Global perspectives on the burden of episodic and chronic migraine

Recent studies have profiled chronic migraine (CM) and episodic migraine (EM) in many regions of the world using samples derived from various sources. In a previous issue of Cephalalgia, Wang and colleagues contrast patients with CM and EM recruited from two headache clinics in Taiwan (1). Their results add to the growing evidence indicating that, in comparison with EM, CM is less common but far more burdensome to individuals, as evidenced by measures of headache-related disability, medical and psychiatric comorbidities, and health-related quality of life (HRQoL), and to society, as indexed by measures of lost productive time as well as direct medical costs. In 2011, the World Health Organization and Lifting the Burden Campaign against Headache published the “Atlas of Headache Disorders and Resources in the World 2011.” The authors of “The Atlas” stated, “Our view of the global burden attributable to headache disorders is incomplete…” (2). They urged researchers around the world to gather and publish data on the prevalence, burden, and other variables related to primary headaches in order to increase the understanding of the global status of headache prevalence, burden, costs, and care. They estimated the prevalence of headache on ≥15 days per month (an approximation of chronic daily headache (CDH)) in Southeast Asia to be 1.7% based on a compilation of data from epidemiological studies. CM is a form of CDH; however, “The Atlas” did not report rates of CM.

In their recent systematic review of CDH and CM in the Asia-Pacific region, Stark and colleagues (3) identified two population studies, both from Taiwan, that reported the prevalence of CM as 1.7% (reported as transformed migraine (TM)) (4) and 1.0% (reported as CDH with migrainous features) (5). They also identified 19 clinic-based studies that reported rates and features of patients with CM. They concluded that Asia may have a lower prevalence of CM and CDH than the international average, although the rates of CM reported in the studies from Taiwan are similar to the prevalence of CM reported by the American Migraine Prevalence and Prevention (AMPP) study (6). Stark et al. also noted that CM was associated with significant headache-related disability, functional impairment, and psychiatric comorbidity in reports of clinic studies from Taiwan, Korea, and China. These findings from the Asia-Pacific region are similar to those reported by large international (7,8), North American (6), and European studies (9), which have also demonstrated that CM is associated with greater disease burden on multiple indices compared with EM.

Making comparisons among studies and variations regions of the world is difficult because of variation in sources of study subjects and data collection methods among other factors. Fortunately, the AMPP study, the International Burden of Migraine Study (IBMS), and the study by Wang and colleagues collected and reported similar data, which allows for comparisons across studies. The AMPP study screened 120,000 United States (US) households selected to be representative of the US population and had a 64.9% response rate. Of 28,261 respondents who reported experiencing “severe headache” in the preceding year, 24,000 were followed annually from 2005 to 2009 with mailed, written questionnaires. The AMPP survey instrument was based on validated items from the American Migraine Studies 1 and 2 (10,11) along with validated questionnaires such as the Migraine Disability Assessment Scale (MIDAS) (12). The IBMS utilized a targeted, Web-based methodology to recruit and survey participants with EM and CM in 10 countries. Invitations were sent via email to 72,059 panelists. Of 23,312 respondents (32.4% response rate), 555 participants with CM and 9160 with EM were eligible and completed the Phase I and Phase 2 surveys (41.7%). The IBMS survey was largely composed of items and instruments from the AMPP study surveys with additional validated instruments added and translated into multiple languages. Though Taiwan was one of the 10 countries surveyed in the IBMS study, the recruited sample was too small for meaningful analysis. Wang and colleagues aimed to gather data from individuals with EM and CM from headache clinics in Taiwan to fill this gap in knowledge. Wang and colleagues based their survey on the IBMS, with the addition of physician interviews for diagnosis and translation into Taiwanese.

In the current study, Wang and colleagues recruited 331 patients with migraine (164 with EM and 167 with CM) from two headache clinics in Taiwan. Diagnosis of EM or CM was made by physicians using Silberstein-Lipton criteria (13,14). Participants completed a questionnaire comprised of items on sociodemographics, headache-related disability, quality of life, medical and psychiatric comorbidities, lost productivity, and healthcare resource utilization, among other topics. The authors found that compared with participants with EM, participants with CM had significantly higher headache-related disability, more medical and psychiatric comorbidities including higher scores on a standardized depression and anxiety screening instrument (Patient Health Questionnaire-4 (PHQ-4)), greater likelihood of endorsing a history of a range of medical and psychiatric comorbidities, increased healthcare resource utilization, and greater work productivity loss. These results are compatible with results seen in the AMPP and IBMS (6–8,15–19). In addition, participants with CM had worse health status scores and reduced HRQoL, findings that are comparable to those observed in the IBMS (7,8).

