Quality of life impairment,disability and economic burden associated with chronic daily headache,focusing on chronic migraine with or without medication overuse: A systematic review

Abstract

Objective : To evaluate the evidence for quality of life (QoL) impairment, disability, healthcare resource use and economic burden associated with chronic daily headache (CDH), focusing on chronic migraine (CM) with or without medication overuse.

Methods : A systematic review and qualitative synthesis of studies of patients/subjects with CDH that included CM, occurring on at least 15 days per month.

Main findings: Thirty-four studies were included for review (25 studies of patients and nine of subjects from the general population). CDH and CDH with medication overuse headache (MOH) were consistently associated with a lower QoL compared to control or episodic headache (EH) and CDH without MOH. CDH was consistently associated with greater disability and productivity loss, more consultations, more or longer hospitalizations and higher direct costs than EH. Data were not amenable to statistical pooling.

Principal conclusions : The findings of this review underline the detriment to QoL and the disabling nature of CDH, and in particular CM and CDH with MOH, and negative impact on workplace productivity compared to other types of headache.

Introduction

Chronic daily headache (CDH) refers to a heterogeneous group of headache disorders characterised by a headache occurring on 15 or more days per month for more than 3 months (1). The overall prevalence of CDH has been frequently reported to be approximately 3–4% of the adult and elderly population in most countries worldwide (2–7).

In practice, CDH is mainly represented by chronic migraine (CM) and medication overuse headache (MOH). In the second edition of the International Classification of Headache Disorders (ICDH-II), CM is classified as a complication of migraine and its original criteria required headaches to meet criteria for migraine without aura on ≥15 days/month, for ≥3 months, without medication overuse (8). Classified as a secondary headache, MOH is defined as a headache present on ≥15 days/month, with regular overuse for >3 months of one or more drugs used for acute and/or symptomatic headache treatment. Criteria also required MOH to worsen during the period of medication overuse, and resolve or revert to its previous pattern within 2 months of discontinuation of the offending agent (8).

Implementation of these classifications was criticized in clinical practice as being overly restrictive for a diagnosis of CM, as it did not apply to most patients with CDH and proved cumbersome for a diagnosis of MOH (9). Subsequently, in 2006, an appendix to ICHD-II defined more inclusive criteria for CM and MOH. So, CM can be diagnosed in migraine patients with headaches on ≥15 days/month, for ≥3 months, of which ≥8 of the days fulfil criteria for migraine without aura or were successfully treated with ergots or triptans, and remaining patients diagnosed using the general system of multiple diagnoses, hence clarifying that CDH and TM (transformed migraine) do not represent a diagnosis but refer to the evolution of headache over time (10). Meanwhile CDH, CM and TM continue to be terms frequently used interchangeably in the literature, as reflected also by this review. Criteria for MOH were also revised to eliminate the requirement that the headache resolves within 2 months of medication withdrawal (10).

Headache and migraine disorders impact a person’s quality of life (QoL) and contribute to a growing financial burden on healthcare systems. However most large reviews in this field have focussed on quantifying the burden of episodic forms of migraine and headache. The objective of this systematic review therefore was to identify, summarize and evaluate the current body of evidence on QoL impairment, disability, healthcare resource use and cost burden associated with CDH, particularly CM with or without medication overuse.

Methods

A systematic review of the literature was undertaken to identify studies describing assessments of the QoL, disability, productivity loss, healthcare resource utilization and financial burden associated with CDH and subtypes including CM and MOH. Systematic searches of the MEDLINE, PREMEDLINE, EMBASE and The Cochrane Library electronic databases were initially conducted in November/December 2008 and updated in July 2009 for the period 1997 forward. The 1997 cut-off was considered appropriate as it allowed for the application of modern criteria for CDH diagnosis from Silberstein et al. forward (11,12). Search terms reflected current and past chronic headache classification terminology. The main concepts for the searches were: chronic daily headache, chronic migraine, transformed migraine, medication or analgesic overuse headache, QoL, disability, healthcare resource utilization, economic burden and costs and burden of illness. A grey literature search was also conducted with assistance from authors GD and MLM. No language restriction was applied.

