Article Text

Download PDFPDF

Cranial autonomic symptoms in migraine: characteristics and comparison with cluster headache
  1. T-H Lai1,2,3,4,
  2. J-L Fuh1,3,
  3. S-J Wang1,3
  1. 1
    Department of Neurology, National Yang-Ming University, Taipei, Taiwan
  2. 2
    Institute of Neuroscience, National Yang-Ming University, Taipei, Taiwan
  3. 3
    Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan
  4. 4
    Department of Neurology, Mackay Memorial Hospital, Taipei, Taiwan
  1. Correspondence to Dr S-J Wang, Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan 11217; sjwang{at}


Background: Cranial autonomic symptoms (CAS) are distinguishing features of trigeminal autonomic cephalalgias, of which cluster headache (CH) is the most common, but they can occur in patients with migraine. For migraine with strictly unilateral headache, the presence of CAS might cause diagnostic confusion with CH. Characteristics of CAS in migraine and comparisons with those in CH have rarely been reported.

Methods: This study prospectively recruited consecutive patients with migraine and CH treated at a headache clinic. Six CAS items were surveyed, including: conjunctival injection, lacrimation, nasal congestion, rhinorrhoea, eyelid oedema and forehead/facial sweating. The CAS characteristics recorded included: laterality, intensity, time sequence and consistency with headache attacks.

Results: A total of 786 migraine patients (625 women/161 men, mean age 40 (13) years) and 98 CH patients (11 women/87 men, mean age 36 (11) years) were recruited. The prevalence of ⩾1 CAS in migraine patients was 56% and did not differ among migraine subtypes. Except for forehead/facial sweating, the features of the other CAS differed between patients with migraine and CH: CAS in migraine tended to be bilateral (OR 5.8–23.8 among different CAS), be unrestricted to the headache sides (OR 5.0–20.4), appear with mild to moderate intensity (OR 1.7–7.7) and be inconsistent with headache attacks (OR 2.8–6.7).

Conclusions: CAS were present in half of our migraine patients and the clinical features may help differentiate migraine from CH.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Cranial autonomic symptoms (CAS), which include conjunctival injection, lacrimation, nasal congestion, eyelid oedema, forehead/facial sweating, miosis and ptosis, are distinguishing features of trigeminal autonomic cephalalgias, of which cluster headache (CH) is the most common.1 These symptoms are also common in patients with migraine but their reported prevalence varies widely from 27% to 73% based on the study settings.2 3 4 Moreover, up to 17% of patients with migraine have “side locked” headache,5 and those with CAS can mimic CH clinically.6 7

Most studies addressing CAS in patients with migraine did not investigate the characteristics of CAS per se.2 3 4 The characteristics of CAS have never been systematically compared in migraine and CH. Identifying CAS features most predictive of migraine or CH is an important clinical issue. This study aimed to characterise CAS in patients with migraine and compare them with those in patients with CH.


From October 2005 to October 2007, this study prospectively recruited consecutive patients with migraine or CH from the headache clinic of Taipei-Veterans General Hospital (Taipei-VGH), a medical centre in Taiwan. Migraine and CH were diagnosed according to the International Classification of Headache Disorders, 2nd edition (codes 1.1, 1.2, 1.6 and 3.1) (ICHD-2).1 Chronic migraine was diagnosed based on the revised criteria (A1.5.1).8 The diagnoses were made during face to face interview with headache specialists. The study protocol was approved by the Institutional Review Board at Taipei-VGH. Each patient gave informed consent before entering the study.

Demographic and headache characteristic data were recorded for each patient using a structured headache intake form and a questionnaire for CAS during their first visit. Patients were asked to rate their headache intensity on a 0–10 verbal numerical scale. Six surveyed CAS items were conjunctival injection, lacrimation, nasal congestion, rhinorrhoea, eyelid oedema and forehead/facial sweating. Miosis and ptosis were excluded because of the poor reliability of self-report data. For each CAS item, the following details were queried: (1) time sequence in relation to headache attacks: before, during or after headache onset; (2) laterality: right, left, bilateral, ipsilateral, contralateral or unrelated to headache side; (3) intensity: grade 1 (mild) to 3 (severe) by a semiquantitative scale (table 1); and (4) consistency: always, often or seldom during headache attacks.