Although Wang et al., the IBMS, and the AMPP all demonstrated that people with CM have significantly greater headache-related disability than individuals with EM, determined by the MIDAS, there were differences among the studies (see the Table 1). Based on the assumption that patients within a clinic generally have more severe disease than individuals within the population, it is not surprising that headache-related disability (MIDAS score/grade) was higher for individuals with CM and EM from the Taiwanese clinic sample than in the US population sample. Among participants with CM, 41.3% of those from Wang et al. had “very severe” headache-related disability (MIDAS Grade IV-B) (7) compared with 24.8% of those in the AMPP study. Similarly, 7.9% of those with EM from the Wang et al. study and 3.2% of those from the AMPP study had scores in the MIDAS Grade IV-B category. Although the absolute rates of “very severe” headache-related disability differed across studies, the proportions of “very severe” headache-related disability within each sample (comparing the CM to the EM group within a study) were fairly similar. In the Wang et al. manuscript, patients with CM were 5.2 times more likely to have “very severe” headache-related disability than participants with EM. In the AMPP study, participants with CM were 7.8 times more likely to have “very severe” headache-related disability than participants with EM.

Table 1.

Selected data from Wang et al., IBMS and AMPP study.

Variable	Wang et al. study		International Burden of Illness Study (IBMS)		American Migraine Prevalence and Prevention (AMPP) study
	CM	EM	CM	EM	CM	EM
Sociodemographics and headache frequency 1,2
Average age (years)	42.2	40.3	41.2	39.7	47.7	46.0
% Female	76.6%	78.7%	87.4%	81.3%	78.6%	80.0%
% Obese (BMI ≥30)	3.0%	4.3%	24.3%	19.4%	25.5%	21.0%
% Employed full time	52.7%	64.0%	35.5%	49.3%	37.8%	52.3%
% Married	60.5%	64.6%	49.9%	46.8%	59.3%	60.8%
% Completed bachelor’s degree and higher	38.9%	45.1%	28.0%	34.5%	28.2%	31.7%
Income	NT$1,300,001 (approx. $43,333 US)  = 8.4%	NT$1,300,001 (approx. $43,333 US)  = 9.8%	>Median county income  = 28.5%	>Median county income  = 31.9%	Income ≥$60,000/ yr = 32.6%	Income ≥$60,000/ yr = 36.3%
Days in the last month with headache of any intensity	22.0	4.6	—	—	—	—
Days in the last three months with headache of any intensity1	—	—	64.4	10.4	—	—
Headache-related disability 3,4
Average MIDAS sum score (headache-related disability)	46.1	14.4	72.6	14.5	—	—
% MIDAS Grade IVB (very severe)	41.3%	7.9%	63.5%	6.3%	24.8%	3.2%
% MIDAS Grade I (none or minimal)	21.6%	43.9%	6.8%	30.7%	42.7%	65.1%
Lost productive time per day with headache (MIDAS score/ number of headache days in past three months)6	46.1/66.0 = 0.70	14.4/13.8 = 1.04	72.6/64.4 = 1.13	14.5/10.4 = 1.39
Quality of life 3,5
MS-QOL- RFR	56.4	70.8	61.4	71.7	—	—
MS-QOL- REP	70.1	81.4	44.4	56.5	—	—
MS-QOL_ EF	62.0	78.1	48.3	67.2	—	—
Comorbidities 1,2
Number of comorbid conditions	2.1	1.7
SR-PD pain comorbidities (group)	40.7%	20.1%	38.0%	18.2%
SR-PD cardiovascular disease risk factors (group)	29.9%	22.0%	41.6%	33.7%
SR-PD cardiovascular disease events (group)	13.2%	5.5%	8.5%	3.5%
SR-PD psychiatric disorders (group)	43.1%	24.4%	45.6%	29.7%
SR-PD chronic pain	38.3%	18.9%	26.5%	10.6%	31.5%	15.1%
SR-PD high blood pressure	19.8%	9.8%	20.4%	13.9%	33.7%	25.6%
SR-PD arthritis	7.2%	1.8%	16.9%	9.0%	33.6%	22.2%
PHQ-4 (depression + anxiety) = severe3	17.4%	1.3%	21.2%	6.6%	Depression with PHQ- 9 = 30.2%	Depression with PHQ- 9 = 17.2%
SR-PD depression	28.7%	6.7%	33.7%	17.8%	41.2%	25.6%
SR-PD anxiety	32.3%	20.7%	34.2%	21.7%	30.2%	18.8%

NT: New Taiwan dollars. Exchange rate as of August 2012 was approximately $1 US = 30 NT dollars; SR-PD: respondent self-report of ever receiving a physician diagnosis of a given condition; CM: chronic migraine; EM: episodic migraine; BMI: body mass index; ¹IBMS data from Payne et al., 2011; ²AMPP data from 2005 AMPP survey, Buse et al., 2010; ³IBMS data from Blumenfeld et al., 2011; ⁴AMPP data from 2004 AMPP survey, Buse et al., 2012; ⁵Quality of life data was assessed with the Migraine Specific Quality of Life instrument, version 2.1 (MSQ v2.1). RFP: role functioning preventive; RFR: role functioning restrictive; EF: emotional functioning. Higher scores indicate better quality of life; lower scores indicate worse quality of life; ⁶Estimated from Wang et al. by multiplying number of days in past month by three. MIDAS: Migraine Disability Assessment Scale; PHQ-4: Patient Health Questionnaire.