Criteria for the inclusion of studies based in the review were:

Must be a study of CDH described within the article as occurring with a frequency of at least 15 days per month over a duration of at least 3 months, or a study where CDH was defined by citation of modern criteria (1,11,12), or a study of any form of long duration CDH as defined by ICHD-II criteria (8,10), or other published criteria concurrent with the study that were consistent with this headache frequency. The umbrella term of CDH therefore allowed inclusion of studies of CDH with or without migraine features, CM, TM, intractable (chronic) headache, CDH or CM with MOH, chronic cluster headache (CCH) and chronic tension-type headache (CTTH), as long as these headaches could be established as fulfilling the specified frequency inclusion criteria. However, since the focus of the review was CDH that included CM, any studies exclusively describing CCH and/or CTTH patients were omitted.

It should be noted that some relaxation of this criterion was exercised to allow inclusion of studies where headache frequency was described as being nearly or almost daily or 14 days or more per month.

Must be a study reporting one or more of the following: QoL outcomes, an economic evaluation, cost of illness, financial burden, health resource utilization, or disability or productivity changes.

Studies of drug treatment interventions were excluded from this review.

Study selection for the review involved a systematic screening process over three phases: title, abstract and full text. At each stage, studies were rejected only if they failed to meet the inclusion criteria. All reviewing and data extraction was undertaken by two independent researchers (authors MM and SC). Disputed inclusions/exclusions/data extractions were decided upon by consensus agreement between the two researchers. At the outset, a quantitative synthesis was intended. However, the heterogeneous nature of CDH and of the respondents/patient populations in the included studies precluded any meaningful statistical pooling for any outcome. Accordingly, the synthesis of data was purely qualitative. A protocol for this review can be provided upon request.

Results

Literature search

The systematic search of electronic library databases and the grey literature identified a total of 43 publications that described 34 unique studies relevant to the review (4,6,13–53). A summary of the characteristics of included studies is provided in Table 1 (included in the online supplementary material).

Quality of life

QoL outcomes were reported in 16 studies (4,15, 21, 23, 25, 29, 33, 37, 41–43, 45, 47,50–52). Twelve studies (11 prospective, 1 retrospective) were of patients enrolled at specialist referral centres or a GP headache clinic (15,21,23,25,29,41–43,45,47,50,52). Four studies were cross-sectional surveys of subjects with headache recruited from the general population (4,33,37,51). The average age of the patient/respondent groups with CDH ranged from around 33 to 50 years. Generally, women greatly dominated men in terms of representation with CDH. The reporting of analgesic overuse and the presence of comorbidity was poor: only six of the 12 patient-based studies reported numbers of patients with medication overuse: two were specifically of patients with medication overuse (23,25), one study was of patients with a history of medication overuse (47) and in the remaining three studies the proportions of CDH patients with analgesic overuse ranged from 41% to 51% (29,43,52). In the three population studies that provided information, proportions of respondents with analgesic overuse were 25%, 32% and 67% (33,37,51), although the latter study cannot be considered representative (51). The proportions of respondents/patients with one or more specifically named comorbidities were reported in six studies only (25,29,33,37,42,45). Named comorbidities were specifically excluded from a further two studies (15,51).

The most widely used QoL instrument was the MOS 36-Item Short-Form Health Survey SF-36 (11 studies) (15,21,23,29,33,41–43,45,47,52). The SF-12, an abbreviated version of the SF-36, was used in one study (37). Other less widely used instruments included the Spitzer QL index (25), the Qualite de Vie (QMV) instrument (4,14,26), the QoL/Wellbeing assessment (50) and the Migraine Quality of Life Questionnaire (MQoL) (47). Brief descriptions of these instruments are provided in online supplementary material. One further study that conducted open-ended interviews is not further discussed within this review (51). For brevity, the review focuses on the 12 studies that used the SF-36/SF-12.

SF-36 comparisons of CDH sufferers with healthy people

Relative to healthy people and normative controls, CDH sufferers show consistently poor scoring across most domains of the SF-36 and in particular the domains of role physical and bodily pain.

Five studies reported QoL comparisons of CDH patients/respondents (or CDH subtypes) with either healthy controls or local normative data (15,23,33,41,52). The data consistently suggested an inferior QoL for CDH sufferers in most domains. Differences in item scores were frequently both statistically significant and clinically meaningful, i.e. a difference in item score of >5 points (54) (Figure 1A).

Figure 1.

