Sage Journals: Discover world-class research

Abstract

Background

Headache has not been established as a risk factor for dementia. The aim of this study was to determine whether any headache was associated with subsequent development of vascular dementia (VaD), Alzheimer’s disease (AD) or other types of dementia.

Methods

This prospective population-based cohort study used baseline data from the Nord-Trøndelag Health Study (HUNT 2) performed during 1995–1997 and, from the same Norwegian county, a register of cases diagnosed with dementia during 1997–2010. Participants aged ≥20 years who responded to headache questions in HUNT 2 were categorized (headache free; with any headache; with migraine; with nonmigrainous headache). Hazard ratios (HRs) for later inclusion in the dementia register were estimated using Cox regression analysis.

Results

Of 51,383 participants providing headache data in HUNT 2, 378 appeared in the dementia register during the follow-up period. Compared to those who were headache free, participants with any headache had increased risk of VaD (n = 63) (multivariate-adjusted HR = 2.3, 95% CI 1.4–3.8, p = 0.002) and of mixed dementia (VaD and AD (n = 52)) (adjusted HR = 2.0, 95% CI 1.1–3.5, p = 0.018). There was no association between any headache and later development of AD (n = 180).

Conclusion

In this prospective population-based cohort study, any headache was a risk factor for development of VaD.

Keywords

Alzheimer’s disease vascular dementia headache migraine

Introduction

Dementia, the most common neurological disorder among the elderly, has a huge impact on quality of life of those affected and a substantial economic impact on society (1). Risk factors for dementia are of great interest to researchers, clinicians and the public, especially with a view to prevention. In the last decades, several modifiable risk factors have been identified (e.g. obesity, diabetes, hypertension and lipid disorders) (2 –4).

Globally, approximately 45% of adults in the general population of the world suffer from an active headache disorder (5). Headache has not been established as a risk factor for dementia. Most research on the relationship between headache and cognition has focused on migraine. A history of migraine is statistically associated with cardiovascular disease (6 –9), and with brain white matter lesions appearing as hyperintensities on magnetic resonance imaging (MRI) (10). Despite this, no previous studies have found migraine (11 –15) or nonmigrainous headache (12) to be associated with cognitive decline. The relationship between other primary headache and cognitive decline and dementia is largely unknown, but of interest since nonmigrainous headache, at least in some population-based studies, also has been found to be associated with unfavorable vascular risk profiles (16,17).

However, no population-based follow-up study has evaluated headache as a potential risk factor for dementia. Our purpose in this large prospective study of the general population was to do this.

Methods

This prospective population-based cohort study used baseline data from the Nord-Trøndelag Health Study (HUNT 2) performed between August 1995 and June 1997 and, from the same Norwegian county, a register of cases diagnosed with dementia during 1997–2010.

HUNT and diagnosis of headache

In HUNT 2 all inhabitants ≥20 years old of Nord-Trøndelag were invited to participate. Each participant had to complete two extensive questionnaires, and participants were also invited to clinical consultations that included measurements of height and weight. The median number of days between the return of the first questionnaire (Q1) and the day the second questionnaire (Q2) was filled in at home was two days (range 0–30) (18). Among the topics in the Q1 were education, physical activity and smoking. Anxiety and depression were assessed by the Hospital Anxiety and Depression Scale (HADS) (19). Educational level was categorized according to duration: ≤9 years, 10–12 years or ≥13 years. Cigarette smoking was categorized as “current daily smoking,” “previous daily smoking” or “never daily smoking.” Reponses to questions on physical activity were categorized in declining sequence according to duration and intensity of exercise per week: ≥3 hours’ hard physical activity, one to two hours’ hard physical activity, ≥3 hours’ light physical activity, one to two hours’ light physical activity, or physical inactivity (0 hours). Body mass index (BMI) was calculated and, for the present study, categorized as <25 kg/m², 25–29.9 kg/m² or ≥30 kg/m².