Table 1

Semiquantification of intensities of cranial autonomic symptoms


The SPSS software package V.15.0 for Windows (SPSS Inc, Chicago, Illinois, USA) was used for statistical analysis. The Student’s t test was used for continuous measures and χ2 or Fisher’s exact test was used for categorical data. The odds ratios (OR) for all CAS items were calculated for migraine and CH patients in those who had at least one CAS item. In this study, all p values were two tailed, and statistical significance was defined as p<0.01 because of the large sample size and multiple comparisons.


Study subjects

During the study period, a total of 786 patients (625 women/161 men, mean age 40.1 (12.9) years) with migraine and 98 patients (11 women/87 men, mean age 36.2 (10.5) years) with CH finished the study after excluding 19 patients (2%) because of data incompleteness. Diagnoses in the migraine group were episodic migraine without aura (n = 375, 48%), migraine with aura (n = 40, 5%), chronic migraine (n = 305, 39%) and probable migraine (n = 66, 8%). In the CH group, 97 (99%) patients had episodic CH and one (1%) had chronic CH. The low incidence of chronic CH is typical of this headache disorder in Asians.9

CAS in patients with migraine

In patients with migraine, 437 (56%) patients had at least one CAS item, which did not differ among different migraine subtypes (migraine without aura 57%, migraine with aura 48%, chronic migraine 57% and probable migraine 47%; p = 0.311). The most commonly reported CAS item was forehead/facial sweating (28.8%) followed by lacrimation (24.6%), while the least common was eyelid oedema (8.7%). Table 2 shows demographic data and headache characteristics for migraineurs with or without CAS. Compared with migraine patients without CAS, those with CAS had higher frequencies of severe intensity headache attacks, nausea, photophobia and phonophobia (p<0.001), and also a tendency towards vomiting (p = 0.010).

Table 2

Comparisons of demographic and headache profiles between migraine patients with or without cranial autonomic symptoms

Comparisons of CAS between migraine and CH patients

The following analyses included only patients with at least one CAS item (ie, 437 patients with migraine and 95 patients with CH). Figure 1A–E shows these comparisons. Because six different CAS were surveyed, the proportions or OR of the characteristics of the CAS are presented with ranges.

Figure 1

Characteristics of cranial autonomic symptoms (CAS) in patients with migraine (M) and cluster headache (CH) who had at least one CAS item. (A) Time sequence: onset of CAS with reference to headache. (B) Location: lateralities of CAS. (C) Location: lateralities of CAS with reference to headache. (D) Semiquantification of CAS intensities. (E) Frequency: CAS occurrence in relation to headache. Nasal cong., nasal congestion.

Frequencies of different CAS

Migraine patients with CAS had a lower number of CAS items than CH patients with CAS (1.8 (1.1) vs 3.5 (1.4); p<0.001) (table 3). In such case, patients with migraine had a lower incidence of all CAS items compared with CH patients except for forehead/facial sweating and eyelid oedema.

Table 3

Comparisons of cranial autonomic symptoms between patients with migraine and cluster headache with at least one cranial autonomic symptom item

Time sequence

CAS occurred most commonly during headache attacks (65–88% for different items of CAS for migraine and 80–96% for CH) rather than before or after headache attacks (fig 1A).


Most patients with migraine (67–95%) reported that their CAS were bilateral; however, except for forehead/facial sweating (67%), only small percentages of patients with CH (20–28%) reported bilateral symptoms (fig 1B). The OR of bilateral CAS for migraine versus CH ranged from 5.8 to 23.8, with the highest one for conjunctiva injection (23.8) followed by eyelid oedema (20.2). Most patients with CH (72–85%, except forehead/facial sweating) reported that their CAS was ipsilateral to their headache side while most patients with migraine (66–84%) reported no such association (fig 1C). The OR of unrestricted CAS (ie, not side locked to headache) between migraine and CH ranged from 5.0 to 20.4 with the highest for rhinorrhoea (20.4) followed by lacrimation (18.7). Among all CAS items, forehead/facial sweating was unique in that, in addition to migraine, more patients with CH reported this symptom to be bilateral (67%, other items of CAS 20–28%) and unrestricted to the headache side (67% CH, other of CAS items 14–26%).


Semiquantitatively, only 8–22% of patients with migraine considered their CAS severe compared with 22–55% of patients with CH (fig 1D). The OR of mild to moderate CAS between patients with migraine and CH ranged from 1.7 to 7.7. The most discrepant items were conjunctival injection (7.7) and lacrimation (7.0).


