World Health Organization (WHO) code and diagnosis: G44.03 Chronic paroxysmal hemicrania

Short description: Paroxysmal hemicrania (PH) sufferers have attacks with similar characteristics of pain and associated symptoms and signs to those of cluster headache (CH), but they are shorter lasting, are more frequent, occur more commonly in females, and respond absolutely to indomethacin. In episodic paroxysmal hemicrania (EPH), attacks occur in periods lasting 7 days to 1 year separated by pain-free periods lasting 1 month or longer. In chronic paroxysmal hemicrania (CPH), attacks occur for more than 1 year without remission or with remissions lasting less than 1 month (21).

Other terms: Sjaastad syndrome

PH is a relatively rare disorder. Data on PH epidemiology are scarce, and many cases are probably still overlooked. Following the original description (58), new patients have been identified in several countries (1,25,57) and isolated PH cases are no longer reported. The incidence and prevalence of PH are not known, but the relative frequency compared to CH is reported to be approximately 1 to 3% (1). Prevalence estimates in CH vary from 56 per 100,000 to 381 per 100,000 (51,61,69). It has been reported that the prevalence of CH in the United States, based on one large study, was 401 per 100,000 (51). However, this was a U.S. study of CH incidence, not prevalence, and converting incidence to prevalence estimates is naturally combined with uncertainties because survival time needs to be conjectured (61,68).

There is no evidence of a genetic etiology in PH. Parents or siblings do not have an increased incidence of CH or migraine compared with the general population (1).

The pathogenesis of PH remains unknown. The unilaterality and pain intensity, as well as the autonomic symptoms and signs, suggest that some mechanisms may be shared with the other trigeminal autonomic cephalalgias (TACs) listed in group 3 of the 2004 IHS headache classification. The absolute indomethacin effect, however, indicates that PH has a distinct pathophysiology. In most of the cases no underlying pathology is detected by any supplementary investigation. Several secondary cases, however, have been described (70) (Table 97-1).

Although the pain in PH is strictly unilateral, autonomic involvement and ocular signs may be bilateral, albeit more pronounced on the symptomatic side. The existence of bilateral ocular signs seems to exclude the possibility of a single lesion of the peripheral nervous system being the cause of PH. A central lesion involving midline structures seems therefore more probable (e.g., hypothalamus, cavernous sinus) (40,56). Functional neuroimaging studies have shown specific activation of hypothalamic areas in some TACs, including CH and SUNCT (short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing) syndrome (35,36,67). Similar neuroimaging findings in PH are not available.

The pain in PH cannot be explained by a disturbance in the autonomic nervous system alone. This is based on the observation that autonomic signs may precede the development of pain in precipitated attacks and that the pharmacologic suppression of autonomic signs does not influence the pain pattern during attacks (66). Besides, the clinical picture does not fit with any classic sympathetic or parasympathetic syndrome. The pathophysiologic mechanisms responsible for the pain in PH remain unknown. It has been suggested that the ipsilateral distribution of trigeminal fibers in the trigeminovascular system may explain the unilaterality of the pain, and evidence of trigeminal-parasympathetic activation during PH attacks has emerged (18).

The ocular findings in PH patients have been studied in detail with dynamic tonometry (22,23). Pulse synchronous changes in intraocular pressure (IOP) may be recorded. They depend on the volume of the pulsatile part of ocular blood flow (23). The pulse synchronous changes in IOP are reflected by the corneal indentation pulse (CIP) amplitudes, which may be measured in micrometers by dynamic tonometry (22). The corresponding changes in intraocular volume in cubic milliliters can be estimated using conversion tables. In PH there is a significant attack-related increase in CIP amplitudes, ocular blood flow, and IOP that is bilateral but more pronounced on the symptomatic side (55). The attack-related increase in IOP cannot be solely explained by changes in aqueous humor dynamics because the increased volume exceeds turnover of aqueous humor and the IOP changes occur so rapidly (<30 seconds). Therefore, these findings are probably the result of an acute vasodilation that increases intraocular volume due to a neurogenic impulse and vasoactive neuropeptide release (55).

Since changes in IOP and CIP may indicate local abnormal hemodynamics, spontaneous PH attacks were studied with transcranial Doppler (54). During attacks, there was hyperventilation, reduction of the end-tidal PCO2, and decrease in blood flow velocity in all insonated arteries, on both symptomatic and nonsymptomatic sides. In the anterior cerebral artery, the reduction was significantly smaller on the symptomatic side. There was no major difference between patients and controls considering reduction of flow during hypocapnia, indicating that vascular reactivity in PH is normal.

Corneal temperature has been found to be increased on the symptomatic side during attacks, a finding that may also be due to increased ocular blood flow following vasodilation (13).

The abruptness with which the accompanying autonomic signs occur in mechanically precipitated attacks suggests that they may be mediated by neurogenic impulses. This would involve neuropeptide-containing perivascular fibers of the trigeminovascular system with its connections to the cavernous sinus plexus and brainstem. The release of vasoactive peptides from sensory fibers, which run in close relationship to other autonomic fibers, may also lead to miosis, increased IOP, and other autonomic disturbances observed in PH (14,37). Calcitonin gene-related peptide (CGRP), a peptide released from trigeminal fibers, and vasoactive intestinal polypeptide (VIP), a parasympathetic peptide, have been found to be abnormally high during a PH attack (18). Values returned to basal levels following indomethacin treatment. This is similar to the peptides’ profile found in CH (17), suggesting that this disorder and PH may share pathophysiologic traits to a certain extent. It is possible that the clinical picture in PH results from the interaction between neurotransmitters and neuromodulators released from sympathetic, parasympathetic, and sensory fibers at the frontal area and local autonomic and vascular mediators.

The ocular vascular findings (increased IOP, conjunctival injection) may also be explained by autonomic changes. The situation seems to be more intricate since experiments show that the IOP increase is inhibited by an α-blocking agent (thymoxamine) as well as satellite ganglion blockade (66). This observation, together with increased sweating and decreased salivation on the symptomatic side during PH attacks, suggests sympathetic stimulation (52,53). The sweat abnormality observed in some patients (53,64,65) suggests a direct sympathetic stimulation rather than a supersensitivity reaction, which has been found in CH (64). Increased tearing, nasal secretion, and miosis may, on the other hand, be due to a parasympathetic stimulation during attacks (52).

Heart rate and electrocardiogram (ECG) changes during attacks show no typical pattern and differ therefore from those observed during CH attacks (50). There is, however, a tendency to marked variations in heart rate in association with attacks of PH. Attack-related heart rhythm disturbances have been observed in some patients: bradycardia and sinoatrial block, bundle branch block with episodes of atrial fibrillation, multiple extrasystoles, and bradycardia (50). These findings may indicate a dysfunction in the central control of the autonomic nervous system during PH attacks.

Pain pressure threshold (PPT), the nociceptive flexion reflex (RIII), corneal reflex, and blink reflex have been studied in a few PH cases (2). The PPT and the subjective pain perception following sural nerve stimulation were reduced in PH, as was the RIII reflex threshold on the symptomatic side as compared to controls. Blink reflexes were normal, and the corneal reflex thresholds were reduced bilaterally, irrespective of indomethacin intake. Interestingly, the RIII threshold was not affected by indomethacin, but the subjective pain perception was significantly more asymmetric in PH on the drug than in controls.