Article Text


The international classification of headache disorders, 2nd edn (ICDH-II)
  1. J Olesen1,
  2. T J Steiner2
  1. 1Department of Neurology, Glostrup University Hospital, 2600 Glostrup, Denmark
  2. 2Division of Neuroscience, Imperial College London, London, UK
  1. Correspondence to:
 J Olesen
 Department of Neurology, Glostrup University Hospital, 2600 Glostrup, Denmark;

Statistics from

Better classification of headache disorders enables better headache research, understanding of headache, communication, and, ultimately, management of a set of disabling neurological disorders

In the past, many have viewed the primary headache disorders as a continuum of which each is part. Better headache classifications have prompted slowly developing awareness of the existence of many discrete entities amongst these disorders. This in turn has contributed not only to their better recognition in clinic (and therefore management) but also to nosological research that has produced better characterisation of many headache disorders.


The first proposals for the classification of headache disorders were put forward in the 1960s, one from an ad hoc committee of the US National Institutes of Health1 and another, quite similar, from the Research Group on Migraine and Headache of the World Federation of Neurology.2 Both proposals merely listed the few headache disorders that were accepted at that time, and gave short descriptions of them rather than diagnostic criteria. The major advance, and the first internationally acceptable—and clinically useful—classification system, came in 1988 when the Headache Classification Committee of the International Headache Society (IHS) published what is now referred to as “The International Classification of Headache Disorders, 1st edition”.3

Several years of work involving more than 100 international headache experts resulted in a much more comprehensive and hierarchical classification of headache types and subtypes, both primary and secondary, using up to four digits to designate the subordinate levels. Explicit diagnostic criteria for at least most major headache disorders were also set out for the first time. This classification became universally accepted,4–6 and was translated into more than 20 languages. No competing headache classifications have existed since that time.

Although the diagnostic criteria were, in many cases, based on the opinions of experts rather than on published and verifiable data, it is testimony to the quality of those opinions that subsequent research has shown most of the criteria to be both valid and reliable.

No research can be done on a disease that is not defined but, conversely, it is difficult to define a disease on which no research has been done. This situation sounds impossible but may be managed by many and often small successive steps of improvement. Thus, the advent of criteria for many headache disorders in 1988 made it possible to undertake research that would confirm, dispute, or occasionally reject the existence of, and the validity of the criteria coupled with, each defined disorder. It is on the basis of a large number of such studies that the second Classification Subcommittee of the IHS has, over the past four years, developed “The International Classification of Headache Disorders, 2nd edition” (ICHD-II).7


When widely used classification systems and their diagnostic criteria markedly change, the knowledge acquired with the earlier criteria becomes of uncertain validity for, and may not be applicable to, disorders diagnosed with the revised criteria. A consequence is that much research would have to be repeated. Fortunately, the diagnostic criteria for the most important headache disorders have not been disputed over the years since the first edition of the classification and, consequently, the criteria for migraine, tension-type headache, cluster headache, chronic post-traumatic headache, and trigeminal neuralgia have been preserved in ICHD-II. Research into these diseases, where it has used the definitions of 1988, remains valid now and will still be in the future. Published evidence has, however, dictated many other important changes.

As before, the classification separates the primary headache disorders (Chapters 1–4) from the secondary headache disorders (Chapters 5–12) (tables 1 and 2). The former include, most importantly, migraine, tension-type headache, and cluster headache. The latter, of which there is a large number, are attributed to some other causative disorder specified in the diagnostic criteria attached to them. The appendix, a new addition in ICHD-II, includes proposed criteria for headaches encountered clinically that are not yet recognised as entities in order to stimulate research into them.

