on activity in sympatheitc neurons. Migraine and headaches with autonomic symptoms are accompanied by autonomic reactions which are dependent on activity in cranial parasympathetic neurons. Whether parasympathetic neurons innervating cranial blood vessels are involved in activation or sensitization of trigemino-vascular afferents is discussed and needs experimental verification.

The background

Under biological conditions, autonomic postganglionic sympathetic and parasympathetic neurones do not communicate with afferent neurones so as to generate sensations, in particular painful sensations; thus, they are not involved in the generation of pain. Autonomic systems (and in particular the sympathetics systems) are involved in mediating various protective reactions of the body that are associated with pain or impending pain. In fact, these autonomic (and neuroendocrine) protective reactions are included in the characterization of pain (1–4). Under certain pathophysiological conditions the sympathetic nervous system may be involved in the generation of pain, and pain being dependent on activity in sympathetic neurones, this is called sympathetically maintained pain (SMP) (5–8). SMP is a symptom that includes generically spontaneous pain and evoked pain (by mechanical and cold stimuli). It may be present in the complex regional pain syndromes (CRPS) type I and type II and in other neuropathic pain syndromes (5,9,11). The idea of the involvement of the (efferent) sympathetic nervous system in pain is based on various clinical observations that have been documented in the literature for decades (6,7,11–14). Representative of these multiple clinical observations are quantitative investigations of patients with CRPS I or II, who had SMP that could be shown to be alleviated by blockade of the sympathetic outflow to the affected extremity, taking the placebo problem and the control of the sympathetic blocks into account (15–18). The results obtained on patients with SMP imply: (i) that activity in sympathetic neurones can be involved in the generation of pain by coupling between sympathetic and primary afferent neurones; (ii) that excitation and/or sensitization of nociceptive afferent neurones is mediated chemically by noradrenaline released by the sympathetic postganglionic neurones; (iii) that the coupling between the sympathetic neurones and the afferent neurones is either direct or indirect (e.g. via blood vessels or other mediator cells); (iv) that the afferent neurones must express adrenoceptors if the coupling is direct; and (v) that sympathetically maintained activity in nociceptive neurones may generate a state of central sensitization/hyperexcitability, leading to spontaneous pain and secondary evoked pain (e.g. mechanical and cold allodynia evoked by stimulation of mechanoreceptive Aβ-fibres or myelinated cold-receptive fibres, respectively), or alternatively, that sympathetically generated activity in non-nociceptive afferent neurones triggers pain on the basis of central hyperexcitability generated otherwise, e.g. in CRPS II.

The experimental and less well-controlled clinical observations on patients with SMP are the basis for the belief that the (efferent) sympathetic nervous system may become involved in the generation of pain. They do reveal that activity in sympathetic neurones is important in the generation of SMP in these patients but not which mechanisms are involved. In fact, the mechanisms by which the sympathetic nervous system may actually be involved in pain may vary considerably between different groups of patients. Possible mechanisms that may operate in patients with SMP and the various behavioural, reduced in vivo and reduced in vitro animal models in which these mechanisms have been investigated have been extensively discussed (8,19–21).

The application to headache

When reasoning about the role of the autonomic nervous system (ANS) in the different types of headache one has to distinguish three different aspects (see also 22).

1. 1

   Is the ANS causally involved in the generation and maintenance of pain as it is supposed to be in CRPS I and II and various other types of neuropathic pain?

2. 2

   Are functional autonomic abnormalities that are associated with different types of headache the consequence of and therefore secondary to headache? This question addresses the observation that any pain is accompanied by autonomic reactions that are based on central reflex pathways in the neuraxis and on the central integration between nociceptive systems and autonomic systems. These autonomic reactions are primarily protective for the body under biological conditions, but not necessarily any longer under pathobiological conditions (see 2).

3. 3

   Are headache and functional autonomic abnormalities parallel events and therefore the consequence of some central abnormalities? It may be assumed that central abnormalities lead to both the abnormal pain and the abnormal autonomic reactions. In this case, it might be useful to investigate the autonomic disturbances in order to elucidate the central pathophysiological changes that may be common for headache and autonomic disturbances.

