Abstract
A population-based sample of 6000 inhabitants of the city of Essen in Germany was screened using a standard questionnaire for possible cluster headache (CH). Fifty-six percent responded (N = 3336, 50.5% of them women, mean age 44.7 ± 12.7 years). All suspected cases (N = 182) were interviewed by a neurologist. Four subjects with CH (three men) were identified. The 1-year prevalence of CH was estimated to be 119/100 000 (95% confidence interval 3, 238/100 000).
Introduction
Cluster headache (CH) has been defined by the International Headache Society (IHS) as a strictly unilateral, severe or very severe retro-orbital or temporal headache, lasting 15–180 min and accompanied by one of the following ipsilateral symptoms: conjunctival injection, tearing, nasal conjunction or rhinorrhoea, eyelid oedema, forehead or facial sweating, miosis or ptosis, a sense of restlessness or agitation (1).
Data on the prevalence of CH are scarce and variable. No data from the German population are available yet.
The aim of the current study was to assess the prevalence of CH in a population-based sample in Germany. The study was embedded into the epidemiological survey of the German Headache Consortium investigating the prevalence of migraine, tension-type headache and chronic headache as well as that of CH in the population-based sample in Germany.
Methods
Questionnaire
The questionnaire was constructed based on the criteria of the IHS (1) and validated prior the survey. The results of the validation of the questionnaire are reported separately (2). Briefly, in a first step 278 patients from our Headache Outpatient Clinic with migraine (n = 97) or tension-type headache (n = 60), or a combination of migraine and tension-type headache (n = 23), as well as patients with CH (n = 98) and healthy subjects (n = 30) were asked to fill out the questionnaire twice, before and after 4 weeks. All of them were examined by a neurologist.
Phase 1: screening
Data for this study were collected during a baseline survey between December 2003 and November 2004. The study population comprised a random sample of 6000 inhabitants of the city of Essen, a town in the Region of North Rhine-Westphalia in the west part of Germany. The town covers an area of 210 360 km2 and has 585 481 inhabitants (305 726 women and 279 755 men). Inclusion criteria were: age 18–65 years and German citizenship (to ensure proper knowledge of the German language). The study was approved by the Ethics Committee of the University of Essen, Germany.
Figure 1 illustrates the screening procedure. Six thousand randomly selected subjects received a questionnaire via postal mail and, in the case of a non-response, a reminder 2 weeks later. Subjects who did not respond were called and asked for an interview by telephone performed by trained medical students based on the same questionnaire. After eight unsuccessful calls, subjects were considered non-responders. Individuals who refused the interview either by postal response or by phone were also considered non-responders.
The screening procedure to detect individuals with possible cluster headache.
The first question was: ‘have you had headache last year?’ The first screening step did not aim to diagnose but rather to screen for possible CH and therefore attempted to increase the sensitivity of the screening questions. Nine questions were asked on headache side, intensity and accompanying symptoms, but disregarded headache duration and frequency of attacks.
Cases of ‘possible cluster headache’ were predefined as fulfilling:
A: two of three criteria:
severe or very severe headache
strictly unilateral headache
retro-orbital or temporal headache
and
B: one of the following criteria:
ipsilateral conjunctival injection or tearing
ipsilateral nasal conjunction or rhinorrhoea
ipsilateral eyelid oedema
ipsilateral forehead or facial sweating
ipsilateral miosis or ptosis
a sense of restlessness or agitation.
Phase 2: neurological examination
Subjects fulfilling the criteria for ‘possible cluster headache’ were interviewed by phone by two experienced headache neurologists. In order to minimize between-examiner variability, both neurologists performed a semistructured interview based on the questionnaire. Subjects suspected to have CH after this interview were invited for a face-to-face examination.
Statistics
Comparison of interval scaled variables (age) was performed by using t-test, of ordinal scaled variables (gender) by using χ2 test. Crude prevalence of CH was expressed as the number of cases per 100 000 inhabitants. Ninety-five percent confidence intervals (CIs) were calculated as suggested previously (3). The level of significance was set at 0.05.
Results
Initial sample
The study sample comprised of 6000 subjects including 2971 (49.5%) men and 3029 (50.5%) women. The mean age was 44.7 ± 12.7 years. The distribution of gender and age was fairly similar to that of the population of the Region of North Rhine-Westphalia.
Responders
Two thousand and four subjects (33.4%) responded per mail, 1332 (22.2%) were interviewed by phone, 2181 (35.1%) could not be reached, 549 (9.2%) refused and 14 subjects (0.2%) were dead.
Responders were slightly older than non-responders (46.3 ± 12.7 vs. 42.8 ± 12.4 years, NS) and comprised more women (53.7% vs. 46.4%, NS).
Comparison of subjects who responded per mail with those interviewed by phone also revealed no significant differences (45.5 ± 12.7 vs. 47.5 ± 12.7 years, 55.4% vs. 51.3% women).