Less than one-third (14–27%) of patients with migraine always experienced CAS during headache attacks while half (44–52%) of patients with CH did (fig 1E). The OR of inconsistency (not always accompanied with each other) between patients with migraine and CH ranged from 2.8 to 6.7. The most discrepant CAS items were rhinorrhoea (6.7) and nasal congestion (4.0).

Sensitivity and specificity of combinations of CAS

To differentiate migraine with CAS from CH, different attributes and/or combinations of CAS characteristics were analysed (complete data not shown). The combination most predictive of migraine was bilateral CAS with either (1) mild to moderate intensity or (2) occurrence of CAS in the absence of headache (table 4). Lacrimation was the most applicable single attribute of CAS with the most affected patients (n = 283, 53.2%), highest positive predictive value (0.92) and specificity (86%), and second highest sensitivity (80%).

Table 4

Sensitivity, specificity, positive predictive value and negative predictive value of migraine with each item of cranial autonomic symptoms with combined diagnostic features*


In this study, about half of the migraine patients (56%) had CAS and the prevalence did not differ among migraine subtypes. Except for forehead/facial sweating, features of the CAS differed between patients with migraine and CH: CAS in migraine tended to be bilateral, less severe, unrelated to the headache side and less consistent with headache attacks. These clinical features of CAS may help clinicians to differentiate migraine patients with CAS from CH patients, especially in the cases of migraine with strictly unilateral headache.

Most of our patients with migraine had bilateral CAS (67–95%). Two previous studies investigated only unilateral CAS in their questionnaires2 3 while a third one reported that 32% of patients had strictly unilateral CAS, which was close to our results.4 In fact, bilateral CAS were not uncommon in patients with CH (20–67%) in the current study. A recent study showed that eyewitnesses reported that eight of nine patients with CH had bilateral conjunctival injection.10 Many physiological studies have revealed that the trigeminal autonomic reflex includes an often minor contralateral component,11 12 13 14 probably due to the crossover in the brainstem.14 Therefore, we believe that the incidence of bilateral CAS in patients with migraine and CH might be higher than thought previously. The reason why higher proportions of patients with CH had unilateral CAS is unknown. It could reflect different underlying pathomechanisms between CH and migraine. One hypothesis suggests that ipsilateral activation of the hypothalamus during cluster headache attacks may stimulate ipsilateral but simultaneously suppress contralateral trigeminal autonomic reflex.15 We also noted that, among all CAS items, bilaterality was the most common for forehead/facial sweating, in both migraine (95%) and CH (67%) patients. Among different CAS items, forehead/facial sweating is unique because it probably originates from both parasympathetic hyperfunction and sympathetic deficit.16 Whether this unique mechanism contributes to bilaterality awaits further study.

This study showed that severe headache and associated symptoms occurred more frequently in migraineurs with CAS than in those without. This finding is compatible with a previous report that associated symptoms correlate with increased headache intensity in migraine patients.17 Therefore, migraine with CAS may represent a more severe attack. It supports the hypothesis that there is a pain threshold above which CAS occur.18 An alternative possibility is that cranial parasympathetic outflow contributes to both pain and CAS in patients with migraine.14 16 19

The strengths of this study are its large sample size and detailed questionnaire regarding CAS characteristics. However, there were some limitations. Firstly, one may argue that it may be a circular argument in interpreting our results. The difference in the CAS features between CH and migraine might be confounded by the diagnostic criteria of the ICHD-2. Except for those patients with restlessness or agitation during attacks, unilateral headache associated with ipsilateral CAS is required for the diagnosis of CH but not migraine.1 Nevertheless, this study was intended to delineate the features of CAS in patients with migraine and, therefore, CAS in patients with CH could also be treated as a reference. Secondly, all CAS were self-reported and based on recollection. Bias of subjective recall is assumed (eg, frequency of eyelid oedema is likely to be underestimated). However, this study presented data from actual clinical practice, which may be more relevant to practitioners. Thirdly, interpretation and generalisation of our results should be cautious because the study was conducted in a headache clinic rather than in the general population.



  • See Editorial Commentary, p 1057

  • Funding The study was supported by grants from the Taipei-Veterans General Hospital (V96C1-041), Taipei, Taiwan.

  • Competing interests None.

  • Ethics approval Approved by the Institutional Review Board at Taipei-VGH.

Linked Articles