Headache questions in Q2 were designed first to determine whether a participant had a current headache disorder and, if so, the frequency of headache, and second to diagnose migraine by applying modified migraine criteria of the International Classification of Headache Disorders, first edition (ICHD-I) (20,21). Subjects were classified as having any headache when they answered “yes” to the screening question, “Have you suffered from headache during the last 12 months?” and otherwise as headache free. Headache that did not fulfill the criteria for migraine was classified as nonmigrainous. For headache frequency, those with any headache were asked “On approximately how many days per month do you suffer from headache?” and categorized as one to 14 days or ≥15 days. The validity of these questionnaire-based diagnoses has been reported previously (18): for any headache, sensitivity was 85% and specificity 83% (kappa value 0.57); for migraine, sensitivity was 69% and specificity 89% (kappa value 0.59); for migraine with aura, sensitivity was 50% and specificity 100% (kappa value 0.64), for nonmigrainous headache, sensitivity was 61%, specificity 81% (kappa 0.43); and for headache on ≥15 days/month, sensitivity was 38% and specificity 97% (kappa 0.44). In our validation study, 80% of individuals with nonmigrainous headache suffered from tension-type headache (18).

Diagnosis of dementia

A dementia registry had been established in Nord-Trøndelag County by a two-step procedure. First, a computerized search of the administrative patient database of the two hospitals in Nord-Trøndelag (Namsos Hospital and Levanger Hospital) identified patients with any suspected type of dementia in the period between 1995 and 2010. Second, the hospital records of all such patients were reviewed by a diagnostic team of researchers and clinical experts in dementia, consisting of an internist (geriatrist) and two psychogeriatrists. Cases without complete electronic medical records (mainly introduced in year 2000) were excluded. The diagnostic team employed used a conservative diagnostic strategy to minimize bias from misdiagnoses: As a minimum, included cases had been evaluated by clinical experts in dementia and investigated by computerized tomography or MRI of the brain. All cases were diagnosed in two steps, applying first the clinical ICD-10 criteria for dementia (22), then the research ICD-10 criteria for AD and vascular dementia (VaD) (23). Patients with AD were required also to fulfill the National Institute of Neurological and Communicative Disorders Association and Stroke-Alzheimer’s Disease and Related Disorder Association (NINCDS-ADRDA) criteria for this disorder (24), and patients with VaD the National Institute of Neurological Disorders and Stroke-Association Internationale pour la Recherche et l’Enseignement en Neurosciences (NINDS-AIREN) criteria for VaD (25). Dementia with Lewy bodies was diagnosed in cases with two of the following three symptoms: fluctuating cognition, visual hallucinations and motor features of Parkinson’s disease. Cases of frontotemporal dementia had documented changes in behavior and personality. Cases of other types of dementia met the general criteria for dementia but not the specific criteria for AD, VaD, dementia with Lewy bodies or frontotemporal dementia.

The dementia registry was linked to data from the HUNT study using the Norwegian 11-digit personal identity number, which is unique to each resident in Norway.

Study populations

In HUNT 2, 64,787 (70%) of 92,936 invited individuals participated, of whom 51,383 (56%) answered the headache questions in Q2. Details of nonresponders have been described previously (21). Individuals who responded to the headache questions tended to be younger, were more likely to be women and had higher socioeconomic status than nonresponders (21).

The population at risk was defined as those susceptible to but free from the dementia in question at the time of HUNT 2. In this study, the population at risk (N = 51,020) included all 51,383 answering the headache questions in Q2 less 363 patients with mild cognitive impairment (MCI, n = 35) or with unvalidated diagnoses of dementia (n = 328) (Figure 1) (26).The population at risk was divided into participants with any headache (subdivided into those with migraine and those with nonmigrainous headache) and those who were headache free. From the dementia registry we identified those who later developed dementia, and categorized them according to dementia type.

Figure 1.

Flow diagram of the study population.

Statistical analysis

Differences between individuals with and without dementia were tested with independent-sample Student’s t-tests for continuous variables and with chi-squared test for proportions. In the population at risk, we used Cox proportional hazards model to estimate hazard ratios (HRs) of different types of dementia with 95% confidence intervals (CI). For each category of headache (any headache, nonmigrainous headache and migraine), the risk of dementia was estimated by comparing with headache free as the reference category. We initially adjusted for age (continuous variable) and gender, and subsequently for other potential confounding factors identified previously (e.g. education (three categories), smoking (three categories), systolic blood pressure (SBP), current use of antihypertensive medication (yes vs no), physical activity (five categories), BMI (three categories), total HADS score (three categories), self-reported cardiovascular diseases (yes vs no) and abstention from alcohol (yes vs no) (27). Further, covariates were included that could potentially influence HRs of dementia, including total cholesterol (continuous variable), triglycerides (continuous variable), nonfasting glucose (continuous variable), daily use of any medication (yes/no), and self-reported stroke (yes/no). Covariates that did not substantially change the estimated HRs separately or together (i.e. less than 1%) were excluded from the final model. An interaction was assessed by including the cross-product of any headache and another variable of interest in a multivariate Cox model (i.e. age, smoking, total HADS score, physical activity); thereafter, the Wald test was performed. Subjects with incomplete data for one or several variables (number stated in Table 1) were nevertheless included (as a separate missing category) in all analyses to reduce the impact of response bias. We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement guidelines in reporting our study cohort. Statistical analyses were performed with the Predictive Analytics SoftWare (PASW) Statistics version 17.0 by SPSS Inc, an IBM Company (Chicago, IL, USA).

