Abstract
Objective
To estimate the one-year prevalence of primary headaches, most importantly migraine and tension type headache, but also other primary headaches, in Estonia.
Methods
A population-based random sample of 2162 subjects in Tartu City and Tartu County were interviewed by telephone or face to face using a previously validated questionnaire.
Results
Of the 2162 contacted participants, 1215 (56%) fully completed the study. Of these, 502 (41.3%) reported headache during the previous year. The prevalences adjusted by weighting by age, gender, education, marital status and habitat were the following: All headaches 41.0%, all migraine 17.7%, all tension-type headache 18.0%, trigeminal autonomic cephalalgias 0.4%, other primary headaches 2.5%, and chronic headaches 2.7%.
Conclusion
The 1-year prevalences of primary headache disorders in Estonia are comparable to the previous findings in other European countries.
Keywords
Introduction
Primary headaches, namely tension-type headache and migraine, are both among the top six most prevalent diseases in the world and migraine is the second leading cause of years lived with disability globally (1). The epidemiologic data on these entities are steadily growing worldwide; however, in some regions, like in Eastern Europe, gaps are still considerable. The epidemiological data on other primary headaches is relatively scarce but the lifetime prevalence of cluster headache has been estimated to be 0.1–0.3% (2) and the lifetime prevalence of all other primary headache disorders combined has been reported to be 2.8% (3). Estonia is a high-income (4) Baltic country with a relatively small but well-defined population of approximately 1.3 million inhabitants. The prevalence of primary headache disorders in Estonia has never been studied.
We present here the results of a survey conducted in Estonia on a population-based adult sample. The goal was to estimate the one-year prevalence of primary headaches, most importantly migraine and tension-type headache, but also other primary headaches in Estonia.
Methods
Survey
The survey was conducted from January 2016 to May 2017. The preselected sample consisted of 3000 subjects aged 18–64 and was derived from the Estonian National Registry. It was a random sample of inhabitants of Tartu city and Tartu county, demographically representative of the Estonian population, and was stratified by gender, age, rural versus urban habitat and marital status. The survey was conducted by telephone or face to face by 14 trained medical students who used a structured questionnaire. If the subject could not be reached by phone on four different occasions, the interviewer visited his/her home address, and if the person was still not found, a note was left that contained a short introduction of the survey and a request to contact the study team.
Questionnaire
We used a structured questionnaire in Estonian or Russian that was previously developed by our study group and had undergone a specificity and sensitivity estimation for most primary headache disorders and also some secondary headache disorders; for example, medication overuse headache (6). The sensitivity and specificity estimates of the questionnaire were based on a validation study conducted in a specialized headache clinic. The validation study questionnaire was self-administered (6). The questionnaire consisted of four parts: a) Demographic data and lifestyle-related possible headache risk factors; b) the headache diagnostic questionnaire; c) headache-related burden and associated factors; and d) enquiry on socio-economic status and willingness to pay for effective headache treatment.
In the beginning of the diagnostic section, the respondent was asked a screening question for headaches: “During the last year, have you had repeated headaches that were not caused by an acute infection, medication side effects, medical procedures, or consumption of toxic substances including alcohol?” If the respondent answered “yes”, they were introduced to questions targeting different aspects of their headache (localisation, laterality, character, intensity, preceding and accompanying symptoms, duration, frequency, response to indomethacin, association with certain situations and activities, precipitating factors, drug consumption, and history of head trauma).
The questions were based on the ICHD 3-beta (5) diagnostic criteria. According to the data acquired from the questionnaire, the diagnosis of a headache disorder was made by a neurologist headache specialist applying the ICHD 3-beta criteria.
In this study, we aimed to evaluate the prevalences of the following entities: Episodic and chronic migraine, episodic and chronic tension-type headache, chronic daily headaches (headache on more than 15 days a month), trigeminal autonomic cephalalgias and other primary headaches except for primary thunderclap headache. We omitted primary thunderclap headache because this diagnosis requires an extensive diagnostic workup in order to exclude secondary causes and thus the prevalence cannot be evaluated by a questionnaire only. The headache had to fit either the definite or probable criteria of ICHD-3 beta to be considered as a case. The criteria were first applied for definite migraine, then definite tension-type headache, then probable migraine, then probable tension-type headache, then definite trigeminal autonomic cephalagias or definite other primary headaches and finally probable trigeminal autonomic cephalagias and probable other primary headaches. If a case did not fit any entity as a probable or definite primary headache syndrome, it was labelled as unidentifiable.