Differences in quality of life measures using SF-36 or SF-12. SF-36 or SF-12 item score differences shown from seven different studies for the following comparisons: (A) CDH vs. healthy or normative controls [CDH vs. HC (33); CDH vs. NC (15); CM vs. NC (23); TM vs. NC (41); TM (females) vs. NC (females) (52)]. (B) CDH vs. EH sufferers [CDH vs. EM (43); CDH vs. EH (15,21); TM vs. EM (33,41,52)]. (C) CDH with migrainous features vs. CDH without migrainous features (CM vs. CCH (23); TM vs. CTTH (33); TM vs. other CDH (52)]. (D) CDH with medication overuse vs. CDH without medication overuse [CDH with medication abuse vs. CDH without medication abuse (33); CM vs. probable CM plus medication overuse headache (37)]. For a comparison X vs. Y, a negative difference indicates relative impairment for X; an item score difference >−5 or +5 is considered clinically meaningful (54). *, ** and *** indicates that the difference in item score is statistically significant (p ≤ 0.05, p ≤ 0.01 and p ≤ 0.001, respectively). NS indicates that the difference in item score is not statistically significant. - indicates that no statistical comparison was undertaken. Abbreviations: PF, physical functioning; RP, role physical; BP, bodily pain; GH, general health; VT, vitality; SF, social functioning; RE, role emotional; MH, mental health. CDH, chronic daily headache; EH, episodic headache; CM, chronic migraine; NC, normative control data; CCH, chronic cluster headache.

Across the five studies, the most obvious and consistent declines in QoL in CDH sufferers were in the domains of role physical and bodily pain, followed by vitality and social functioning. Absolute scores in these domains were noticeably lower in CDH patients in studies in a clinical setting (15,23,41,52). Discrepancies between the scores for physical functioning and role physical were a consistent feature throughout the studies.

SF-36 comparisons of CDH and EH sufferers

Relative to EH sufferers, CDH sufferers show consistently poor scoring across most domains of the SF-36 and in particular the domains of general health, vitality and role emotional.

Six studies reported QoL comparisons between patients/respondents with CDH (or CDH subtype) and patients/respondents with EH (15,21,33,41,43,52). The results consistently showed an inferior QoL for CDH sufferers. Across the studies, most consistent differences were observed in the domains of general health, vitality and role emotion (Figure 1B).

SF-36 comparisons of CDH, with and without migrainous features

Three out of four studies suggest that CDH sufferers who exhibit migrainous features have lower SF-36 QoL scores than those who do not exhibit migrainous features.

Four studies reported QoL comparisons for CDH without migraine or migrainous features (23,29,33,52). Differences in domain scores for three studies are presented in Figure 1C [actual scores were not reported for the remaining study (29)]. Three out of the four studies suggested an inferior QoL for those patients who had CDH with migraine or migrainous features compared to some other CDH subtypes (23,29,52). This was particularly pronounced in the D’Amico et al. study (23) where all of the patients with migrainous features were also medication overusers. In contrast, Guitera et al. (33) reported neither a statistically significant difference nor any consistent directional difference between scores across the domains between respondents with CTTH and TM. Furthermore, respondents who were CTTH sufferers had noticeably lower scores for physical functioning and general health compared to respondents who were TM sufferers, and these differences were large enough to be clinically meaningful. The authors suggested that this could be related to the older age of the CTTH sufferers. The discrepancy between these studies may also be explained by the differences in study setting: the Guitera et al. (33) study took respondents from the general population, whereas the other three studies were of patients recruited from treatment centres who may overall represent more severe headache sufferers.

SF-36/SF-12 scores in CDH sufferers with medication overuse

Three out of five studies suggest that CDH sufferers who are medication overusers have lower SF-36 QoL scores than those who are not medication overusers.

Five studies reported QoL comparisons between CDH subtypes with and without analgesic overuse, using the SF-36 (29,33,43,47) or SF-12 (37). Domain scores were available for only two studies (33,37) and these are presented in Figure 1D. Three out of five studies showed that QoL was more impaired in patients/respondents who were medication overusers compared to those that were not. Differences across the studies in terms of setting and design make between-study comparisons difficult.

