The prevalence of primary headache disorders in Nepal: a nationwide population-based study

The Nepal Health Research Council, the Institutional Review Committee of Kathmandu University School of Medical Sciences, Dhulikhel Hospital (IRC-KUSMS) and the Regional Committee for Health and Research Ethics in central Norway all approved the study protocol.

All participants were informed about the nature and purpose of the study. Written consent was obtained by signature or fingerprint in accordance with requirements of IRC-KUSMS.

This was a cross-sectional, population-based survey using structured interviews administered by trained health workers making unannounced door-to-door visits to households in May, 2013. We randomly selected one eligible adult (aged 18–65 years, Nepali-speaking and living in Nepal) from each household. To ensure adequate representation from the country as a whole, we used multistage stratified cluster sampling, including all three physiographic divisions and, within each, all five development regions (Far-Western, Midwestern, Western, Central and Eastern). The details of the sampling and data collection procedure, including the steps taken to ensure a very high participation rate, have been published elsewhere [10]. The sample size (N = 2,100) was estimated assuming a headache-type prevalence of ≥10 % and absolute margin of error of 1.3 % with 95 % confidence interval (CI).

We used the Headache-Attributed Restriction, Disability, Social Handicap and Impaired Participation (HARDSHIP) questionnaire developed by LTB for similar studies [13]. The English version was translated into Nepalese according to LTB’s translation protocol for lay documents [14] and adapted to fit the characteristics of the Nepalese culture [15].

The full questionnaire has been published previously [10]. It consisted of five parts. For all participants there were (i) personal and demographic enquiry and (ii) a headache screening question (“Have you had a headache in the last year?”). Those who answered “no” were classified as headache-free. Those who answered “yes” were asked whether their headaches were of one or more types and, if more than one, to focus only on the most bothersome type. Only those who answered positively to the screening question were also asked (iii) diagnostic and (iv) burden and health-care questions relating to their headache. To the standard HARDSHIP questions we added others relating to use of herbal therapies. Finally, there were (v) questions on certain comorbidities asked of all participants [10].

We used culturally-validated Nepali-translations of the Hospital Anxiety and Depression Scale (HADS) [16] and the Eysenck Personality Questionnaire Neuroticism Short Form Revised version (EPQRS-N) [17] to assess psychiatric comorbidity. We measured height, weight and waist-circumference, and calculated body-mass index (BMI). We measured blood pressure (BP) using a digital device (3BM1-3® by Microlife).

We recorded the altitude of each household using a portable altimeter (SAL 7030® by Sunoh).

Diagnoses were not made during the interviews but later by an algorithm [13]. Participants reporting headache on ≥15 days/month were first separated as a distinct group because they cannot be fully diagnosed by questionnaire. Those who were also overusing acute medication were considered to have probable MOH (pMOH); the remainder were diagnosed as “other headache on ≥15 days/month”. Medication overuse was diagnosed in those who: a) reported the use on ≥15 days/month of either a single class of analgesics, or one type of herbal medicine, as acute headache treatment; or b) reported using on ≥10 days/month (i) a combination of analgesics of different classes, or (ii) more than one type of herbal medicine, or (iii) a combination of analgesics and herbal medicines. Triptans and ergots were not used.

To all others, reporting headache on ≤14 days/month, the algorithm applied modified criteria of the International Classification of Headache Disorders (ICHD-3 beta) [18] in the following order: definite migraine, definite TTH, probable migraine and probable TTH. We found that two additional adaptations were necessary. Firstly, photophobia was reported in association with more than three-quarters (75.8 %) of all headaches and therefore offered little discriminative value diagnostically. Accordingly, we ignored it when diagnosing headache types. Secondly, according to ICHD-3 beta [18], attacks lasting <4 h when untreated in adults may be compatible with a diagnosis of probable migraine when other criteria are met. Many of our participants could report attack durations only after taking acute medication, and some were very short. We decided to disallow a diagnosis of probable migraine (in favour of probable TTH) whenever headache duration was <1 h. We took the view that so short a duration, even with acute treatment, was very unlikely to be migraine, given that the adult participant was asked to describe a “usual” attack [19].

Cases of definite and probable migraine were combined as, likewise, were cases of definite and probable TTH in the estimations of prevalence and in association analyses. Remaining cases were considered unclassifiable.

We estimated crude 1-year prevalence with 95 % CI for all headache, migraine and TTH, and point prevalence for all headache on ≥15 days/month and pMOH. We adjusted prevalences for gender and age according to the general population distribution (within the range 18–65 years) in Nepal [20].