The specificities and sensitivities for these diagnostic entities are reported elsewhere (6). Shortly, the sensitivities and specificities for most prevalent headache diagnoses, namely migraine and tension-type headache, were 90% and 80% for migraine and 92% and 60% for tension-type headache, respectively.
If a respondent had more than one type of headache, they were asked to describe up to three most bothersome ones.
The headache-related burden and associated factors were assessed using the headache impact test, HIT-6 (7), and insomnia and depression assessment questionnaires (8). The respondents were also asked about their household income per capita. Their willingness to pay for treatment was assessed by a bidding game method (9). The results of these enquiries will be published elsewhere.
Statistical analyses
The main outcome variables of the study were the one-year prevalences of primary headaches in Estonia. Data analysis was performed using R (10). Sample weights were calculated using the ANES raking algorithm implemented in R package anesrake (11).
Ethics
The study was approved by the Research Ethics Committee of University of Tartu (permission no. 252T-15). All the respondents were informed of the purpose of the study and gave their written informed consent for participation.
Results
Of the 3000 preselected subjects, 838 (27.9%) had insufficient contact data. Of the contactable sample of 2162 subjects, 919 (43.2%) refused and 1243 (56.8%) respondents consented to participate in the study. Of those 1243 consenting subjects, 28 had missing data or gave unusable answers, so the participating sample consisted of 1215 subjects (Figure 1). Hence, applying the recommendations for the methodology of population surveys of headache prevalence from the Global Campaign against Headache (12), the participation rate of our study was 56%.
Flowchart of study sample.
Comparison of Estonian population and survey sample for distribution of gender, age, marital status, education and habitat.
Weighted one-year prevalences of primary headaches in Estonia (weighted by age, gender, marital status, habitat and education).
TTH: tension-type headache. The bold font is to make it visually easier to follow the table – the statistical significance does not apply here.
Migraine was reported by 233 respondents and tension-type headache by 228 respondents, and the weighted one-year prevalences were 17.7% and 18.0% respectively. Definite episodic migraine was diagnosed in 93 (weighted one-year prevalence of 6.6%) and probable episodic migraine in 128 respondents (weighted one-year prevalence of 10.2%). Chronic migraine was diagnosed in 12 respondents (weighted one-year prevalence 0.9%). Definite episodic tension-type headache was diagnosed in 154 (11.8%) and probable episodic tension-type headache was diagnosed in 60 (4.7%) respondents. Chronic tension-type headache was reported by 15 participants (1.5%). Distribution of all migraine and all tension-type headache by age and gender are shown in Figures 2 and 3 respectively.
Distribution of all migraine by age and gender. Distribution of all tension-type headache by age and gender.
Weighted one-year prevalences of identified other primary headaches in Estonia (weighted by age, gender, marital status, habitat and education).
Discussion
Here we present the estimates of one-year prevalences of primary headache disorders in the Estonian adult population. This is the first headache prevalence study conducted in this country.
The general one-year prevalence of headache in Estonia (41%) is somewhat lower than the mean prevalence in Europe (53%) (14). It is comparable to the headache prevalences in Georgia (15) and Italy (16) and higher than the prevalences in Sweden (17), Greece (18) and France (19).
The one-year prevalence of both probable and definite episodic migraine combined in Estonia (16.8%) is comparable to the mean prevalence of migraine in Europe (14.7%) (14), being closest to the respective one-year prevalences in France (19), Croatia (20) and the Netherlands (21).
