By Dara Jamieson, MD

Associate Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Jamieson reports she is a consultant for Bayer and Boehringer-Ingelheim.

SYNOPSIS: Non-migrainous headaches, for which there are many causes, appear to be a risk factor for stroke in an elderly population, but the mechanism is uncertain.

SOURCE: Norton J, Portet F, Gabelle A, et al. Are migraine and non-migrainous headache risk factors for stroke in the elderly? Findings from a 12-year cohort follow-up. Eur J Neurol 2016;23:1463-1470.

The authors expanded the proposition that migraine is a risk factor for stroke, examining the incidence of stroke in an elderly population with migraine and with the more common non-migrainous headache (NMH). Invitation letters were sent to randomly selected community-dwelling persons, aged 65 years and over, living in Montpellier, France, between March 1999 and February 2001. These elderly persons were invited to attend a half-day clinical examination to check eligibility for the retrospective study. The 2,259 eligible subjects who responded were interviewed about their medical history and underwent a neuropsychiatric interview and a neurological examination. Among subjects reporting headaches, a diagnosis of NMH was made only after excluding a diagnosis of migraine, as based on the International Headache Society (IHS) criteria. After recruitment, all subjects were to be followed up at 2, 4, 7, 9, and 11 years. The 136 subjects who were lost to follow-up, and thus excluded from the analysis, were older, more disabled, and less educated with lower income, with more vascular risk and cognitive impairment. Despite more medical and social impairment, these excluded participants had no significant differences in current or lifetime NMH and migraine. The 1,919 remaining subjects with no history of stroke at baseline and no missing values for the main covariates were followed for stroke incidence for a median follow-up period of 8.8 years. At each follow-up examination, subjects reported neurological events that occurred since the previous visit. Strokes, either hemorrhagic, ischemic, or unknown, were adjudicated, but brain imaging, mainly computerized tomography, was available for “more than 80% of validated stroke cases.” Lifetime migraine by IHS criteria was reported in 17.4% and current migraine was reported in 5.4% of the elderly subjects. The diagnosis of NMH, made during their lifetime, was 11.4% of subjects, and was diagnosed in 8.9% currently. The majority of subjects were said to report only one type of headache. The NMH diagnoses were varied: “tension headaches” 36.5%, “rheumatology-related” 25.1%, “Arnold’s neuralgia” (occipital neuralgia) 12.9%, “hypertension-related” 4.5%, “glaucoma-related” 3.3%, “trigeminal neuralgia” 3.3%, “intra-cranial” 3.3%, “ear, nose, and throat-related” 2.8%, “histaminic cephalalgia” 2.2%, and “other aetiologies” 6.1%. In the elderly subjects with a migraine history at study recruitment, 1.9% (2/106) had a stroke during the follow-up period, as compared to 6.2% (10/161) of the baseline NMH sufferers, and 4.3% (11/258) of subjects with a past history of migraine or NMH. Cox proportional hazard models indicated that current migraine history in the elderly population (mean age for migraine and NMH: 72 years) was not a risk factor for stroke; however, NMH sufferers were twice as likely to have a stroke (hazard ratio, 2.00; 95% confidence interval, 1.00-3.93; P = 0.049).


Convincing epidemiological studies have shown that migraine with aura is a risk factor for ischemic stroke in a younger population, most notably in women. Yet stroke is more common in an older population, whose headaches, including migraine, are likely to have dissipated with age. Norton et al evaluated a possible correlation between headaches, both migraine and non-migraine, and stroke in an older population. Unfortunately, a small, selected population of older individuals and a lack of granularity in disease categories, lead to few viable conclusions from these data.

The authors commented on some of the limitations of this retrospective epidemiological study. The small number of strokes detected in follow-up (n = 73), with only two strokes in elderly subjects with migraine, may obscure the correlation that was found in studies with a larger population. Because of the small number of elderly migraineurs in the study, the population was not stratified according to migraine without or with aura or to sex, variables of importance in other epidemiological studies. The authors noted the heterogeneous mix of NMHs, including both primary and secondary headache types. The number of subjects with each headache type was too small to do a sub-group analysis. Likewise, because of the small number of events, there was no sub-analysis according to stroke type. The correlation with migraine is much more robust for ischemic stroke than for intraparenchymal hemorrhage. Given that these details may differentiate between a connection and a coincidence, the study population was too small to make a clear conclusion about a migraine and stroke linkage in the elderly. Another limitation of the study is that the population analyzed was self-selected participants who volunteered for the study, and more disabled individuals were lost to follow-up. The population followed over the long-term were more likely to be the healthy elderly with headache, as opposed to a more representative cross-section of the elderly.

One important inference that can be made from these data is that headache in the elderly deserves investigation and monitoring. The authors suggested that elderly individuals with NMH be followed closely because of an increased stroke risk. In general, headaches are less common in the elderly and, if present, are more likely to be secondary to an underlying systemic disease or identifiable brain lesion. Therefore, secondary headaches in the elderly may be a marker for poor health in general, and may be indicative of greater cerebrovascular risk specifically. Why elderly individuals with multiple headache types should have a somewhat greater stroke risk than those with migraine, the tenuous conclusion from this study, does not have any mechanistic explanation.