Risk Factors for Headache Chronification


  • Ann I. Scher PhD,

    1. From the Uniformed Services University – Preventive Medicine Biometrics, Bethesda, MD, USA (Scher, Midgette); Albert Einstein College of Medicine – Neurology, Bronx, NY, USA (Lipton).
    Search for more papers by this author
  • Lynn A. Midgette MPH,

    1. From the Uniformed Services University – Preventive Medicine Biometrics, Bethesda, MD, USA (Scher, Midgette); Albert Einstein College of Medicine – Neurology, Bronx, NY, USA (Lipton).
    Search for more papers by this author
  • Richard B. Lipton MD

    1. From the Uniformed Services University – Preventive Medicine Biometrics, Bethesda, MD, USA (Scher, Midgette); Albert Einstein College of Medicine – Neurology, Bronx, NY, USA (Lipton).
    Search for more papers by this author

  • Conflict of Interest: None

Ann Scher, Preventive Medicine Biometrics, Uniformed Services University, 4301 Jones Bridge Road, Bethesda, MD 20814, USA.


About 4% of the adult population and about 1% to 2% of children experience chronic attacks on a daily or near daily basis. While there is uncertainty about the biological mechanisms that lead to headache “chronification,” the epidemiologic literature has provided some insight into modifiable and nonmodifiable factors that appear to influence risk of headache progression. This review summarizes the evidence from population-based studies related to the chronic daily headache phenotype, natural history, and risk factors that may influence incidence, prevalence, or prognosis.


While most people who suffer from headaches experience attacks once or twice a month, some individuals (about 4% of the adult population and about 1% to 2% of children) chronically experience attacks on a daily or near daily basis.1-3 While there is uncertainty about the biological mechanisms that lead to headache “chronification,” the epidemiologic literature has provided some insight into modifiable and nonmodifiable factors that appear to influence risk of headache progression.* The purpose of this review is to summarize the evidence from population-based studies related to the chronic daily headache (CDH) phenotype, natural history, and risk factors that may influence incidence, prevalence, or prognosis.


We use the definition of CDH as headache of any type occurring on 15 or more days per month, which has been the frequency cut-point used in most studies. As noted later in this review, there is evidence that this headache frequency is too low and probably not optimal for epidemiologic research.

CDH can be secondary (attributable to an underlying disorder) or primary (not attributable to an underlying disorder). Primary CDH is divided into shorter duration headaches (<4 hours, such as chronic cluster headache, chronic paroxysmal hemicrania, and hypnic headache) and longer duration headache types.4 Long duration types include chronic tension-type and chronic migrainous headache and, less commonly, hemicrania continua and new daily persistent headache.4 Most CDH sufferers in the general population have chronic tension-type headache or chronic migrainous headache (below).

The classification of the very frequent headache disorders has been a subject of controversy and has undergone recent revision in the International Classification of Headache Disorders (ICHD-II).5,6 The First Revision (ICHD-IIR1) provides diagnostic criteria for some frequent headache disorders, including chronic tension-type headache and chronic migraine. We note that in this nomenclature, the diagnosis of chronic migraine is related to the presence or absence of “medication overuse headache” (MOH). MOH is not established until medication is withdrawn for 2 months and the headache resolves or reverts to its episodic pattern (see ICHD-IIR1 diagnostic criteria for MOH in Fig.). This latter requirement is impractical in nonclinical populations and the prevalence of MOH – as defined in ICHD-IIR1 – is unknown in the general population.

Figure Figure.—.

Adapted from International Classification of Headache Disorders, Revision IIR1.


Prevalence.— The prevalence of CDH has been fairly consistent in adult populations worldwide. Table 1 summarizes the population data related to the prevalence of CDH, stratified when possible into the overall population prevalence, prevalence in childhood or adolescence, and prevalence in the elderly. Data are extracted from the referenced publications in some instances as noted in the table.

Table 1.—. Prevalence of Chronic Daily Headache
Author (year of publication)CountryAge range (years)Prevalence (%)Case definition
  • Extracted based on published data.