Table 1.

Baseline characteristics of cohort members (n = 51,859).

Variables	Without headache (n = 29,988)	With any headache (n = 21,871)
Age (mean (SD))	52.3 (17.7)	46.2 (15.2)
Gender, female (%)	46.5	64.3.
Education >12 years (%) (missing^a= 1648)	19.7	20.8
Married (%)	61.4	61.1
Body mass index (mean (SD))	26.3 (3.9)	26.3 (4.2)
Diastolic blood pressure (mean (SD))	80.8 (12.3)	79.3 (11.8)
Systolic blood pressure (mean (SD))	133.9 (20.2)	133.9 (22.0)
Total cholesterol (mean (SD))	5.94 (1.27)	5.79 (1.22)
HDL cholesterol (mean (SD))	1.37 (0.39)	1.39 (0.39)
Triglycerides (mean (SD))	1.78 (1.13)	1.70 (1.11)
Glucose (mean (SD)) (missing^a= 404)	5.52 (1.59)	5.33 (1.30)
HADS A score (mean (SD)) (missing^a= 6370)	3.58 (2.93)	4.92 (3.55)
HADS D score (mean (SD)) (missing^a= 3546)	3.18 (2.87)	3.71 (3.18)
Total HADS^b score (mean (SD)) (missing = 1520)	6.83 (5.11)	8.76 (6.05)
Current use of antihypertensive medication (%)	12.6	8.8
Self-reported diabetes mellitus (%)	3.3	2.4
Self-reported heart infarction (%)	4.1	1.8
Self-reported angina pectoris (%)	5.6	3.6
Self-reported stroke (%)	2.0	1.5
Current smoking (%) (missing^a= 286)	25.3	30.3
Abstainers from alcohol (%) (missing^a= 634)	12.5	10.9
Daily use of medication (%) (missing^a= 1809)	37.9	44.6
Physical inactivity (%) (missing^a= 4139)	21.5	23.3

Numbers of individuals with missing values are given when different from zero. ^bHospital Anxiety and Depression Scale (HADS) anxiety (A) depression (D).

Ethics

The study was approved by the Regional Committee for Ethics in Medical Research, and by the research group Health and Memory in Nord-Trøndelag.

Results

Baseline characteristics of the cohort members are given in Table 1. Compared to those with any headache, individuals without headache were older, more likely to be men and less likely to be current smokers and alcohol users, had somewhat shorter education, higher SBP and diastolic BP, higher total cholesterol and triglyceride levels, higher nonfasting glucose, more self-reported cardiovascular diseases, and, lower HADS scores, and were more likely to be using antihypertensive medication.

From the dementia registry we identified 912 patients with dementia (all criteria present), of whom 476 (52%) had participated in the HUNT 2 study; of these, 378 patients (224 women) had answered the headache questions: 180 with AD, 63 with VaD, 52 with mixed dementia (AD + VaD), 13 with Lewy body dementia, 12 with frontotemporal dementia and 58 with other types of dementia (alcohol related (three), Parkinson’s disease (11), Alzheimer’s combined with Down’s syndrome (two), Huntington’s disease (two), other types or unspecified (40)). Mean age at diagnosis was 79.9 years (standard deviation (SD) 7.7; range 50–95) with no difference between men and women (79.3 vs 80.2 years, p = 0.28). On average, dementia was diagnosed 8.6 years (SD 2.8; range two to 15) after participation in HUNT 2. A total of 211 patients (56%) of the 378 had died before March 2010, 44 before 2005. Any headache at baseline in HUNT 2 was associated with slightly higher risk of dementia of any cause (age- and gender-adjusted HR = 1.3; 95% CI 1.1–1.7, p = 0.009), although this risk failed to reach statistical significance when the genders were considered separately (women: p = 0.054; men: p = 0.08) (Table 2).