We categorized age in five groups (18–25, 26–35, 36–45, 46–55 and 56–65 years). We classed habitation as rural or urban, and categorized altitude of the household as <500 m, 500–999 m, 1,000–1,499 m, 1,500–1,999 m, 2,000-2,499 m or ≥2,500 m. We took household consumption per year in USD (at the time of the survey, USD 1 = NPR 100) as an indicator of the economic wellbeing of the participant and categorized it in three groups: poorest (<USD 950/year); poor (USD 950–1,200/year); intermediate and above (>USD 1,200/year).

We used bivariate and multivariate logistic regression analyses (with odds ratios [ORs] and adjusted ORs [AORs] respectively, each with 95 % CIs) to investigate associations of demographic, lifestyle, environmental and other health factors with each of migraine, TTH and pMOH. Gender, age, household consumption, habitation, altitude, systolic and diastolic BP and EPQRS-N and HADS scores were entered as covariates in the multivariate logistic regression. BP readings and EPQRS-N and HADS scores were treated as continuous variables.

Statistical analyses were carried out using the Statistical Package for Social Science software (IBM SPSS Statistics 21, Chicago, USA).

None of the headache disorders was associated with household consumption (Table 4). TTH was weakly associated with urban dwelling (OR = 1.4 [95 % CI: 1.2–1.7]; p = 0.001; AOR = 1.3 [95 % CI: 1.1-1.6]; p = 0.003) (Tables 3 and 4). Migraine was strongly associated with living at an altitude of ≥1,000 m in both bivariate (OR = 1.7 [95 % CI: 1.4-2.0]; p < 0.001) (Table 3) and multivariate analyses (AOR = 1.6 [95 % CI: 1.3–2.0]; p < 0.001) (Table 4). In view of this finding, we estimated the prevalence of migraine by altitude category, standardising for age and gender against census data for the Nepali population [20] (Table 5). This analysis revealed that the age- and gender standardised prevalence of migraine was 27.9 % in the low-lying Terai (<500 m; there were no participants in our sample living between 500 and 999 m). Prevalence increased in an almost linear relationship with altitude up to 2,000 m. Thereafter it levelled, and indeed declined. This relationship was evident in both genders.

A negative association between TTH and altitude was indicated by both bivariate (OR = 0.7 [95 % CI: 0.6–0.8]; p < 0.001) and multivariate analyses (AOR = 0.7 [95 % CI: 0.6–0.8]; p < 0.001) (Tables 3 and 4).

Our most obvious finding was that the prevalence of migraine in Nepal is very much higher than the mean global estimate of 14.7 % [22]. Our first comment on this is to note that more recent studies, in all regions except the Far East, have generally yielded higher values than 14.7 %. The mean global estimate is based on a large number of heterogeneous studies, performed with varying methods during a period of >30 years. Many reports are silent on the crucial issue of how they applied diagnostic criteria with respect to definite and probable migraine; while some explicitly excluded the latter, it is highly probable that more did so without making this evident, because only recently has there been clear guidance and explanation of why this is misleading [21]. The consequences, in our view, are that many of these studies significantly underestimate migraine prevalence and, therefore, so does the global mean. Nonetheless, our finding of 34.1 % in Nepal is considerably higher even than the 25.2 % reported from neighbouring India [23] (although this was from the single State of Karnataka in the south). We used the same methodology and diagnostic questionnaire as Karnataka [24]; indeed, LTB has supported studies using similar methods and the same questionnaire in many other countries, cultures and languages [13]: in Russia [25], China [26], Zambia [27], Ethiopia (unpublished), Pakistan (unpublished), Saudi Arabia (unpublished), Morocco (unpublished). None has discovered such a high prevalence of migraine; in fact, India (Karnataka) was the highest [23].

Nepal is a poor country, but we discovered no association with economic wellbeing that might offer an explanation (neither was there one in India [23]). The clearly relevant factor is altitude: over half our sample were resident at or above 1,000 m, and nearly one quarter above 2,000 m. Altitude, we discovered, is an environmental factor so strongly associated with migraine that living above 1,000 m led to 60 % increased odds of having the disorder. In practice this meant that, among the adult population living above this altitude, an additional 11.4 % had migraine (Table 2). Since more than half the population were in fact living above 1,000 m, this explained a very large part of the overall excess prevalence. For the remainder, we should look to the distinctive age and gender distributions of Nepal: more precisely, to the high proportions of young adults (median age is 22 years [9]), and females among these adults, the two strongest determinants of migraine prevalence [1]. Furthermore, large numbers of healthy people have consistently been going abroad for work, around half of them in the age group 15–35 years, and mostly males [9], leaving behind a skewed adult distribution perhaps more at risk of migraine. It is entirely possible to see this as the explanation of the difference between 27.9 %, the standardised prevalence in the Terai of Nepal, and 25.2 %, the standardised prevalence in Karnataka [23] (which is at a mean altitude of about 900 m [28]).