  • Unpublished data from referenced study.

 Alders (1996)48Malaysia5+  2.4>180 days previous year
 Bigal (2006)16USA18-895.02.1 15+/month, 4+ hours/day
 Castillo (1999)7Spain14+, 4+ hours/day
 Colas Chacartegui (2004)49Spain14+  1.515+/month
 Deleu (2002)50Oman10+  5.4180+/year
 Hagen (2000)51Norway20+
 Henry (2002)52France15+  3.0Daily
 Ho (2001)53Singapore12+  3.3>180 days in the preceding year
 Lanteri-Minet (2003)54France15+  3.0Daily
 Lavados (1998)55Chile15+  2.6180+/year
 Lu (2001)8Taiwan15+, 4+hours/day
 Mitsikostas (1996)56Greece15-756.82.14.5Daily
 Queiroz (2006)57Brazil15-6410.12.16.4≥180 days per year
 Rasmussen (1991)58Denmark25-64  3.0≥180 days per year
 Scher (1998)9USA18-655.02.84.115+/month
 Takeshima (2004)59Japan15+ criteria for chronic tension-type headache
 Tekle Haimanot (1995)60Ethiopia20+ criteria for chronic tension-type headache
 Wiehe (2002)61Brazil18+ criteria for chronic tension-type headache
 Wiendels (2006)18Netherlands25-55  3.7>14 days/month for at least 3 months
 Anttila (2002)62Finland12  1.615+/month, 6+months
 Bugdayci (2005)63Turkey8-16  1.5Met criteria for chronic tension-type headache
 Castillo (1999)7Spain14-190.90.0 15+/month, 4+hours/day
 Fendrich (2007)64Germany12-≥15 days/month
 Laurell (2004)65Sweden7-15  2.815+/month
 Wang (2006)66Taiwan12-, 3+ months, 2+hours/attack
 Zwart (2004)67Norway13-18  0.5Daily
 Castillo (1999)7Spain60+11.30.5 15+/month, 4+hours/day
 Hagen (2000)51Norway60+2.82.2 15+/month
 Prencipe (2001)10Italy65+
 Scher (1998)9USA56-653.71.83.115+/month
 Wang (1999)11Taiwan65+, 6+ months

The prevalence rates of CDH in the general population average 3% to 4%. The prevalence in childhood or adolescence is lower, averaging about 1% to 2% (Table 1). Prevalence in the elderly is similar to that seen in the general adult population.

Headache Subtypes.— In the population, most CDH sufferers can be classified as either chronic migrainous headache (pre-ICHD-II criteria) or chronic tension-type headache, with roughly equal prevalence of both.7-11 CDH patients in tertiary referral centers primarily have chronic migrainous headaches, presumably due to higher rates of consultation.9 Regarding headache frequency, in one study, 68% of CDH sufferers had frequent but not daily headaches, 15% had daily headaches, and 17% had continuous headaches.1 The subgroup with continuous headaches were older, had a longer duration of CDH, and were more likely to have unclassified headaches (eg, either tension-type or migraine) compared with the other CDH sufferers.1

Incidence.— In a US population sample, the 1-year incidence of CDH in adult episodic headache sufferers aged 18 to 65 years was approximately 3%.1 We note that this is the incidence rate among episodic headache sufferers overall and that incidence rates may be different in those with episodic tension-type vs episodic migraine headache. Incidence rates were much higher (14%) in a German clinic-based study of migraine sufferers.12 The difference in the incidence rates reported in the 2 studies is likely due to the presumably higher risk population seen in clinic vs population samples.

Remission.— The CDH population, at least in nonclinical samples, is highly fluid (Table 2). In the previously mentioned US study, more than half (57%) of CDH cases at baseline had remitted to less than 180 headache days per year at follow-up.1 Similar remission rates were reported in Taiwan (65% remitted to fewer than 15 headaches per month after 2 years, and 60% after 1 year).8 The remission rate was lower in an elderly Taiwanese population (29% had remitted over a 4-year period).11 Most recently, 60% of adolescent CDH sufferers had remitted over 1 year and 75% over 2 years.13 The rate of remission to a more “normal” headache frequency, defined as either <1 headache per week or <7 headaches per month, was 12% to 16% (Table 2).