Table 2.

Multivariable-adjusted^a HRs (95% CIs) of different types of dementia by end of follow-up according to headache status at baseline in HUNT 2.

	Number	Dementia, all types		Alzheimer’s disease (AD)		Vascular dementia (VaD)		Mixed dementia^b		VaD plus mixed dementia^c		Dementia with Lewy bodies		Frontotemp. dementia		Other types of dementia
	Number	No	HR	No	HR	No	HR	No	HR	No	HR	No	HR	No	HR	No	HR
Both genders	51,859	378	180	63	52	115	13	12	58
No headache	29,988	249	1.0 (ref.)	127 xx	1.0 (ref.)	35	1.0 (ref.)	31	1.0 (ref.)	66	1.0 (ref.)	9	1.0 (ref.)	6	1.0 (ref.)	41	1.0 (ref.)
Any headache	21,871	129	1.3 (1.1–1.7)	53	1.0 (0.7–1.4)	28	2.3 (1.4–3.8)	21	2.0 (1.1–3.5)	49	2.1 (1.5–3.1)	4	1.3 (0.4–4.5)	6	2.9 (0.9–9.1)	17	1.0 (0.5–1.8)
Men	23,857	154	57	32	21	53	8	8	28
No headache	16,044	113	1.0 (ref.)	46	1.0 (ref.)	21	1.0 (ref.)	12	1.0 (ref.)	33	1.0 (ref.)	8	1.0 (ref.)	4	1.0 (ref.)	22	1.0 (ref.)
Any headache	7813	41	1.4 (1.0–2.0)	11	0.9 (0.5–1.8)	11	2.1 (1.9–4.5)	9	3.3 (1.4–7.8)	20	2.5 (1.4–4.4)	0	–	4	3.2 (0.8–13)	6	1.0 (0.4–2.4)
Women	28,002	224	123	31	31	62	5	4	30
No headache	13,944	136	1.0 (ref.)	81	1.0 (ref.)	14	1.0 (ref.)	19	1.0 (ref.)	33	1.0 (ref.)	1	1.0 (ref.)	2	1.0 (ref.)	19	1.0 (ref.)
Any headache	14,058	X 88	1.3 (1.0–1.7)	42	1.1 (0.7–1.6)	X 17	2.4 (1.2–4.9)	12	1.4 (0.7–3.0)	29	1.8 (1.1–3.1)	4	8.6 (1.0–78)	2	2.2 (0.3–16)	11 ×xxx	1.0 (0.5–2.1)

HR: hazard ratio; CI: confidence interval; HADS: Hospital Anxiety and Depression Scale; HUNT 2: Nord-Trøndelag Health Study.

Adjusted for age, gender, education, total HADS score, and smoking. The following covariates were excluded from the final model (change the estimated HRs with less than 1%): body mass index, triglycerides, non-fasting glucose, total cholesterol, systolic blood pressure, use of antihypertensive medication, self-reported stroke, and physical inactivity. ^bMixed dementia: merged groups of both AD and VaD (n = 52). ^cMerged groups of VaD (n = 63) + mixed dementia (n = 52).

In multivariate analyses, age, gender, education, smoking and total HADS score were identified as important confounders; these were included in the final analyses. Any headache in HUNT 2 was associated with an increased risk of VaD (HR = 2.3; 95% CI 1.4–3.8, p = 0.002), this risk being similar and remaining significant both in men (p = 0.043) and women (p = 0.018) (Table 2). There was a significant interaction between age and any headache regarding VaD (p < 0.0001): In subsequent analyses stratifying by age, any headache increased the risk of VaD more among individuals ≥75 years of age at baseline (HR = 3.0; 95% CI 1.3–6.7, p = 0.007) than among those <75 years (HR = 1.8; 95% CI 1.0–3.6, p = 0.059). Any headache in HUNT 2 was also associated with increased risk of being diagnosed with mixed dementia (HR = 2.0; 95% CI 1.1–3.5, p = 0.018), evident in men (HR = 3.3; 95% CI 1.4–7.8, p = 0.001) but not in women (Table 2). When both groups with VaD (VaD only (n = 62) and mixed dementia (n = 52)) were pooled, the increased risk was more marked for men (HR = 2.5; 95% CI 1.4–4.4, p = 0.001) than for women ((HR = 1.8; 95% CI 1.1–3.1, p = 0.018) (Table 2). No relationship was found between any headache in HUNT 2 and a later diagnosis of AD (Table 2).