We should acknowledge two diagnostic issues. First, it was a limitation of the study that we were not able to perform a prior validation of the headache diagnostic questionnaire in its Nepali translation. Because there are no headache specialists in the country, we had no means of applying a “gold standard” for this purpose [21]. We had to rely on the fact that the questionnaire had been validated in many other languages and countries [13], including India with not too dissimilar cultural settings [24].

Second, having said this, we encountered a problem that was not met in India: such a large proportion (75.8 %) of respondents with any type of headache reported photophobia that this symptom had virtually no discriminative value as a diagnostic criterion, and we could not use it within the framework of ICHD [18]. Photophobia is a technical concept, not easy to convey to lay participants (even by trained interviewers) [21]. Because of this we had taken great care in translation to convey not merely an aversion to bright light but the idea of a wish to withdraw into darkness from ordinary light. Our eventual solution was to disregard photophobia altogether, and in our view this was necessary: the prevalence estimate for migraine would otherwise have been much higher. What this suggests is that ICHD criteria – or at least this particular one – may not be universally applicable, and not for linguistic reasons.

The estimated prevalence of TTH in Nepal at 41.1 % is in line with the estimated global average [2], although considerably lower than some national estimates [29, 30]. Two factors are relevant here. One is whether or not survey participants report infrequent TTH, which can have a marked effect on the prevalence estimate and is likely to be determined in part culturally and in part by the insistence (and purpose) of the interviewer [23, 24]. In our case, not only did we not consider infrequent TTH to be of public-health significance, but also we focused only on the most bothersome headache in participants identifying more than one type. Those with both migraine and TTH would tend to regard the former as the more bothersome [31], leading to a partial neglect of TTH. Our prevalence estimate for TTH was therefore somewhat conservative.

We suspect this was a factor in the observed negative association between TTH prevalence and altitude: as migraine prevalence increased with altitude, reporting of TTH became less likely. In other words, it was artefactual. We do not know this, but on the other hand can offer no explanation for a true negative association.

TTH was significantly more prevalent in urban areas, an association not observed for migraine or pMOH. Possible explanations for the higher prevalence of TTH may be the noisy and stressful environments of the cities. Also, physical inactivity is a risk factor specifically for non-migraine headache [32], and people in rural areas in Nepal are more physically active working in the fields. These are speculative proposals.

According to our estimate, the prevalence of headache on ≥15 days/month in Nepal is more than twice the global average [2] and much higher than in neighbouring China [26] and India [23]. Several factors such as low socio-economic status, poor access to health services with a paucity of health-care providers and lack of standard protocols for diagnosis and treatment of headache disorders, all against a background of high levels of episodic headache, offer what may be sufficient explanation.

We do not know what the cases with headache on ≥15 days/month were diagnostically, other than the proportion diagnosed as pMOH. We estimated the prevalence of pMOH in Nepal at 2.1 %, which is towards the upper end of the range for most countries studied (0.5–2.6 %) [33], and much higher than the 0.6 % and 1.2 % in neighbouring China [26] and India [23]. Surprisingly, the proportion of pMOH among those with headache on ≥15 days/month was only 29 % (46/161); in other countries where studies have used similar methodology, this proportion is closer to two thirds [23, 25, 26]. This may be due to poverty: most pMOH was associated with overuse of simple analgesics sold over-the-counter (OTC), but many people in Nepal cannot afford these. They have recourse to alternative treatments, which may include plant-based remedies that are used regularly across the country, especially in the rural areas. Many of these have phytochemical or pharmacological properties [34], and may be able to cause transformation of episodic headache to pMOH. We endeavoured to include these in our enquiry, and identified six cases, but herb-based remedies may not be perceived by lay people as therapy for headache and it is likely that they were underreported.

Headache disorders are very common in Nepal. Only the prevalence of TTH is in line with the global average. That of headache on ≥15 days/month is double, while the prevalence of pMOH is towards the upper end of the range for most countries studied. The prevalence of migraine, however, appears uniquely high, explained, we believe, by some of the distinct characteristics of this country. One is the combination of mountainous and hilly terrain over much of the country, important in the light of the previously unreported but strong association between migraine prevalence and altitude. Another is the demographic make-up.

The new evidence from this study will inform national health policy, and provide a basis for health-care needs assessment. Meanwhile, research of a different type is needed to find explanations for the association between migraine and altitude, since these may be relevant to health-care interventions. Continuous long-duration exposure to high altitude compromises oxygen uptake and results in haemodynamic changes, with elevated haemoglobin levels, increased blood viscosity and reduced oxygen delivery to brain tissues [35, 36]. Migraine has been linked with the consequences of such changes [37–40], but their relevance at lower altitudes, between 500 and 2,000 m, needs to be investigated.