Abstract
Ferrante et al. (2012) report on the epidemiology of migraine in a non-clinical sample of adults living in Parma, Italy (1). A face-to-face interview of each of the 904 participants was conducted by a headache specialist using a validated questionnaire followed by a neurological examination. Migraine, including chronic migraine and migraine with aura, was diagnosed in accordance with the International Classification of Headache Disorders, 2nd edition (ICHD-2) criteria (2). They found an age-adjusted, past-year prevalence of migraine of 24.7% (13% for men and 32.9% for women) (1). In the worldwide review by Stovner et al. (3), looking only at the 14 European studies, ICHD-1 migraine (4) was prevalent in about 15% of adults (range 10–25%), thus suggesting that migraine is almost twice as prevalent in Parma compared to these earlier studies.
Factors that may affect migraine prevalence include study methodology (interview type, screening questions, diagnostic criteria) as well as population differences (socio-economic status, some lifestyle factors, genetics, ethnicity). A nationwide, multicenter study of 71,588 patients interviewed by 902 GPs in Italy using an ICHD-1-based questionnaire found a 11.6% prevalence of migraine (6.9% definite migraine + 4.6% probable migraine), which is much lower than that found in the Ferrante et al. (2012) study; results were not reported by geographical region (5). Might the prevalence of migraine in Italy vary by geographical region? For example, a nationwide, population-based study of migraine in Spain showed an overall migraine prevalence of 12.6% but significant geographical variation (> twofold) was reported, from 7.6% in Navarra to 18% in the Canary Islands (6).
Some studies have found a higher prevalence of migraine among those of lower socio-economic status (7–10). However, central Parma, where the study sample is derived from, may be associated with higher socio-economic status and Parma overall is considered to be one of the more affluent cities in Italy (11). If that is the case, it is possible that prevalence of migraine in this part of Parma would be expected to be lower than in the rest of Parma and lower than in Italy in general. On the other hand, the sensitive screening question (‘Have you had a headache in the last year?’) followed by a headache specialist examination could contribute to a higher prevalence.
While the epidemiology of migraine has been studied in more than 100 studies using modern diagnostic criteria (3), the measurement of two clinically important migraine subtypes (migraine with aura and chronic migraine) in population samples remains problematic. This is due, we believe, to some degree of uncertainty about these phenotypes as well as the difficulty of operationalizing these clinical constructs in large-scale population studies (12–15). This latest study is one of a fairly small number of migraine epidemiology studies using physician interviewers exclusively (16–20), and a clinical examination of 904 participants from a non-clinical population is a commendable effort.
The Ferrante et al. (2012) study reports a prevalence of 0.2% (0.0% for men and 0.4% for women) for (ICHD-2) chronic migraine, possibly on the low side compared to earlier studies using these restrictive criteria (6,21). The definition and hence epidemiology of very frequent headache/chronic migraine has of course evolved over the last 20 years or so. First, there was transformed migraine described by Mathew (22) and Mathew et al. (23) followed by the criteria outlined by Silberstein et al. (24) for chronic daily headache (CDH) with migraine features, and then ICHD-2 chronic migraine. ICHD-2 chronic migraine is perhaps one-third as common as the CDH/migraine or equivalent described by Silberstein et al. (24), which is typically in the range of 1.4–2.2 (12). Revised criteria for chronic migraine were published as an appendix to ICHD-2 (25) and called the International Classification of Headache Disorders, 2nd edition, revised (ICHD-2R). The revised criteria have been reported to address most of the criticism towards ICHD-2 (26) and will be included in ICHD-3 which is tentatively scheduled for release in January 2013 (27).
The different case definitions of chronic migraine mentioned above have been field tested and their epidemiological profile has been shown to be similar (28). Subjects with chronic migraine (all definitions) had a different epidemiological profile compared to subjects with episodic migraine; chronic migraine subjects were older, had higher body mass index (BMI), had lower socio-economic status, and were more likely to smoke (28). The American Migraine Prevalence and Prevention (AMPP) study has also shown that chronic migraine is associated with increased comorbidity compared to episodic migraine (29). Future observational and interventional studies may provide justification for further expansion or refinement of the chronic migraine case definition.
There has been less consistency in the prevalence of migraine with aura (MA) compared to migraine without aura (MO), possibly due to genetic or other differences between populations but also likely due to the aforementioned difficulty in measuring the migraine aura in large-scale population studies. Needless to say, this methodological problem is increasingly important given the body of recent literature linking MA with cardiovascular disease and other outcomes (30–32). In the Ferrante et al. (2012) study, the crude 1-year prevalence of MO, MA, and MO + MA was 17.0%, 2.9%, and 1.5% respectively. A Danish study by Rasmussen and Olesen (18), where a clinician diagnosis was also performed, showed a 1-year prevalence of 5.3% for MO, 3.4% for MA, and 0.7% for MO + MA among 740 participants. A Croatian study by Zivadinov et al. (16), where diagnoses were made by trained medical students supervised by a neurologist, reported a 1-year prevalence of 9.4% for MO, 5.9% for MA, and 1.5% for MO + MA among 3794 participants. If subjects with both MA and MO are counted as MA, then the proportion of migraineurs with aura is 20.4% in the Ferrante et al. (2012) study, which is lower than the Danish and Croatian studies (43.5% and 44.2% respectively) (16, 18).There are two population-based studies which may be added for comparison, the first one is a study from the United States (US) using an extensive lay interview (33) and the second one a study from the Netherlands using physician-supervised lay interviews where subjects drew their visual auras (34). In the US study by Stewart et al. (33), the proportion of migraineurs with aura was 24% (derived from Stewart et al. (33)) while in the Dutch study by Launer et al. the proportion was 31% (34). Thus, overall, the proportion of migraineurs in these general adult populations classified as having MA has varied considerably from 20–40% in the clinician interview studies and 24–31% in the two lay/neurologist interviews.
There is still a lot to be explored in mapping the prevalence of all forms of migraine, particularly chronic migraine and migraine with aura, and to examine the determinants of the observed variability across countries as well as within countries.