The association with VaD was almost identical for migraine and nonmigrainous headache and between migraine with and without aura (Table 3). However, frequency of any headache influenced the risk: HR was greater for any headache on ≥15 days/month (HR = 4.2; 95% CI 1.7–10.8, p = 0.002) than for any headache on <15 days/month (HR = 2.0; 95% CI 1.2–3.5, p = 0.009) (Table 3). This was more marked for migraine on ≥15 days/month (HR = 9.1 (2.2–40.1), p = 0.003) than for nonmigrainous headache on ≥15 days/month (HR = 3.2 (1.0–10.3), p = 0.057).

Table 3.

HRs (95% CI) for vascular dementia (VaD) (n = 63) by end of follow-up according to headache status at baseline in HUNT 2.

	Overall			Men			Women
	Total number	VaD	Age adjusted^a	Multivariable adjusted^c	VaD	Age adjusted^b	Multivariable adjusted^c	VaD	Age adjusted^b	Multivariable adjusted^c
No headache (reference)	29,988	35	1.0	1.0	21	1.0	1.0	14	1.0	1.0
Any headache	21,871	28	2.0	2.3 (1.4–3.8)	11	2.0	2.1 (1.0–4.5)	17	2.3	2.4 (1.2–4.9)
Headache <15 days/month	20,589	23	1.8	2.0 (1.2–3.5)	8	1.6	1.8 (0.8–4.0)	15	2.2	2.3 (1.1–4.7)
Headache ≥15 days/month	1282	5	4.0	4.2 (1.7–10.8)	3	4.6	5.2 (1.5–17.4)	2	3.4	3.5 (0.8–15.2)
Migraine (all)	6740	7	2.3	2.9 (1.3–6.6)	2	2.0	2.4 (0.6–10.4)	5	2.7	3.1 (1.0–8.1)
Migraine with aura	1019	1	2.7	3.0 (0.4–22.5)	1	10.4	13.3 (1.6–108)	0	–	–
Migraine without aura	5721	6	2.8	2.7 (1.1–6.7)	1	1.4	1.4 (0.2–10.6)	5	3.4	3.3 (1.2–9.5)
<15 days/month	6335	5	1.8	2.1 (0.8–5.4)	1	1.1	1.4 (0.2–10.4)	4	2.3	2.5 (0.8–7.7)
≥15 days/month	405	2	8.7	9.1 (2.2–40.1)	1	12.5	18.3 (2.2–151)	1	7.3	5.6 (0.7–47.7)
Nonmigrainous headache	15,131	21	1.9	2.1 (1.2–3.7)	9	1.9	2.2 (1.0–4.8)	12	2.1	2.1 (1.0–4.6)
<15 days/month	14,254	18	1.8	2.0 (1.1–3.5)	7	1.7	1.9 (0.8–4.4)	11	2.1	2.2 (1.0–4.8)
≥15 days/month	877	3	2.9	3.2 (1.0–10.3)	2	3.5	4.0 (0.9–17.5)	1	2.2	2.3 (0.3–16.6)

HR: hazard ratio; CI: confidence interval; HADS: Hospital Anxiety and Depression Scale; HUNT 2: Nord-Trøndelag Health Study.

Adjusted for age and gender. ^bAdjusted for age. ^cAdjusted for age, gender, education, total HADS score, and smoking. The following covariates were excluded from the final model (changed the estimated HRs less than 1%): body mass index, triglycerides, nonfasting glucose, total cholesterol, systolic blood pressure, use of antihypertensive medication, self-reported stroke, and physical inactivity.

Discussion

In this large population-based cohort study, any headache was associated with an increased risk of later diagnosis of VaD, most evident among individuals aged ≥75 years at baseline. In contrast, no association was found between any headache and the later development of AD alone.

Comparison with other studies

To the best of our knowledge, no large population-based follow-up study of adults, with similar objectives and study design, has hitherto been published. According to a recent preliminary report, the influence of migraine on the risk of dementia has been evaluated during follow-up of 716 participants in a health study in Canada (28). In marked contrast to our results, people with a history of migraine were four times more likely to develop AD, whereas no clear relationship was reported between migraine and VaD. However, too few details are available in the Canadian study to compare with our study population, and the quality of these data remains to be seen in the full publication.

