Clinical Characteristics and Burden of a Large Series with Cluster Headache From Turkey: A Cross-Sectional Study From Headache Centers


 Background: Our purpose was to investigate the demographics, diagnosis patterns, clinical characteristics, triggers, treatment experiences, and personal burden of patients with Cluster headache (CH) in Turkey, a country located between Europe and Asia.Methods: The study is a cross-sectional investigation based on data from eight headache centers in Turkey. All patients completed the semi-structured survey either face to face or by phone interview with a neurologist.Results: A total of 209 individuals with a mean age of 39.8 (11.3) completed the survey (176 males; 188 episodic, 21 chronic). The mean age at disease onset was 28.6 (10.2) years. The diagnostic delay was 4.9 years and misdiagnosis before CH was 57.9%. Of participants, 9.1% reported a positive family history for CH. Male patients with CH showed higher rates for being current smokers in comparison to females (59.7% vs. 24.2%; p<0.0001) and they also had significantly more past history of smoking at the time of first diagnosis (60.8% vs. 21.2%; p<0.0001). Females with CH had a previous diagnosis of migraine more frequently (57.6% vs. 27.3; p=0.001). Attack duration without treatment was significantly longer in female patients with CH compared to males (112 min vs. 87 min; p=0.029). Female participants had more migrainous features (57.6 % vs. 36.9%; p=0.033) and nausea/vomiting (48.5% vs. 30.1%; p=0.045) during their attacks. Only 42.1% of all participants reported satisfying treatment experiences. Of the participants, 85.9 % reported that oxygen was efficient for abortive treatment of CH; however, only 22 % of them had an oxygen tube at home. Female participants, as well as chronic CH patients, reported a higher likelihood of preventive treatment experiences. In this study, 49.3% of all participants appeared to be disabled by their headaches. Over one-quarter percent of our cohort reported that CH caused job-related burden.Conclusion: Remarkable diagnostic delay is an ongoing problem for CH and migraine was the most common misdiagnosis. Nearly half of the patients suffered from a burden of CH regardless of chronicity. Both past abortive and preventive treatment experiences of the participants highlight the insufficient efficacy of available choices and the necessity of more specific treatments for CH.


Introduction
Cluster headache (CH) is the most common form of the trigeminal autonomic cephalalgias and de ned by short lasting attacks of excruciating unilateral headache associated with ipsilateral autonomic features and/or restlessness or agitation [1]. The diagnosis of CH is rmly based on clinical history because of the lack of a diagnostic marker. The prevalence of CH is estimated at 0.5-3/1000, with male predominance [2]. Even the prevalence of CH is fairly rare compared with migraine, more than 500.000 individuals are probably experiencing this "suicidal" primary headache syndrome, in the United States of America (USA) alone [3]. The neurobiological mechanism underlying CH remains incompletely understood, so far. Hypothalamic activation along with secondary activation of the trigeminal-autonomic re ex is the leading hypothesis in CH pathophysiology [4][5][6][7].
The investigation of burden is important to detect not only for ictal but also for interictal consequences of the headache disorders. Therefore, individual problems may be targeted selectively and then it is much easier to act against them. So far, studies on headache burden mostly focus on migraine. Unfortunately, the substantial burden and consequences of living with CH have received less attention. Results of the Eurolight CH Project showed that the disease can have a huge and potentially irreversible impact on patients' lives even during interictal periods [8]. Considering the gender differences in the clinical presentation of CH, studies have reported that females with CH suffered from increased associated migrainous features, longer duration of untreated attacks, association with hormonal uctuations and tended to have a positive family history of migraine [9][10][11][12][13][14][15][16][17]. It is well-known that headache disorders show geographic and ethnic differences between Asian and Western countries. Studies from Asian population disclosed that CH patients had a stronger male predominance, lower rates of clinical presentation with restlessness, extremely rare aura rate, a lower circadian rhythmicity, and lower headache attack and bout frequencies, and rare presentation with the chronic form [10,[18][19][20][21][22][23].
There is still no published study about CH from Turkey, which has a unique geographical location in the intersection of Asia and Europe. Therefore we aimed to investigate the demographics, diagnosis patterns, clinical characteristics, triggers, treatment experiences and personal burden of CH patients in Turkey. The second purpose was to search for gender differences in CH. Lastly, episodic CH (ECH) patients and chronic CH (CCH) patients were compared to elaborate similarities and disparities between two forms.