Table 2.—. Prognosis of Chronic Daily Headache in Nonclinical (Population-Based) Samples
 Age range (years)(A)
Remission to <1 HA/week (%)
Remission to <7 HA/month (%)
(C) Remission to <15 HA/month (%)
1 year2 years4 years5 months5 months1 year2 years4 years
  1. Proportion of CDH sufferers who remit to episodic headache (as defined in A, B, C) over periods of observation ranging from 5 months to 4 years.

  2. CDH = chronic daily headache.

Lu (2001)815+      65 
Scher (2003)118-6514    57  
Wang (2000)1165+ 1613   4029
Wang (2006)1312-14     6075 
Wiendels (2006)1825-55   1235   

From these data, it is evident that the CDH population is highly fluid when defined by a headache frequency of 15 or more headache days per month; the 1-year incidence rate is half or more than the 1-year prevalence and the 1-year remission rate is probably 50% or greater. This has important implications for treatment trials in this population: uncontrolled trials (such as trials of medication withdrawal) will overestimate the benefit of treatment and, in controlled trials, a high placebo response (or, alternatively, regression to the mean or natural history) may make it difficult to detect a treatment response.


Age and Gender (Table 1).— In contrast to migraine, CDH prevalence appears to be fairly constant throughout adulthood. CDH affects women more than men by a factor of roughly 2 to 1. The prevalence of CDH is higher in women than men even after the average age of menopause. Gender-stratified data on CDH in childhood or adolescence are scarce, but suggest a higher prevalence in adolescent girls than adolescent boys (Table 1).

Socioeconomic Status.— The prevalence of CDH appears to be inversely related to socioeconomic status (SES). Lower SES is not only a risk factor for CDH incidence or prevalence1,8,9,11,14,15 but is also associated with a poorer prognosis.1

Marital Status.— In a US prospective study,1 currently married individuals had a lower risk of CDH at baseline and a better prognosis at follow-up.


Potentially modifiable risk factors for CDH include obesity, snoring and sleep problems, comorbid pain conditions, head or neck injury, major life events, smoking, and possibly caffeine intake.

Obesity.— Obesity, defined as a body mass index ≥30, was predictive of 1-year CDH incidence or prevalence in 2 US studies.1,16 Scher et al found that obesity was a risk factor for the 1-year incidence of CDH (OR = 5.53, [1.4-21.8]).1 Meanwhile, Bigal et al found that in comparison with individuals of normal weight, a significantly higher proportions of obese (OR = 1.5, [1.1-2.1]) and morbidly obese (OR = 1.7, 1.1-2.6) individuals had CDH.16

Snoring.— Chronic daily headache subjects were more likely to be habitual (daily) snorers than control subjects. The odds ratio for CDH (daily snorers vs nonsnorers) was 3.3 (< .005).17 This association was independent of headache type and factors known to be associated with snoring and sleep apnea, such as male gender, obesity, and increased age. In this study and in a more recent population study from the Netherlands,18 sleeping problems in general were also associated with CDH.

Comorbid Pain.— There is evidence that CDH co-occurs with other pain syndromes.19 For example, in a Norwegian study,20 individuals with CDH were more than 4 times more likely to have musculoskeletal symptoms than those without CDH (RR = 4.6 [1.0-5.3]). In a US study,1 CDH sufferers over the age of 40 were more likely to report physician-diagnosed arthritis at baseline, compared with episodic controls (OR = 2.41 [1.8-3.3]). In a study of third and fifth grade Finnish schoolchildren21 pediatric headache sufferers were more likely to have persistent musculoskeletal pain at follow-up compared with children not suffering from headache (RR = 1.28 [95% CI 1.08-1.51]).