Until now, no relationship between headaches and VaD had been established, and no association between migraine and cognitive decline had been found in population-based cross-sectional (11) or prospective studies (13 –15). A lifetime diagnosis of migraine was not associated with cognitive deficits among middle-aged Danish twins (11). In a study of 780 participants with a mean age of 69 years, a history of severe headache was associated with an increased volume of brain white matter hyperintensities, but these abnormalities were not related to cognitive impairment (12). This lack of association between migraine and cognitive decline is in contrast to our results, but the reason is unclear. It should be emphasized that none of these earlier studies focused on dementia and, because patients with cognitive impairment are less likely to participate in such studies, the possibility of selection bias could not be ruled out. Furthermore, two of the prospective studies had relatively short follow-up durations of <7 years, and the migraine patients included in all three were relatively young from the point of view of developing cognitive decline (mean ages of 47, 66 and 69 years) (13 –15). All these factors may have influenced the divergent results since any headache was a greater risk factor among those ≥75 years in our study, when incipient dementia is more prevalent.

Interpretation

Headache status and the outcome (dementia) came from two separate registries. Thus, it is unlikely that information in the patients’ medical records about headache disorders could have influenced diagnoses of dementia, since no clear relationship between dementia and migraine or other headache existed when the dementia registry was established, and a history of headache is not a part of the diagnostic criteria for any types of dementia. However, those making the dementia diagnoses would still have knowledge of other risk factors. Thus, we cannot rule out that the relationship between any headache and VaD may be an artifactual, because headache is associated with vascular risk profiles and vascular lesions, and these lesions are indicative of whether a demented individual has AD or VaD. For seven individuals, the interval between headache reporting in HUNT 2 and the diagnosis of dementia was less than five years. For those few individuals it is possible that a condition of dementia was causing headache, and not headache that later caused dementia. However, the total follow-up period was 15 years, and the mean interval until diagnosis of dementia was nearly nine years.

Nine of 63 cases of VaD self-reported previous stroke at baseline. We cannot rule out the potential mediating role of stroke in the observed association between migraine and dementia, because migraine is a risk factor for stroke (6) and stroke is a risk factor for dementia (4). Adjustment for self-reported previous stroke at baseline was performed and did not influence the estimates. However, information about stroke during follow-up was not available in our data file. Because stroke is a major part of the diagnostic criteria of VaD, stroke during follow-up could be an important explanation of the relationship between any headache and VaD. This possible explanation may seem more plausible since we found the strongest association between any headache and VaD in those over the age of 75 years, where the risk of stroke is high. It might be argued that people who experience headache, particularly frequent headache and migraine, are be more likely to seek medical care and therefore have more opportunities to be diagnosed with dementia. If this is true, the observed increased risk of dementia among those with any headache may not be real. Given the small number of cases of VaD, even a small bias in the diagnosis of dementia could cause a large effect. However, although this explanation cannot be ruled out, it should be mentioned that very few of the headache sufferers interviewed at HUNT 2 had consulted a doctor (18).

Although an association between any headache and VaD has never before been demonstrated, some indirect evidence exists that makes such a relationship plausible, at least for migraine. White matter hyperintensities and cardiovascular disease are both more prevalent in migraineurs, in particular those with migraine with aura, than in the general population (6 –9). This is of relevance since white matter hyperintensities and cardiovascular disease are both associated with increased risks of dementia (6,29 –32). The underlying mechanism of the increased cardiovascular risk in migraineurs is still unknown, but probably not mediated by atherosclerosis (33).

The fact that any headache in the present study was associated only with VaD (and mixed dementia for men) and not with other types of dementia may give credibility to our results. A lack of differential influence between migraine and nonmigrainous headache was found. Notably, in other studies based on data from the HUNT 2 study, both migraine and nonmigrainous headache were associated with unfavorable vascular risk profiles (16,17). Accordingly, the possibilities exist of shared environmental factors or shared underlying genetic susceptibilities (34), even though migraine and nonmigrainous headache stood out as independent risk factors when adjusted for available environmental factors.

The observed relationships between migraine and nonmigrainous headache on the one hand and VaD on the other need to be confirmed in other studies because of their major implications. If these headache disorders are indeed risk factors for VaD, the underlying mechanisms must be elucidated. A key question will be whether pharmacological treatment of headache (which changed, at least for migraine, with the introduction of triptans in the early 1990s) may induce vascular changes. Ultimately, treatments with potential efficacy against both headache and the development of VaD might be considered more frequently.