Impact of comorbidity on SF-36 score

Four studies (data not shown) demonstrated the statistically and clinically significant negative impact of specific comorbidities [anxiety and depression (45,52), mood disorders (42) and stress (30)) on the QoL of CDH patients compared to CDH patients without the comorbid disorder across most domains of the SF-36.

QoL studies using instruments or methodologies other than SF-36

Included studies showed that CDH was consistently associated with a lower QoL than EH.

Three additional studies reported QoL outcomes using instruments other than the SF-36 (4,25,50). The results of these studies are shown in the online supplementary material Table 1.

Table 1.

Studies of disability and productivity loss using MIDAS

Table 2.

Studies of resource utilization

Table 3.

Studies of economic impact – direct costs

Disability

Loss of productivity due to headache-related disability was reported in 17 studies (4,6,17,20,22–24,27,31,34,37,38,41,44,46,49,53). Ten (8 prospective, 2 retrospective) were of patients enrolled at specialist referral centres or a GP headache clinic (17,20,22–24,31,38,41,46,49). Six studies were cross-sectional surveys of respondents with headache recruited from the general population (4,6,27,34,37,44). One study was a community-based follow-up study of adolescents (53).

With the exception of the study of adolescents, where reported, the average ages of the patient/respondent group with CDH ranged from around 36 to 50 years. With the exception of those studies where patient numbers were small, women dominated men in terms of representation with CDH. Two studies were specifically of analgesic overuse patients (23,24) and the proportion of respondents/patients abusers with analgesic overuse was reported in five studies only (ranging from 17 to 70%) (17,37,46,49,53). The proportion of respondents/patients with named comorbid psychiatric disorders was reported in only three studies (23,37,49). Named comorbidities were specifically excluded from one further study (38).

The most widely used instrument to measure productivity loss was the Migraine Disability Assessment (MIDAS) (13 studies) (4,17,20,22–24,31,37,38,41,44,49,53). This instrument measures disability due to migraine and also provides an indirect estimate of productivity loss (a description of the instrument is provided in the online supplementary material). One further study employed the paediatric MIDAS. Other studies that measured productivity loss did so using a quantitative questionnaire. For brevity, this review mainly focuses on the studies that used the adult MIDAS. Data from 11 of these studies that compared headache subtypes are presented in Table 1. Data from one further study is discussed only within the text (22). The remaining study (53) that included the MIDAS instrument reported no actual data.

Impact of CDH on productivity and MIDAS severity grade

Studies included in this review show that CDH impairs productivity.

All the studies using MIDAS consistently demonstrated that patients/respondents with CDH or subtypes of CDH had impaired productivity, with missed days and/or days of reduced effectiveness across some or all the activities covered in the MIDAS assessment. However, mean number of work days missed, mean total MIDAS scores and spread of patients/respondents across the MIDAS severity grades were quite variable among the CDH (and CDH subtype) patient/respondent samples included in the studies. For example, the proportion of CM patients/respondents with Grade 4 severity ranged from 16.6% to 83%, total MIDAS scores for CM patients/respondents ranged from 34.9 to 84.1 and the number of work days missed in the previous 3 months ranged from 0.23 to 16.4.

MIDAS comparisons of CDH patients/respondents with EH patients/respondents

Studies included in this review consistently demonstrated that CDH is associated with greater disability and productivity loss than is associated with EH.

Seven studies (4,17,20,24,41,44,49) reported comparisons of CDH (or CDH subtype) patients with EH patients using MIDAS. The results consistently demonstrated disability and productivity loss associated with CDH worse than that associated with EH.

Two studies reported comparisons of CDH subtypes that were with and without migraine or migrainous features using MIDAS. Data reported by Lanteri-Minet et al. (4) showed that CDH subjects with migrainous features scored significantly worse than CDH subjects without migrainous features (p ≤ 0.01). In contrast, Gesztelyi et al. (31) showed no significant difference in total MIDAS scores between patients with CM and those with CTTH. However, this study included only a very small number of subjects with CM (n = 8), compared to the Lanteri-Minet et al. study (n = 109) and reporting of data was incomplete.

Impact of medication overuse and comorbidity on MIDAS score

The one available study (37) that compared productivity loss in CDH respondents with and without analgesic overuse showed that medication overuse was associated with greater disability and productivity loss. In addition, one study examined the impact of specific comorbidities on the MIDAS score of patients with CDH. Corchs et al. (22) reported no significant differences between MIDAS scores for phobic and non-phobic CM patients, although numerically scores were higher in non-phobic patients (mean total score 69.03 ± 69.30 vs. 92.50 ± 90.45).