Study population
The study is a cross-sectional investigation (performed between January and June 2020) based on data from eight headache centers in Turkey. Participants were recruited from the headache centers by two ways. First, patients diagnosed with CH were searched for retrospectively in the from records of the headache centers.
Then, they were invited by phone to participate in the study. One-hundred-sixty-eight patients with CH volunteered to participate into the study. Second, newly diagnosed patients with CH were also enrolled from the outpatients or emergency clinics of these centers during the recruitment period. Forty-one participants with episodic CH were enrolled into the study in that way. Eleven individuals rejected to participate in the study. We did not reach out to 21 patients with CH by theirs phones or emails. All patients were evaluated by an experienced headache specialists and their diagnoses of CH were checked according to the International Classi cation of Headache Disorders-3 criteria [1].

Inclusion and exclusion criteria
Inclusion criteria for the study were willing to participate in the study and being diagnosed with ECH or CCH by a headache expert. Exclusion criteria were diagnosis of secondary CH, unwillingness to participation, illiteracy, unstable medical and psychiatric condition. Informed consent was obtained from each participant following a detailed explanation of the aims of the study which was conducted in accordance with the ethical principles stated in the "Declaration of Helsinki". The study was approved by the Acıbadem University Ethics Committee.

Assessments
All patients completed the semi-structured survey either face to face or by phone interview with a physician, due to the restrictions after the pandemic. The survey was composed of 85 questions which addressed sociodemographic characteristics as well as clinical features, delay of diagnosis, triggers for attacks, treatment experiences and personal burden in CH (Appendix 1). Majority of the questions were adopted from the USA Cluster Headache Survey [14,24].

Statistical Analysis
No statistical calculation of power was performed prior to the study. The sample size was based on available data. All analyses were planned by authors PYD, BB and ME. For missing data, the percentages were calculated from valid cases. Normality of data was evaluated by using Shapiro Wilks test. Data expressed as mean (Standard deviation (SD)) and percentages (% In this study, the participants were enrolled from 8 headache centers located in ve different geographical regions in Turkey (Marmara, Aegean, Mid-Anatolian, Mediterranean, South-East Anatolian regions). Table 1 shows demographics, past and family history characteristics as well as, comorbidities and diagnostic issues of the main group as well as the subgroup comparisons in terms of gender and ECH vs CCH. In the main CH group, the mean age at rst diagnosis was 33.5 (11.1) years and the diagnostic delay before the correct diagnosis was 4.9 (6.3) years.

Neuro-radiological imaging
In the current study, all patients had brain magnetic resonance imaging (MRI) and 32.1% of them had at least more than one radiological imaging (3.6 ± 2, 2-16). Twenty-four patients (11.5%) had nonspeci c ndings in their brain MRIs regardless to their diagnoses. 5. Cluster headache treatment (Table 3)

Acute treatment experiences
The most common choice for acute treatment was oxygen (73.6%), followed by triptans (50%) and nonsteroidal anti-in ammatory drugs (NSAI) (47.6). Our study showed that female patients with CH had signi cantly higher use of triptans compared to male counterparts (68.8% vs. 46.6%; p = 0.033). Oxygen use was more often reported by CCH patients than episodic ones (91.3% vs. 71.4%; p = 0.045).