Strengths and limitations of the study

The major strengths were the prospective design, a follow-up period of 15 years and the large cohort recruited from the adult population of an entire county. Furthermore, we had validated the questionnaire-based headache diagnoses (18), which had good sensitivity and high specificity for migraine. The diagnoses of dementia and its different subtypes were carefully verified in all patients by a team of clinical experts in dementia. In the multivariate analyses, there were data available to allow adjustments for a large number of potential confounding factors. Nevertheless, the possibility of residual confounding by unrecognized factors cannot be wholly excluded.

Two limitations should be considered. First, despite the large study size, some subgroups of headache and dementia were relatively small. Thus, separate evaluation of, for example, medication-overuse headache was not possible, although this would have been of interest. Furthermore, few individuals who suffered from headache on ≥15 days/month or migraine with aura developed VaD, and the results for frontotemporal dementia and dementia with Lewy bodies should also be interpreted with caution. Second, there was a relatively low participation rate at baseline in HUNT 2 (56%), introducing a possibility of unknown selection bias, although the wide scope of the HUNT study, which addressed a broad range of medical disorders, made biases with particular relevance to headache unlikely (21,27). Individuals with cognitive impairment might have been less likely to participate in HUNT 2, but this type of selection bias is not a limitation of the present study, since the population at risk was defined as those susceptible to but free of dementia at the time of HUNT 2.

Conclusion

In this large population-based cohort study, any headache was a risk factor for the development of VaD, and the relationship was similar for migraine and nonmigrainous headache. No associations were found between headache and the development of AD.

Clinical implications

Headache is associated with an increased risk of vascular dementia (VaD).

The relationship between any headache and VaD is similar for migraine and nonmigrainous headache, and most evident for any headache on ≥15 days/month and age ≥75 years.

No association was found between headache and Alzheimer’s disease.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

None declared.

Contributors

KH designed the study, made the statistical analyses, drafted and revised the paper. He takes responsibility for the integrity of the data and the accuracy of the data analysis. ES designed the study and revised the paper. ML, TS and LJS all drafted and revised the paper.

Acknowledgments

The Nord-Trøndelag Health Study (HUNT) is a collaboration between the HUNT Research Centre, Faculty of Medicine at the Norwegian University of Science and Technology (NTNU), the Norwegian Institute of Public Health and the Nord-Trøndelag County Council.

References

Olesen

Gustavsson

Svensson

. The economic cost of brain disorders in Europe. Eur J Neurol 2012; 19: 155–162.

Anstey

Cherbuin

Budge

. Body mass index in midlife and late-life as a risk factor for dementia: A meta-analysis of prospective studies. Obes Rev 2011; 12: e426–e437.

Qiu

Gatz

. Mid- and late-life diabetes in relation to the risk of dementia: A population-based twin study. Diabetes 2009; 58: 71–77.

Sahathevan

Brodtmann

Donnan

. Dementia, stroke, and vascular risk factors; a review. Int J Stroke 2012; 7: 61–73.

Stovner

Hagen

Jensen

. The global burden of headache: A documentation of headache prevalence and disability worldwide. Cephalalgia 2007; 27: 193–210.

Etminan

Takkouche

Isorna

. Risk of ischaemic stroke in people with migraine: Systematic review and meta-analysis of observational studies. BMJ 2005; 330: 63–63.

Scher

Terwindt

Picavet

. Cardiovascular risk factors and migraine: The GEM population-based study. Neurology 2005; 64: 614–620.

Schürks

Rist

Bigal

. Migraine and cardiovascular disease: Systematic review and meta-analysis. BMJ 2009; 339: b3914–b3914.

Bigal

Kurth

Santanello

. Migraine and cardiovascular disease. A population-based study. Neurology 2010; 74: 628–635.

10.

Kruit

van Buchem

Launer

. Migraine is associated with an increased risk of deep white matter lesions, subclinical posterior circulation infarcts and brain iron accumulation: The population-based MRI CAMERA study. Cephalalgia 2010; 30: 129–136.

11.

Gaist

Pedersen

Madsen

. Long-term effects of migraine on cognitive function: A population-based study of Danish twins. Neurology 2005; 64: 600–607.

12.

Kurth

Mohamed

Maillard

. Headache, migraine, and structural brain lesions and function: Population based Epidemiology of Vascular Ageing-MRI study. BMJ 2011; 342: c7357–c7357.

13.