This view is important because it asks the question whether the pain – here the headache – is an expression of changes in the central nervous system which are reflected in sensory (painful and nonpainful), somatomotor, autonomic and other (e.g., neuroendocrine) changes. Recently the hypothesis was put forward, on the basis of experimental investigations on patients and animals and of clinical observations, that the complex regional pain syndrome (CRPS) type I is a disease of the central nervous system (7, 23).

Testing autonomic disorders

Diagnosis and management of autonomic disorders depend very much on the testing procedures used, as does the diagnosis of whether the ANS is involved in the generation and maintenance of different types of headache. The major aims of these testing procedures are: (i) to determine if an individual autonomic function is normal or abnormal; (ii) to assess the degree and site of dysfunctions; and (iii) to ascertain if the autonomic abnormality is primary or secondary (24).

Clinical procedures to assess autonomic deficits have been extensively described in the recent literature (see 24–28). These clinical autonomic test procedures should ideally fulfil the following criteria:

• they must be practical, noninvasive and reproducible,

• they should lead to results about the disturbance of one sympathetic or parasympathetic autonomic subsystem,

• they must be standardized in order to allow the comparison of results from different clinical institutions,

• they must be quantified,

• they must exclude deficient functioning of the effector organs.

Results obtained on patients with autonomic disturbances must be compared with age- and sex-matched control groups of healthy human beings in order to allow for interindividual variance and changes of autonomic functions with age.

Evidence that the autonomic nervous system is involved in headache

General aspects

The rationale to study functions of the ANS in headache is primarily based on clinical observations. Changes of ANS functions are obvious in cluster headache with autonomic symptoms (so-called trigeminal autonomic cephalgias [TACs]) (29–31) and have been extensively investigated for these pains in the last 20 years. Changes of ANS function are suspected to be important in migraine but are less obvious in this headache. Kruszewski et al. (32) have extensively discussed the literature on headache with autonomic symptoms.

In most studies small samples of patients have been investigated. Usually patients have been classified according to the diagnostic criteria of the International Headache Society (33) and groups of patients were paralleled by healthy sex- and age-matched control groups. Test–retest reliability has seldom been studied adequately and blinding has not been used. Most published studies focus on group differences in order to understand more about headache pathogenesis. The group differences have seldom been large, and it is generally not proven that ANS tests can help in diagnosing the headache disorder. Sensitivity and specificity is almost never discussed although some exceptions occur (34).

Migraine

Based on a literature review and an extensive investigation of patients with migraine, Thomsen, Olesen and coworkers conclude that ‘Clear dysfunction of the sympathetic nervous system remains to be shown. Mild parasympathetic hypofunction with denervation supersensitivity may be present in migraine’ (35, 36). Results are variable, however. Boiardi et al. (34) reported for instance that the diastolic blood pressure response to sustained handgrip was impaired in 61% of migraine patients. In a recent study, no heart-rate variability differences between migraine patients and control subjects were found (37). Micieli et al. (38) found increased basal pupillary diameter as well as increased light reflex contraction and dilatation velocities in migraineurs. Increased pupillary dilatation to phenylephrine eyedrops has also been found in migraine (39, 40).

Recently it has been proposed that activation parasympathetic neurones in the pterygopalatine and otic ganglia leads to vasodilation of cranial blood vessels, release of inflammatory mediators and activation and/or sensitization of trigeminovascular afferents (41, 42). Consistent with this idea appears to be that intranasal application of lidocaine may relieve migraine attacks (43, 44). It is suggested that this treatment blocks impulse transmission in the pterygopalatine ganglion. However, this idea of involvement of the efferent parasympathetic neurones in (direct or indirect) activation or sensitization of trigeminovascular afferents needs better experimental verification. Furthermore, relief of migraine attacks by intranasal lidocaine requires a double-blind, placebo-controlled approach. Finally, the latter results could probably also be explained by blockade of activity in trigeminovascular afferents passing through the pterygopalatine ganglion.

Cluster headache and other headaches with autonomic symptoms

A ‘Horner-like’ pupillary dysfunction has been found in cluster headache (45, 46). Fanciullacci et al. (46) found decreased responses to indirectly and directly acting sympathomimetic eyedrops. Salvesen et al. (47) found ipsilateral miosis, decreased mydriatic responses to directly acting drugs (tyramine, OH-amphetamine); however, they also found increased mydriatic responses to phenylephrine that they interpreted as denervation supersensitivity.