There are extreme variations in the prevalences of episodic tension-type headache across regions and cultures (22). For example, the mean one-year prevalence of episodic tension-type headache in Europe is 62.6% (14), but it is estimated to be 7% in Africa (22) and about 30% in the Americas (23). The one-year prevalence of episodic tension-type headache in Estonia (16.5%) appears to be about three times lower than the mean prevalence in Europe. Of course, the variation between other European countries is large too – ranging from 30.9% in Russia (24) to 86.5% in Denmark (25). The reason for this could be the differences in methodological approaches across studies (questionnaire vs. personal interview, different time frames, etc), diagnostic overlap with probable migraine, differences in headache awareness, socioeconomic situation across countries, or unknown genetic or environmental factors (22). We speculate that the relatively low prevalence of episodic tension-type headache in our study could be mainly due to two factors. Firstly, in Estonia it is not customary to complain about milder pain or headaches nor to consider them as diseases or noteworthy health issues, thus the infrequent or subtler forms of tension-type headache might go unnoticed or unreported. Additionally, in the case of milder headaches, as tension-type headaches generally are, there is always the problem of recall bias and thus infrequent faint headaches might simply not be remembered over the period of a year.
A possible set of specific reasons for this low prevalence of episodic tension-type headache in our study are methodological. As stated in the methods section previously, we included both definite and probable categories for both migraine and tension-type headache. If a case fulfilled both criteria for migraine and tension-type headache, the former was favoured over the latter. This means that, in cases of doubt, the prevalence of tension-type headache could be slightly underestimated in favor of the prevalence of migraine. Another methodological aspect that is generally accepted as a large problem in epidemiological studies is the occurence of multiple headache types in the same subject. Respondents tend to report the most bothersome headache, despite being offered the opportunity to describe more than one. Thus, respondents who have comorbid migraine with tension-type headache might report the former and omit the latter. This is also supported by the data from our study, since the proportion of participants reporting more than one type of headache was small (6.8%) and the real comorbidity after the cases had been diagnosed was half that figure. As under-reporting is not evident in clinical experience, it again points to the possibility of underestimation of milder comorbid headache disorders, most prominently infrequent episodic tension-type headache. Similar problems with underestimating the prevalence of episodic tension-type headache in the case of comorbid migraine has been reported before (26).
Lastly, one of the possible methodological reasons for underestimating the prevalence of episodic tension-type headache is the limitations related to the screening question used in our study, which will be discussed further on in this section.
Aside from the episodic tension-type headache, the prevalence of chronic tension-type headache in Estonia (1.5%) is comparable to that of Europe (3.3%) (14), being closest to Denmark (0.9%) (27) and Germany (1.3%) (28).
The mean one-year prevalence of chronic daily headache (headache on more than 15 days per month) has been reported to be around 4% in Europe (14); however, the prevalences vary across regions, being somewhat lower in Western Europe (19,29–31) and higher in Eastern Europe (15,24,32). Our finding of the one-year prevalence of chronic daily headache of 2.7% is closer to the Western European prevalences.
Trigeminal autonomic cephalalgias are rare and therefore it is impossible to validate a questionnaire with sufficient power to make a reliable diagnosis by that alone. However, there were three cases in our sample in which the reported headaches fulfilled the definite or probable criteria for trigeminal autonomic cephalalgias, corresponding to a one-year prevalence of 0.4%, which is comparable to the previous studies on the prevalence of cluster headache in European countries (14,33–37).
The same problem of rarity and lack of a reliably validated questionnaire exists when it comes to other primary headaches. Data on the prevalences of these entities is even more scarce. A prospective population-based study in Northern Italy, the Bruneck Study, estimated the lifetime prevalence of all other primary headaches combined to be 2.8% (3). This is in concordance with our finding – 2.5% of our participants reported different headaches that did not fulfil ICHD 3-beta criteria for either migraine or tension-type headache but did for either the definite or probable criteria for the entities described in section 4 of the ICHD 3-beta.
Slight differences in the prevalences found here compared to the previous studies could have occurred due to the fact that we used the ICHD 3-beta criteria to diagnose the cases whereas the previous studies used the ICHD 2.
This study has several limitations.
