In this systematic review and meta-analysis of cohort studies, we compiled current evidence on the association between headache disorders and the future risk of dementia in 291,549 individuals from 6 population-based studies. We found that any headache is a potential risk indicator for all-cause dementia.
Our results were further supported by findings of 2 other large population-based studies that we did not include in this meta-analysis (reasons for exclusion are shown in Additional file
1: Table S3). Stræte Røttereng et al. used data from Nord-Trøndelag Health Surveys conducted during 1995–1997 (HUNT2) and 2006–2008 (HUNT3), and found that both any headache and non-migrainous headache were more likely to be reported in dementia patients, when compared with the control group (any headache, odds ratio [OR] = 1.24; 95% CI: 1.04–1.49 and non-migrainous headache, OR = 1.49; 95% CI: 1.24–1.80) [
29]. Tzeng et al. analyzed 10 years of follow-up data from the National Health Insurance Research Database of Taiwan and suggested that patients with primary headaches had twice the normal risk of developing dementia in the future (hazard ratio [HR] = 2.06; 95% CI: 1.72–2.46) [
30]. All these findings indicated that headache disorders might be a potential predictor for dementia.
It should be noted that no statistically significant result was found in the pooled analysis of the association between migraine and all-cause dementia. However, the result of this analysis was based on only three studies with two were abstracts, and as such, should be interpreted with caution. The available data on migraine and risk of AD were also unsatisfactory. Although 1 study indicated that history of migraine was associated with increased risk of AD [
25], there was insufficient evidence to draw any conclusion about this association. The current available evidence on migraine and dementia were scarce but could suggest us that the possible association might exist, and highlight the need for more population-based research on this association.
The pathological association between headache disorders and dementia remains largely unknown, but several mechanisms are speculated to be involved. First, headache is a common pain disorder. A previous study found that several brain structures involved in the pain network, such as the thalamus, insula, anterior cingulate, amygdalae, and temporal cortex, undergo morphometric changes during the disease process [
31]. Interestingly, these brain regions also play important roles in the memory network [
32]. In addition, a previous structural–neuroimaging study of chronic headache showed that the gray-matter volume of memory network structures, including the cingulate cortex, insula, prefrontal area, and parahippocampus, decreased significantly in individuals who suffered from headache compared with those who did not [
33]. These significant changes in the overlapping pain and memory networks explain the potential correlation between chronic pain and memory impairment in headache patients. Second, a previous meta-analysis found an association of white-matter hyperintensity with an increased risk of dementia [
34]. Incidentally, headache patients were reported to have an increased risk of white-matter hyperintensity [
35]. Therefore, subtle changes in the brain white-matter might contribute to an increased risk of dementia in headache patients. Third, depression is common, with approximate 20% of the general population experiencing a depressive episode during their lifetime [
36]. An association between depression and dementia has been suggested in previous studies. Headache, especially migraine, is often comorbid with depression [
37]. Specifically, previous studies found that earlier-life depression or depressive symptoms were associated with a significantly increased risk of developing dementia [
38,
39]. Vascular disease, alterations in the cortisol–hippocampal pathway, increased amyloid plaque formation, inflammatory changes, and deficits in nerve growth factors or neurotrophins are predicted to be the potential biological mechanisms linking depression to dementia [
40]. Thus, an increased risk of dementia in headache patients might be partly due to comorbidity with depression. Finally, stress and mental tension have been identified as predictors of headache disorder [
41]. A previous cohort study found an association between psychological stress in middle-aged women and the development of dementia, especially AD [
42]. The underlying mechanism remained unclear, but the hypothalamic–pituitary–adrenal axis and the effects of glucocorticoids on the brain are thought to be behind the association [
43].
To the best of our knowledge, our meta-analysis was the first to summarize the currently available evidence of the association between headache and the risk of dementia, and to indicate that any headache is a risk factor for developing all-cause dementia. Sensitivity analysis verified the stabilization of the results. However, there are some limitations as well. The number of studies included was small, and the meta-analysis of the association between any headache and the risk of AD included only 3 studies. As such, the results are likely to be imprecise [
44], and, consequently, the conclusions drawn from this study should be considered preliminary. In addition, the studies included show quite high heterogeneity in terms of study design, population sizes and population age range. The potential effect of the heterogeneity should be taken into account when interpreting the findings of the review. Even so, our findings are still of great significance. The association we found might aid in identifying people prone to dementia or cognitive decline. This emphasizes the need to reveal the mechanisms underlying the link between headache and dementia, which may become all the more evident while improving the quality-of-life of patients with headache disorder. The information is critical to finding new preventive and treatment strategies for dementia. It is also of crucial importance that we figure out whether treatment for headache disorder might intervene in the overlapping pathways and subsequently reduce the risk of dementia.