Head or Neck Injury.— There are scarce data measuring the extent to which CDH follows head or neck injury in the general population. In a US population sample, 20% of male cases reported a head or neck injury the same year or the year before CDH onset (OR = 3.3 [1.0-19.8]).22 The association between head or neck injury and CDH was marginally significant for women (OR = 2.4 [1.0-10.8]). Interestingly, in this study, the relationship between head or neck injury and CDH did not appear to be strongly related to the proximity of the injury; that is, the risk of CDH was increased in individuals with lifetime injuries to the head or neck, even if the injuries were remote to the onset of CDH. Furthermore, the severity of the injury – assessed by whether the subject reported fainting or loss of consciousness – was not obviously related CDH. These results need to be replicated in other samples with a more detailed assessment of head and neck injury.

Contrary results regarding persistent headache as a sequelae of head or neck injury were found in studies based on occupants of vehicles involved in rear-end collisions,23,24 or in emergency room patients with head injury with loss of consciousness.25,26 These discrepant results could be due in part to differences in study subjects, severity or type of injury, and duration of follow-up or latency between injury and development of headache.

Life Events.— Life changes have been considered a precipitant of chronic headache in clinical or other selected populations27-37 although population-based data are scarce and most studies did not separate life events into pre- vs post-CDH events.38,39 In one study, CDH cases were more likely to report certain major life changes in the years proximate to CDH onset compared with controls (2.7 events cases vs 2.0 events controls, < .005).40 Precipitating life events included moves, death of family/friends, changes in marital status, and ongoing “extremely stressful” life events.40 There was no difference in the number of post-CDH events between cases and controls.

Two additional population-based studies compared the occurrence of life changes in CDH cases and controls. In a large school-based study from Taiwan,38 adolescent students with and without CDH were interviewed about the presence of childhood stressors using the Global Family Environment Scale (GFES).41 The GFES measures negative family environmental factors and adverse family events such as parental absence, parental separation or divorce, or abuse. The time period over which events were assessed is not specified, but presumably includes events occurring both before and after CDH onset. Results showed that both physical abuse and parental divorce were more common in the families of the adolescents with CDH than the control group, with the CDH subjects overall having a 10% higher GFES.

In a study of a large Dutch population,39 the occurrence of life events in the year before the study was compared between a group with current frequent headache (defined as headache at least weekly) and a control group. Life events were measured using Paykel's life events scale.42 Results showed that both major and minor life events were more common in the frequent headache cases than controls for the younger (<50) age group. It is likely (but not specified) that the life events primarily followed the onset of CDH in this study.

Caffeine Intake.— The role of dietary or medicinal caffeine as an aggravating factor in the development of CDH is of particular interest since caffeine is the only substance shown to result in withdrawal headache in a blinded, placebo-controlled trial.41,43 In a case control study, the association between high caffeine use, as defined by being in the top quartile for dietary caffeine use or by the use of a caffeine-containing medication for headache as a preferred treatment, with CDH was investigated.44 Although current caffeine use was not a factor, pre-CDH use of caffeine was found to be a modest risk factor (OR = 1.5 [1.0-2.3]).44 Secondary analyses found that pre-CDH caffeine consumption as a risk factors was most evident for those <40 (OR = 3.4 [1.7-6.6]), women (OR = 1.9 [1.2-3.1]) and those with chronic episodic (vs chronic continuous) headache (OR = 1.8 [1.1-2.7]).44 Consistent negative results (for current caffeine consumption) were found in a large population study by Wiendels et al from the Netherlands.18 Wiendels et al18 also found smoking to be associated with CDH.

Medication Use.— MOH is defined when a headache sufferer experiences headache 15 or more days per month in conjunction with the overuse for 3 months or more of acute medication taken to treat headache pain, including triptans, analgesics, opiods, and combination medications.5,6 In addition, headache must remit or revert to its previous episodic pattern within 2 months after discontinuation of the overused medication(s).6 As a trial of medication withdrawal is required to make this diagnosis, the prevalence of MOH in the general population is not known and will not be known until a population-based sample is treated with medication withdrawal and followed for 2 months. However, the proportion of CDH sufferers in the general population who meet criteria for medication overuse (in terms of quantity of medication), as defined by Silberstein/Lipton or other criteria, is approximately 30% (Table 3). (Proposed Appendix criteria eliminate this latter requirement of medication withdrawal followed by headache improvement – making them more practical.45 However, these criteria would overestimate the prevalence of MOH as they implicitly assume that all CDH with medication overuse is CDH due to medication overuse.)