Disability and loss of productivity using questionnaires other than MIDAS

Three additional studies examined productivity losses due to CDH using quantitative questionnaires (the results from these studies are included in the online supplementary material) (6,27,34).

Four studies reported on the disability impact of CDH on patients using instruments other than MIDAS (see online supplementary Table 2).

Healthcare resource utilization

Resource utilization data were available for two studies only: the GRIM 2000 study from France (4,14) and the AMPP study from the USA (44). The GRIM 2000 study evaluated the use of healthcare resources by respondents with CDH over the previous 6 months compared to respondents with migraine and respondents with other EH (Table 2). The average number of consultations about headaches per 6 months was 5.38 per CDH subject, the majority of which were to specialists. This was four times the number of consultations compared to migraine subjects and 15 times compared to other EH subjects. The difference between the headache groups was most marked when considering the number of specialist consultations. CDH subjects had made nearly six times more specialist visits than migraine subjects and 20 times more visits than other EH sufferers. Psychiatrists were the specialists most frequently consulted by CDH subjects (data not shown). However, the medical speciality consulted by the highest proportion of CDH subjects was ophthalmologists. Compared to other headache sufferers, visits to GPs were significantly more frequent for CDH subjects (p < 0.01; Wilcoxon test). The extent and nature of physician consultation was similar in CDH subjects with and without migrainous features (data not shown). Ten subjects from the CDH sample had been hospitalized for their headaches for an average of 12.9 days. In comparison, seven subjects with EM were hospitalized for an average 2.29 days. However, this difference was not statistically significant. There were no headache-related hospitalizations among other EH sufferers.

Economic impact

Direct costs

The two most comprehensively reported economic studies included in this review demonstrated higher direct costs associated with CDH, particularly CM, than with EM.

Direct costs for CDH or CM were reported for four different studies from three different countries. Data reported from the GRIM 2000 study (4,14) conducted in France (Table 3) showed that total direct costs per individual with CDH in France were over 10 times greater than per individual with EM and nearly 50 times greater than for subjects with other EH. Hospitalizations contributed most (>60%) to the total direct costs for subjects with CDH. Compared to an individual with migraine, average hospitalization costs were more than 40 times higher for an individual with CDH. Average out-of-pocket drug consumption costs per individual for headache relief were nearly five times higher for CDH (€171.25 per year) than for migraine (€35.70 per year) and nearly 35 times higher than for other EH. Analgesics and non-steroidal anti-inflammatory drugs were the most widely used drug treatments in each case (data not shown). Expenditure on medication was somewhat higher in CDH subjects with migrainous features than those without, although not significantly so (data not shown). Average cost per subject for consultations were more than four times higher for CDH subjects compared to EM subjects and about 15 times higher compared to subjects with other EH. Similarly, diagnostic work-up was approximately five times more costly per subject with CDH than per subject with EM and more than 10 times more costly than per subject with other EH. It is interesting to note that this study showed that two-thirds of all headache-induced direct costs in France were due to CDH subjects and that the main part of this was attributable to hospitalizations (data not shown).

Information obtained from the Mazzotta et al. study (40) is limited in respect to CM because of the very small number of patients and it was also a study of children and adolescents. Nonetheless, consistent with the French study, this Italian study showed an approximate 10-fold increase in drug costs per CM patient compared to those per patient with EM with aura. The US-based AMPP study reported by Munakata et al. (44) showed that total direct costs for a subject with TM were approximately three times greater per year than for a subject with EM. This higher cost per subject was reflected in approximately proportionate increases across primary care, outpatient care, drug treatment and hospital costs.

Discussion

This systematic review set out to examine the QoL, disability and productivity loss, healthcare resource utilization and financial burden associated with CDH and, in these terms, compare the burden of CDH with that of EH and of CDH with migraine with other subtypes of CDH.