Baars

van Boxtel

Jolles

. Migraine does not affect cognitive decline: Results from the Maastricht aging study. Headache 2010; 50: 176–184.

14.

Rist

Dufouil

Glymour

. Migraine and cognitive decline in the population-based EVA study. Cephalalgia 2011; 31: 1291–1300.

15.

Rist

Kang

Buring

. Migraine and cognitive decline among women: Prospective cohort study. BMJ 2012; 345: e5027–e5027.

16.

Winsvold

Hagen

Aamodt

. Headache, migraine and cardiovascular risk factors. Eur J Neurology 2011; 18: 504–511.

17.

Winsvold

Sandven

Hagen

. Migraine, headache and the development of metabolic syndrome: An 11-year follow-up in the Nord-Trøndelag Health Study (HUNT). Pain 2013; 154: 1305–1311.

18.

Hagen

Zwart

Vatten

. Head-HUNT: Validity and reliability of a headache questionnaire in a large population-based study in Norway. Cephalalgia 2000; 20: 244–251.

19.

Mykletun

Stordal

Dahl

. The Hospital Anxiety and Depression (HAD) scale: Factor structure, item analyses, and internal consistency in a large population. Br J Psychiatry 2001; 179: 540–544.

20.

Headache Classification Committee of the International Headache Society. Classification and diagnostic criteria for headache disorders, cranial neuralgias, and facial pain. Cephalalgia 1988; 8(7 Suppl): 1–96.

21.

Hagen

Zwart

Vatten

. Prevalence of migraine and non-migrainous headache—Head-HUNT, a large population-based study. Cephalalgia 2000; 20: 900–906.

22.

World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland: WHO, 1992, pp. 12--67. Available at: http://www.who.int/classifications/icd/en/bluebook.pdf (accessed 12 November 2013).

23.

World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for Research. Geneva, Switzerland: WHO, 1993. Available at: http://www.who.int/classifications/icd/en/GRNBOOK.pdf (accessed 12 November 2013).

24.

McKhann

Drachmann

Folstein

. Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under auspices of Department of Health and Human Services Task force on Alzheimer’s Disease. Neurology 1984; 34: 939–944.

25.

Roman

Tatemichi

Erkinjuntti

. Vascular dementia: Diagnostic criteria for research studies: Report of the NINDS-AIREN International Workshop. Neurology 1993; 43: 250–260.

26.

Bergh

Holmen

Saltvedt

. Dementia and neuropsychiatric symptoms in nursing-home patients in Nord-Trøndelag County [in English and Norwegian]. Tidsskr Nor Laegeforen 2012; 132: 1956–1959.

27.

Hagen

Stovner

Zwart

. Potentials and pitfalls in analytical headache epidemiological studies. Lessons to be learned from the Head-HUNT Study. Cephalalgia 2007; 27: 403–413.

28.

Morton

Tyas

. Does a history of migraines increase the risk of Alzheimer’s disease or vascular dementia? Alzheimers Dement 2012; 8(2 Suppl): P504–P504.

29.

Debette

Markus

. The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: Systematic review and meta-analysis. BMJ 2010; 341: c3666–c3666.

30.

Silbert

Nelson

Howieson

. Impact of white matter hyperintensity volume progression on rate of cognitive and motor decline. Neurology 2008; 71: 108–113.

31.

van Straaten

Harvey

Scheltens

. Periventricular white matter hyperintensities increase the likelihood of progression from amnestic mild cognitive impairment to dementia. J Neurol 2008; 255: 1302–1308.

32.

Duron

Hanon

. Vascular risk factors, cognitive decline, and dementia. Vasc Health Risk Manag 2008; 4: 363–381.

33.

Stam

Weller

Janssens

. Migraine is not associated with enhanced atherosclerosis. Cephalalgia 2013; 33: 228–235.

34.

Bigal

Kurth

. Migraine and cardiovascular disease. Possible mechanisms of interaction. Neurology 2009; 72: 1864–1871.

Headache as a risk factor for dementia: A prospective population-based study

Abstract

Background

Methods

Results

Conclusion

Keywords

Introduction

Methods

HUNT and diagnosis of headache

Diagnosis of dementia

Study populations

Statistical analysis

Ethics

Results

Discussion

Comparison with other studies

Interpretation

Strengths and limitations of the study

Conclusion

Clinical implications

Footnotes

Funding

Conflict of interest

Contributors

Acknowledgments

References