Table 3.—. Proportion of CDH Sufferers with Medication Overuse* in Nonclinical (Population-Based) Samples
StudyAge range (years)% with medication overuseMedication overuse definition and citation
  • *

    Medication overuse definitions:

  • (A) Silberstein/Lipton criteria:68 At least one of the following for at least 1 month: simple analgesics >5 days/week; combination analgesics >3 days/week; narcotics >2 days/week; ergotamine >2 days/week.

  • (B) Revised ICHD-IIR criteria for medication overuse:5 One of the following for at least 3 months: simple analgesics 15+ days/month or ergotamine; combination analgesics, triptans, opioids, or combination of acute medications 10+ days/months.

  • (C) Analgesics 3+ days/week; triptans 2+ days/week; ergots 1+ day/week; narcotics 10+ days/month; 5+ caffeine units/day.18

  • CDH = chronic daily headache; ICHD-IIR = International Classification of Headache Disorders.

Castillo (1999)714+25%Silberstein/Lipton criteria (A)68
Wang (2000)1165+25%(A)68
Lu (2001)815+34%(A)69
Scher (2001)7018-6523%Daily medication use
Prencipe (2001)1065+38%(A)69
Wang (2006)6612-1420%(B)5
Dyb (2006)7113-1836%Daily HA/almost daily medication use
Wiendels (2006)1825-5562%(C)

The potential aggravating or causal effect of medication overuse on headache frequency is a complex issue; much of the clinical data related to medication withdrawal as a treatment for MOH come from uncontrolled trials – often with multiple interventions.46 Thus, the success of withdrawal may be related to concomitant treatment. The recent study by Zeeburg et al is of particular interest due to its relatively large size, data on noncompleters, and the fact that medication withdrawal was the only intervention used.47 In this study, 337 patients with CDH and high medication use (eg, “probable medication overuse headache”) were treated with medication withdrawal for 2 months. Of these, 216 patients (64%) stayed medication-free during this time. After two months, 45% of the 216 completers improved, 48% had no change, and 7% worsened. As improvement appears to be defined as 1+ fewer headache days per month, it is uncertain what proportion of these patients no longer met criteria for CDH. Further, including the noncompleters in the analysis (intention to treat) results in a considerably lower success rate. Thus, even in this study, it is difficult to determine the degree to which medication withdrawal alone will “cure” CDH – if cure is defined as reducing headache frequency to below 15 headache days per month.


Chronic daily headache is a common neurological disorder. Modifiable risk factors identified in nonclinical populations include female gender, low SES, and previous marriage. Potentially modifiable risk factors include obesity, snoring, comorbid pain conditions, head or neck injury, and life events.

The accepted cut-point (for CDH) of 15 headache days per month may be too low, given the highly fluid nature of the population thus identified. This likely complicates the interpretation and conduct of observational and interventional studies for this disorder. There may be subtypes of CDH (eg, continuous headache) that represent a distinct phenotype; future observational and interventional studies should present results separately for this group if possible.


  • *

    While the term “chronification” does not appear in any dictionary to our knowledge, it is used – primarily in the pain literature – to refer to the process by which episodic pain becomes chronic. A Pubmed search of English language publications using the word “chronification” resulted in 51 citations related to: headache: n = 12; back pain: n = 5; other pain: n = 8; psychiatric disorders: n = 7; hepatitis: n = 4; atopic dermatitis/eczema: n = 3. The term chronification appeared more frequently in German-language publications, with 123 publications related to: pain (general): n = 31; back pain: n = 20; headache: n = 3; miscellaneous specific pain conditions: n = 9; psychiatric disorders (general): n = 15; depression: n = 5; eating disorders: n = 4; other specific psychiatric conditions: n = 6; hepatitis: n = 9.