Levels of headache-related disability were highest in the IBMS. Higher levels of disability in the IBMS compared with the AMPP study are not surprising owing to possible effects of the targeted recruitment strategy, which likely led to a greater participation of participants with more severe disease in the IBMS. However, higher rates of disability in the IBMS than in Wang et al. are surprising as the Taiwanese sample was derived from a specialty-care setting where increased disability is likely. Interestingly, when comparing disability per headache day (MIDAS sum score for three months/number of headache days in three months), participants from the Asian sample reported less lost productive time than those in the IBMS (Wang et al. sample: CM = 0.70 lost day per headache day, EM sample = 1.04 lost days per headache day) compared with those from the IBMS (CM = 1.13 lost days per headache day, EM = 1.39 lost days per headache day). Within both studies, participants with CM reported less lost productive time in days than participants with EM per headache. The finding that participants in the Wang et al. study had less lost productive time in days per headache may reflect a reduced burden of CM and EM in Asian populations, a cultural difference in symptom reporting or illness behavior, the benefits of treatment provided in headache centers, selective participation in the IBMS, or a combination of factors.

A comparison of similarities and differences of these three studies of persons with EM and CM raises interesting questions. It is difficult to disentangle three possible influences: 1. race, 2. reporting and 3. the source of the sample. Variation between the three samples may reflect genuine epidemiological differences based on factors such as race-related genetic differences, social norms, healthcare access, diet, and symptom reporting. Differences between the Asian and Western populations could arise because of genetic differences in pain expression or illness severity. A US study found that the prevalence of migraine in Asian-Americans was 50–60% that of Caucasians (20). Some research has also identified race or ethnic differences in pain perception and response, although differences in physiological pain thresholds have not been consistently replicated (21). Differences could also arise because of cultural differences in symptom reporting, though that is very difficult to examine without specifically designed studies. Differences may also be due to the sources of the samples. In general, patients who seek specialty medical care have more severe disease than those in the general population (22). Interestingly, this was not the case in terms of comorbidities, where 38.0% of those with CM in the IBMS study and 40.7% of those in the Wang et al. study reported one or more pain conditions (other than headache), and 45.6% of those with CM in IBMS and 43.1% of those with CM in Wang et al. reported being diagnosed with depression or anxiety by a physician. These findings are in line with a recent Brazilian study that found similar rates of depression between people with CM from a clinic (n = 43) and those sampled from the community (n = 41) (23). Using a structured clinical interview, psychiatrists found that 39.3% of patients with CM in a tertiary-care setting and 32.6% of residents in a city borough met Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria for depression.

The study by Wang and colleagues is an important addition to the scientific literature. It offers valuable data from Taiwan, expands our perspective on patterns of disease severity, disability, and comorbidity of CM and EM, and allows for comparisons with studies with similar variables and instruments. Comparisons across cultures and regions of the world are challenging because of possible differences in illness biology, culture and treatment. However, the use of similar items and instruments between the AMPP study, the IBMS and the present report by Wang et al. facilitates comparisons across samples of different types (clinic vs. population) and among different regions of the world. In this review, we find that patients with CM and EM in Wang et al. had greater headache-related disability than seen in the population-based AMPP study. However, disability and comorbidity rates were higher in the IBMS than in Wang et al. Wang and colleagues also report data on health resource utilization, quality of life, medical and psychiatric conditions, and other important indices of disease severity and burden. Broadly, the results show that EM and CM are both disabling and both disrupt quality of life and are associated with medical and psychiatric comorbidities, yet the burden is consistently greater among those with CM. Based on this study, the studies summarized in the Table and on an expanding body of literature, we conclude that the burden of CM greatly exceeds that of EM in North America, Europe, and Asia, particularly Taiwan. In comparison to people with EM, those with CM have greater individual burdens, as measured by lost time due to headache, unemployment, and underemployment as well as reduced HRQoL. They also produce a greater societal burden, as measured by direct and indirect medical costs. Future work should continue to gather data on the burden, comorbidities, and direct and indirect costs of EM and CM in geographic regions around the world in a variety of settings to continue to expand our knowledge and understanding of migraine.