Firstly, the participation rate of our study was 56%. This is a moderate reponse rate (12). We speculate that in addition to Estonian people being rather reserved and conservative when it comes to communication on health-related topics, one of the reasons for refusing to participate in the study is the limited knowledge about headaches and their impact in the general population. Hence, the problem is underappreciated and people are not motivated enough to participate in such a study. Another reason for refusal to participate, particularly over the phone, may be the negative influence of advertisement by phone which is rather prevalent in Estonia. It is thus possible that people are put off by surveys that are carried out over the telephone. Quite a large proportion of the preselected sample (27.9%) could not be reached because of insufficient contact data. This could partly be due to the fact that the Estonian National Registry was last updated in 2011 when the latest census was conducted, while our survey started in 2016. This leaves a time gap of five years during which a proportion of contact data had inevitably expired. It is also possible that of the 838 people whom we considered as having insufficient contact data, a number would actually have qualified as non-responders either because they did not want to answer a phone call from an unknown number, or because they chose not to contact the study team after they received the note that was left at their home address by one of the investigators. The size of the proportion of these subjects is impossible to determine retrospectively and it is a possible source of selection bias. It must also be taken into account that people with headache are more willing to participate in a headache study. Given that 43.2% of the contactable sample refused to participate in the study, there is a possible interest bias that may lead to an overestimation of the true prevalence of headache disorders. However, we have tried to minimize these possible biases by weighting the sample by age, gender, habitat, education and marital status.
Secondly, a possible source of bias is created by using a questionnaire that is validated in a clinical setting. This creates the risk of underestimating the true prevalences in the population, since headache clinic patients are pre-educated and know how to answer the questions more precisely. In order to compensate for these differences between the validation sample and the general population, the interviews were carried out by medical students so the participant could ask for clarification if they were in doubt or confused about the questions asked. Thus, although during the aforementioned validation study the questionnaire was self-administered and in the current epidemiological study the questionnaire was administered face-to-face or by telephone by medical students, we believe this will not lead to significant over- or underestimation, because face-to-face or telephone interview gives the subject the possibility of asking specifying questions in case of doubt, resulting in more reliable responses than with a self-administered questionnaire.
One source of possible underestimation of the prevalences can be hidden within the screening question. We aimed to avoid contamination of the data by secondary headaches using the specific wording; however, it has been shown that such a non-neutral screening question may produce false negatives (12). As mentioned before in this section, this can also be one of the reasons why the prevalence of episodic tension-type headache in our study was lower than in other countries nearby. The subjects were asked if they had had “recurrent” headaches, and it may be suspected that people with infrequent and milder headaches do not consider these attacks as recurrent and thus do not report them. This may cause the underestimation of all types of headaches (26) but most of all episodic tension-type headache, which is by definition a milder and less bothersome headache.
The main strength of our study is that we were able to estimate the prevalence of almost all primary headaches in our country, including migraine and tension-type headache as the socioeconomically most bothersome entities as well as the rare trigeminal autonomic cephalalgias and other primary headaches. The results of our study are representative of the Estonian population aged 18–64 years. Another strength of the study is the low proportion of unidentified headaches (2%) in the sample. We used the most up-to-date classification, while the vast majority of prevalence studies published in recent years were based on the previous (ICHD-2) classification, published in 2004.
In conclusion, while there are substantial uncertainties in our estimates of the one-year prevalences of primary headache disorders in Estonia, they are nonetheless comparable to previous findings in other European countries, except for episodic tension-type headache. However, we believe that this is an underestimation due to the limitations of our study addressed earlier. This in turn means that, as in other countries, primary headaches in Estonia are an important cause of morbidity, loss of quality of life and personal suffering as well as a prominent socioeconomic burden. Thus, more attention and appropriate intervention is needed on medical, educational and political levels.
Footnotes
Public health relevance
The one-year prevalences of primary headache disorders in Estonia are comparable to the previous findings in other European countries.
More attention to headache disorders and appropirate intervention is needed on medical, educational and political levels in Estonia.
Acknowledgements
The authors are very grateful to the participants of the present study. We thank Kaja-Triin Laisaar, MD, PhD (Institute of Family Medicine and Public Health, University of Tartu) for valuable methodological advisory expertise in epidemiologic questions.
Declaration of conflicting of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Estonian Headache Society and MyFitness (public limited company). Aire Raidvee was supported by institutional research funding (IUT02-13) from the Estonian Ministry of Education and Science to Jüri Allik. The funders had no role in study design, data collection and analysis or preparation of the manuscript.