Sweating (evaporimetry, electrodermal responses) and vascular responses of the forehead skin have also been studied in cluster headache. The most interesting findings are ipsilateral reduced responses to body heating and exercise and increased (sweating) response to parenteral pilocarpine, similar to first neurone (central) Horner patients (denervation supersensitivity?) (48–50). However, both responses were enhanced in response to ocular stimulation (50). These findings are interpreted to mean that the ipsilateral forehead skin is sympathetically denervated and reinnervated by parasympathetic cholinergic fibres (which normally innervate the lacrimal glands) branching into the vacant sympathetic pathways (50). Finally, heart-rate variability may be slightly reduced in cluster headache (51); however, it is unclear whether results addressing disturbances of the autonomic regulation of the heart give information on autonomic abnormalities in cluster headache.

Overall, data regarding reliability, sensitivity and specificity are incompletely reported and the variation between subjects seems to be large. Tests for the different autonomic parameters should be investigated further and refined, with the aim of developing protocols that are helpful in the clinical situation and give reliable quantitative data.

Recently it has been proposed by Goadsby and Lipton (30) to classify cluster headache (episodic, chronic), paroxysmal hemicranias (episodic, chronic), short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT) and hemicrania continua under the generic term trigeminal autonomic cephalgia (TAC). This term is rather unfortunate because it implies that these types of headache are autonomic. A better and neutral term would be headache with autonomic symptoms. The reasons for this proposed reclassification are: (i) the clinical finding that these headaches are associated with autonomic symptoms, such as conjunctival injection, lachrymation, nasal congestion, rhinorhoea, ptosis or oedema; and (ii) the assumption that these headaches with autonomic symptoms share a common pathophysiology in which parasympathetic systems, projecting through facial cranial nerve VII and synapsing in the pterygopalatine (and otic) ganglion, are involved. Postganglionic neurones in these ganglia innervate lacrimal, palatine, pharyngeal and nasal glands as well as cerebral and extracerebral blood vessels. It is believed that these pre- and post-ganglionic parasympathetic neurones are reflexly activated during activation of trigeminovascular afferents, leading in this way to a correlation between the headache and the autonomic changes in patients with headaches with autonomic symptoms (TACs), and that ‘the trigeminofacial reflex is a key part of the pathophysiological expression of the [TAC] syndromes’ (30). The nature of the pathophysiology and whether reflex activation of the parasympathetic neurones innervating blood vessels can enhance the excitation or excitability of trigeminovascular afferents, thus being involved in their excitation and being an important component in a peripheral positive feedback circuit, awaits further experimentation and verification (31).

Future research

Many studies of autonomic functions in migraine and cluster headache concentrated on autonomic systems innervating specific target organs (e.g. heart, resistance vessels, sweat glands, pupils, etc.), which are anatomically and functionally not necessarily related to the hypothetical autonomic targets that may be the origin of the pain. Conclusions were drawn about hyper- or hypo-function of the sympathetic and the parasympathetic system. The implicit assumption is made that changes in the functioning of one autonomic subsystem allows conclusions to be made regarding the functioning of another autonomic subsystem. This assumption is most likely invalid because functionally different autonomic systems have distinct organizations in the periphery as well as in the neuraxis (52, 53). Furthermore, in many studies measurement of autonomic parameters are confounded by response characteristics of the effector organs.

In most (if not all) studies, no sound and testable hypothesis has been formulated about the involvement of the ANS in headache. Any hypothesis must start with the clinical observations, but must break down the putative mechanisms that are behind these observations into component parts that can be tested in animal models in vivo and in vitro (see 6, 21). Mechanisms of headache as such (e.g. migraine or cluster headache) cannot be tested in animal or human models. In analogy, CRPS I per se cannot be tested in an animal model but only its components, such as mechanism of SMP, swelling, trophic changes or motor deficits. Progress about the involvement of the ANS in the generation of headache can only be expected by basic research on animal models in interaction with research on human patients and with detailed investigations of the clinical phenomenologies. Furthermore, largely negative results as to the involvement of the ANS in headache (e.g. migraine; see 35, 36) does not exclude that specific ANSs that elude the testing procedures used are involved in the generation of pain. Finally it is possible that the ANS (e.g. the sympathetic system and the sympatho-adrenal system) is involved in the generation of pain, including different types of headache, by mechanisms that are entirely different from those hitherto described (see 19,54,55).