Preventive treatment experiences
Rates of past preventive treatment in male CH patients were lower than females (66.3% vs. 84.8%; p = 0.040).
The participants with CCH reported more common past preventive treatments compared to episodic ones  years, in Spain: 4.9 years, in Italy and East European countries: 5.3 6.4 years, in Denmark: 6.2-9 years, in the USA: 6.6-8.5 years, in Japan: 7.3 6.9 years) [20,[28][29][30][31][32][33][34][35]. In our study, diagnostic delay was 4.9 years. In Greece and Flanders, neurologists missed the diagnosis in 40% and 80% of the patients [34,36]. Even though a majority of our patients were diagnosed by a neurologist in Turkey, a correct initial diagnosis of CH occurred in 42.1% of them. Indeed, this rate was still higher than a previous large internet American survey (21%) [24]. Our nding was probably related to the fact that neurology was the mostly consulted specialty for headache disorders in Turkey because of the health care organization [37]. Our rate of family history for CH (9.1%) was higher than Eastern countries (0-6.7%) and more similar to Western countries (5-17%) [10,18,21,24,[38][39][40].
It is well-known that the rates of being a previous or current smoker were high in patients with CH, as 73-81% [14,26,41]. Our male CH patients had statistically higher rates of past and current smoking compared to females in line with the USA study [24]. The percentage of active smokers in the male CH patients was higher than the average rate of overall active smokers (29.3 %) in Turkey (2018) [42]. Current smokers had higher numbers of attacks with longer bouts than patients with CH who report never having smoked [26,41]. Although there is no strong evidence between quitting smoking and improvement of CH, smoking may enhance alcohol consumption and alcohol may trigger CH attack [41,43]. Therefore, it might be wise to advice to quitting smoking for CH patients. But even clinicians gave advice about quitting smoking, our CH patients seemed not to follow it. Thus more strong suggestions might be necessary for those patients.

Gender comparisons
Females with CH had a longer mean duration of untreated attacks than males (112.8 min vs 87.9 min), this nding was compatible with previous studies [44,45]. Migraine and CH have overlapping features that they share as different primary headache disorders. It was understandable but still interesting to note that females with CH are more frequently misdiagnosed as migraine [31]. Migraine was the leading misdiagnosis regardless of gender differences in Turkey, a pattern similar with the USA ndings (32.1% vs. 34%). An important confusing factor in misdiagnosis is the accompanying symptoms during attacks. There is a need for increased awareness, since CH patients can also experience the same accompanying symptoms well-known in migraine, as also seen in Table 2. Migrainous features and nausea/vomiting were frequently reported by females with CH in this study compatible with previous reports [14,15,25,44,46], explaining the increased misdiagnosis rate in women along with the well-known male dominance of CH.
We observed that female patients were statistically more likely to experience pain in the temple and in the ear compared to men, for unclari ed reasons.
Menstruation was cited as a trigger for CH attack in females similar to migraine, but with a low rate of 3.3% of females, in this study. Moreover, autonomic features can also occur in migraine, but usually bilaterally. Hence ± ± the occurrence of either ipsilateral or bilateral autonomic features needs to be carefully questioned in headache patients.

Clinical features
In our study, the most common cranial autonomic symptoms were lacrimation (79.9 %), followed by nasal congestion (55%), and agitation (55 %). In the USA study, lacrimation (91%) and nasal congestion (84%) were also the leading two autonomic symptoms reported in more frequent rates [24]. On the other hand, in Asian studies, lacrimation, conjunctival injection and rhinorrhea were the most common cranial autonomic symptoms [10,18,21,22]. In the USA study, men experienced more frequently lacrimation (92% vs. 88%, p = 0.03), while woman were more likely to experience nausea (41% vs. 34%, p = 0.03). In the Italian study, ptosis and nasal congestion were more prevalent in females [46]. In contrast to aforementioned studies, we did not see any statistical difference between two genders in regard to occurrence of autonomic symptoms. Moreover, we did not see any difference of these symptoms between episodic and chronic CH patients in contrast to previous studies [38,47].
It is worth to emphasize that the presence of aura is not particularly helpful in the differentiation between migraine and CH. Intriguingly, aura occurs in 14-23% of Western CH patients, but only <%1 of Asian patients [10,14,18,21,22,25,38,39,48]. Our nding of 26.8% with aura was pretty similar with the Western cohort studies. Agitation is also the most striking difference between migraine and CH, it was reported up to 93% of patients in the USA population. More than half of our participants (55%) reported a sense of restlessness or agitation during their attacks, remarkably.
January and February were the most frequently reported months of the year that cluster bouts would start in Turkey. Seasonal propensity has been reported partly discordant in studies, this might be related to geographical location of countries [10,22,24]. Seasonal changing, stress and alcohol were the most common triggers for attacks in our study. Sleep deprivation was more likely to be reported by chronic CH patients. The chronobiological features of CH have been extensively studied [13,49]. Higher risk was reported at 21.41, 02.02 and 06.23 [49]. However, the highest peak was during the afternoon in an Italian population [13]. In our cohort, 82.3% of the participants reported that they had the exact same time of the day for CH attacks and an increased risk peak was found at the night (57.7%). Many factors might be related to this timing like light exposition in different altitudes and different sociocultural habits [13]. In Western, Japan and Korean studies, nocturnal CH attacks were frequent (58-73%), whereas CH patients (65%) had both diurnal and nocturnal attacks in some Asian studies [10,20,22,24,38,40,47].