The review identified a relatively small number of studies that met fairly broad inclusion criteria to define the CDH, CM, TM or CDH with MOH population included within the study. The review also highlighted a number of general issues that limit the interpretation, generalizability and comparability of the study data. These issues relate to differences in study design, setting and the respondent/patient populations. In particular, the use of different criteria used in the studies to define different CDH types and the different headache comparisons undertaken made the outcomes difficult to compare across the different studies. Studies of patients recruited through headache centres are likely to represent the more severe end of the headache spectrum compared to those recruited from the general population. The importance of each setting is dependent on the perspective of the audience, i.e. that of a GP, a specialist physician or a wider healthcare perspective. A lack of distinction between data derived from the different settings could lead to confusion and cloud interpretation of data. The quality of the data is also an issue as most of the studies included in the review had a high dependency on patient/responder self-reporting. Few studies characterized the study populations well in terms of presence of psychiatric and other comorbidities, analgesic medication overuse and/or adjusted or controlled for these factors or others such as headache severity, age, gender and duration of illness, which may influence outcomes and may be important in distinguishing between headache subgroups. Furthermore, small sample size also limited the quality of the data in certain cases. Nonetheless, the review served to identify patterns across the headache spectrum examined.

Quality of life

The studies included in this review, which were from both general and treatment centre populations, consistently demonstrated that QoL is significantly impaired in CDH sufferers compared to non-headache sufferers, and this is particularly in relation to role physical, bodily pain, social functioning and role emotional scales of SF-36. Studies of headache centre populations also consistently indicated that CDH sufferers have a worse QoL compared to patients with corresponding EH and, among CDH sufferers, those with migraine (CM, TM, CDH-M) have the lowest QoL of all. This reflects the relative impact of headache chronicity and attack severity, with CM representing the most severe form of CDH.

There are some data to suggest that patients with CDH who are analgesic medication overusers have even greater QoL impairment. However, more quality studies using modern criteria to define headache in patients with MOH are required to confirm the QoL status of these patients. A limited number of studies have shown that comorbid depression, anxiety, stress or suicide intent can negatively impact on the QoL of CDH patients. However, the presence of these comorbidities has not been considered as possible confounder or controlled for in most of the studies included in this review.

Disability and loss of productivity

Assessments of absenteeism and days of reduced workplace effectiveness due to CDH using the MIDAS instrument or MIDAS-like questioning revealed substantial variation across studies and/or between countries. This variation may reflect the population setting studied and/or the retrospective self-reported nature of the data. It is important to note that the MIDAS instrument is limited in that it depends on the respondent’s recall of their disability and absenteeism over a 3-month period. The lowest levels of absenteeism and days with reduced workplace effectiveness were reported in GRIM 2000 (4,14), which was a study in the general population setting. The other studies were of patients recruited from headache centres. It is likely the patients from the headache centres would be more severely afflicted and hence likely to have higher levels of productivity loss. Nonetheless, analogous to the QoL study findings, the studies using MIDAS were consistent with greater impairment with CDH compared with corresponding EH. Quality data concerning the impact of CDH with migraine features on productivity loss are limited to the GRIM 2000 study and are consistent with greater productivity loss with CDH in the presence of migraine. The results from the FRAMIG study also suggested that productivity loss was greater in patients suffering with medication overuse (37).

Healthcare resource utilization and economic burden

The four studies reporting the economic burden of migraine were difficult to compare due to the different settings of each of these studies. Treatment practices, local prices, different patient populations and small patient numbers in some studies meant that any differences in costs observed between the different types of migraine across these studies could not necessarily be attributed to the nature of the migraine itself.

Implications for further research

Taken together, the QoL data and productivity loss data presented in this review underline the disabling nature of CDH, particularly CDH with migraine features and with medication overuse, and demonstrate the extent to which physical health and health emotional problems may interfere with workplace productivity. However, studies to demonstrate how these impairments translate to increased use of healthcare resources and additional financial burden in comparison with other types of headache are limited. Of particular concern is the paucity of information relating specifically to patients with an intractable or refractory form of CM. These are likely to be those patients with the greatest QoL impairment, who present most frequently at hospital emergency departments and GP clinics, who suffer from medication overuse, and for whom there is no further therapy option currently available to successfully provide relief from their condition. These are the patients in whom new therapeutic possibilities may have the greatest potential to be cost-effective.

Contributors

MLM provided any additional unpublished data; SC and MM extracted and analysed the data, and wrote the manuscript. All authors critically revised the manuscript and provided final approval of the version to be published.

Study guarantor MLM accepts full responsibility for the finished article.