Suspected CH cases
One hundred and eighty-two subjects were identified as subjects with possible CH. All of them were interviewed by phone and 65 subjects were invited and examined by two experienced headache neurologists.
Confirmed CH cases
Table 1 demonstrates demographic and clinical characteristics of individuals with CH. Four individuals with CH were identified. Hence, the 1-year prevalence of CH was estimated to be 119/100 000 (95% CI 3, 238/100 000). Three of them were men. All presented typical clinical features of strictly unilateral, very severe headache attacks. None of them was correctly diagnosed as suffering from CH and none of them had received adequate acute or preventive treatment.
Demographic and clinical characteristics of individuals with cluster headache
VAS, Visual analogue scale.
Cases not confirmed
In 178 subjects, we rejected the suspected diagnosis of CH. One hundred and eighteen subjects had migraine, 40 had migraine and tension-type headache, 17 had tension-type and three subjects had trigeminal neuralgia. Table 2 summarizes the demographic and clinical characteristics of this population.
Demographic and clinical characteristics of individuals suspected of, but not having cluster headache
VAS, Visual analogue scale; TTH, tension-type headache; TN.
Discussion
The study achieved its main aim of estimating the prevalence of CH in the general population of Germany. We investigated 3336 subjects and identified four subjects with CH. This corresponds to the prevalence rate of 119 per 100 000 with the 95% CI of 3, 238 per 100 000.
A combination of postal mail and telephone interviews was used, achieving an acceptable response rate of 56%. Comparison of demography of responders and non-responders showed a trend toward a selection bias, which, however, is usual in population-based surveys. Comparing ‘postal mail’ vs. ‘telephone’ responders, no significant differences were found. Finally, the distribution of genders across different age groups was comparable to the data for the general population of the Region of North-Rhine Westphalia. Hence, we think that with a reasonable approximation the studied sample can be considered as representative of the general population of Germany.
A clear discrepancy between the questionnaire and physicians' diagnoses should be stressed. The screening questionnaire suspected CH in 182 subjects. We were able to confirm this in only four cases. This fact clearly indicates the use of our questionnaire (and probably any questionnaire) is quite limited in epidemiological surveys of CH. The vast majority of subjects were diagnosed to have migraine with trigeminal autonomic symptoms. This finding is in line with a previous report (4) as well as with our recent study demonstrating that 27% of patients with migraine had at least one unilateral autonomic symptom (5).
To the best of our knowledge, this is the first study to provide data on the prevalence of CH in the German population. Worldwide, few studies on the prevalence of CH are available, probably because of the rarity of this disorder. The Vaga study from Norway studied a population-based sample of 1838 subjects and found seven subjects with CH, estimating a prevalence rate of 326 per 100 000 (95% CI 120, 720 per 100 000) (6). Another population-based study in Copenhagen County investigated the prevalence of idiopathic headache syndromes in a sample of 1000 men and women aged between 25 and 64 years and found one case of CH (7). An Italian study screened 7522 patients registered in the lists of general practitioners and found 21 patients with CH, providing a prevalence rate of 279 per 100 000 (95% CI 120, 720 per 100 000) (8). A study from San Marino, Italy revealed a much lower prevalence rate of 69 per 100 000 (9). The follow-up study 14 years later revealed a prevalence rate of 56 per 100 000 (95% CI 31, 92 per 100 000) (10). The data, however, were obtained indirectly by scrutinizing the files of neurological, ear, nose and throat and eye disease services. Another population-based study in Ethiopia revealed a much lower prevalence of 30 per 100 000 (11). This difference could reflect true variations in the prevalence of CH throughout different geographical regions. On the other hand, it could be explained by methodological difficulties such as different language and cultural environments and, most importantly, by challenges in training of lay persons who carried out the study. A further study evaluated the prevalence of CH in a defined and circumscribed population: a Swedish study found the prevalence of CH to be 90 per 100 000 (95% CI 180, 240 per 100 000) in population of 18-year-old men (12).
The limitations of the study are: (i) the age of the screened population was limited to 18–65 years; (ii) response bias possibly resulting in an overestimation of prevalence; (iii) use of a questionnaire as a screening instrument; and (iv) the low number of identified cases with CH resulting in a broad CI.
Acknowledgements
Supported by the Federal Ministry of Education and Research (BMBF), Germany.
Appendix 1
Screening questions for cluster headache, translated from the German language:
Is this headache severe or very severe?
Does this headache always affect one side of the head?
Is this headache localized behind or above the eye?
Is this headache accompanied by any of the following symptoms?
One-sided injection of the eye or tearing (on the headache side)?
One-sided nasal conjunction or rhinorrhoea?
One-sided eyelid oedema (on the headache side)?
One-sided miosis or ptosis (on the headache side)?
One-sided forehead sweating (on the headache side)?
A sense of agitation or restlessness?