Treatment experiences
We noted that triptans were more widely preferred by our female patients with CH, partly explained by longer attack durations. In regard to the effectiveness of abortive treatment, only 42.1% of all participants reported satisfying treatment experience. CCH patients (33.3%) gave lower scores about effectiveness of acute therapies, as expected. But still the rates of oxygen use for attacks were statistically higher in CCH patients compared to episodic ones. Oxygen has been well-known as an acute treatment of CH since 1985 [50]. Studies have shown that oxygen therapy frequently was found to be effective by more than 75% of patients in both Western and Asian countries [19,21,[51][52][53][54]. In the current study, 85.9 % of all participants reported that oxygen was e cient for an abortive treatment of CH; however only 22 % of them had oxygen tube in their home. In our hospital-based population, only 57.2% of all participants remembered that they had a previous advice for having an oxygen tube at home. Moreover 11% of all participants had been using a nasal cannula instead of a non-breather mask during oxygen treatment and 33.1% of our cohort did not know about an exact oxygen treatment protocol for CH attacks [55]. These ndings may be related to insu cient patient education, di culty to obtain durable medical equipment of home oxygen because of insurers and unreluctance of patients to have this equipment in spite of enough suggestion and encouragement from clinicians.
Subcutaneous sumatriptan 6 mg has been shown to be effective as an abortive treatment of CH [55].
Zolmitriptan and sumatriptan spray can both be used as an alternative treatment of CH attacks, but they are not available in Turkey [55]. In the current study, e cacy of triptan treatment was reported by 39.2% of the participants. Indeed, females had statistically more frequently used triptans in their attacks and also reported to effectiveness of triptans more superior compared to males ( Thirdly, it is possible that coexisting migraine diagnosis in our cohort may create problems. This comorbid condition might blur some our results such as the presence of aura, associated symptoms, and triggers. But investigation of the les and the interviews were realized by headache experts and we tried to isolate CH ndings from migraine as far as we can. Fourth, it might be hard to precise conclusions regarding treatment experiences retrospectively without any established guide or previous consensus among the centers.
Nevertheless, the study has some obvious strengths. This is the rst large-sized multicenter study about CH from Turkey and our ndings were gathered on face-to-face or detailed phone interviews due to pandemic by experienced headache specialists. Moreover, we compared results of two genders and two forms of CH to get more detailed picture of this ominous disease.
In conclusion, remarkable diagnostic delay is an ongoing problem for CH and migraine was the most common misdiagnosis, especially for females with CH due to longer attacks and higher rates of associated symptoms.
Therefore females who have confounding features about a diagnosis of CH need to be examined in detail. In the treatment part, even though higher oxygen e cacy for attack treatment, only 22% of patients had oxygen tube in their homes. We think that the availability of oxygen tube may reduce ER utilization of the patients for abortive treatment. Finally, nearly half of the patients suffered from a personal burden of CH and at least onethird of them had job related burden in our country. Past treatment experiences of the patients underscore insu cient e cacy of available choices and need for more speci c abortive and preventive treatment options.

Declarations
Acknowledgements: The authors would like to express their gratitude to the patients with Cluster Headache who contributed the study. Funding: Non Availability of data and materials: The supporting data is available.
Ethical approval and consent to participate: Informed consent was obtained from each participant following a detailed explanation of the aims of the study which was conducted in accordance with the ethical principles stated in the "Declaration of Helsinki". The study was approved by the Acıbadem University Ethics Committee.