Table 1

Classification of primary headaches (first level, with selected disorders at second and third levels)

Table 2

Classification of secondary headaches (first level, with selected disorders at second and third levels)

The most important changes to the classification and diagnostic criteria of migraine relate to migraine with aura. The revised criteria do not fundamentally change the way we diagnose aura but are, we believe, more easy to understand and apply. Furthermore, they allow a subdivision of migraine with aura into subtypes, all characterised by the occurrence of typical aura (which itself has detailed diagnostic criteria) but, respectively, followed by migraine headache, non-migraine headache, or no headache at all. Other changes to migraine with aura have considerably sharpened the diagnostic criteria for familial hemiplegic migraine (FHM), the dominantly inherited subform of migraine for which two causative genes have now been found. Sporadic hemiplegic migraine (SHM) is a new entry, and both FHM and SHM are more distinctly separated from basilar-type migraine, a new term for what was previously basilar migraine.

An important new inclusion is chronic migraine (CM), classified under complications of migraine. This diagnosis is given to patients who fulfil the criteria for pain and associated symptoms of migraine without aura for 15 or more days per month over three months or longer, without medication overuse. By far the most common cause of migraine-like headaches occurring on 15 days or more per month is medication overuse, described below, giving rise to medication-overuse headache (MOH). CM, in which by definition this is absent for at least two months, is probably a rare syndrome and no good cases have been published in detail, but all experts have seen indisputable cases. If the criteria for CM are otherwise met but medication overuse is present or suspected, the diagnoses given should be probable MOH and probable CM. In addition, because a principle of the classification is that all headache disorders present in a patient are separately diagnosed and coded, the antecedent migraine subtype (almost invariably migraine without aura) is also diagnosed. If, after medication withdrawal, the patient improves, a definite diagnosis of MOH is made and, if not, the confirmed diagnosis should be CM.

In tension-type headache (TTH) the only significant change has been the subdivision of the episodic subtype into infrequent episodic TTH and frequent episodic TTH. The former is defined as TTH occurring less than once a month and was separated out because such a mild headache disorder can hardly be regarded as a health problem, more as a normal variant in the general population (although still requiring classification). TTH is a medical problem only in its frequent episodic and chronic variants.

The major change to cluster headache and other trigeminal autonomic cephalalgias is the inclusion of short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT). In addition, an episodic subtype of paroxysmal hemicrania has been newly recognised.

Under other primary headaches are a number of newly included headache disorders that are fairly rare but should be known to neurologists: hypnic headache, new daily-persistent headache (NDPH), hemicrania continua, and primary thunderclap headache.

A clear improvement to the second section of the classification is that the criteria for all secondary headaches are now built over the same frame: criterion A specifies the headache characteristics, criterion B requires the presence of the causative disorder (sometimes, when this is important, with another set of diagnostic criteria), criterion C defines the causal relationship (often just a close temporal relation) and criterion D demands that the headache greatly improves or disappears after cure or remission of the causative disorder. When A–C are fulfilled but not D, it is generally recommended to diagnose headache probably attributed to [the presumed causative disorder]. There are exceptions, although only chronic post-traumatic headache, chronic headache attributed to whiplash injury, and chronic post-bacterial meningitis headache are currently recognised entities of this sort. Others are included in the appendix for lack of evidence to support their existence. This scheme allows a much better characterisation of each headache and will hopefully stimulate more nosological research into the secondary headaches.

Causation is the key issue in the classification of the secondary headaches. The terminology used for the secondary headaches has been strengthened in ICHD-II to reflect better evidence of causation. Previously these headaches were described as associated with the causative disorder but this term has been replaced throughout by attributed to. Nevertheless, causation is not always clinically apparent or certain. When, for example, a patient develops a headache for the first time with or shortly after head trauma and this headache persists for months or longer, few would disagree with a diagnosis of chronic post-traumatic headache. Such de novo headache, and only such, occurring in close temporal relation to another disorder that is a recognised cause of headache, was recognised in the 1988 classification. This led to some unacceptable situations. A patient who had infrequent episodic TTH, but after head trauma, experienced frequent and severe headaches that phenomenologically still conformed to the definition of TTH could not in the past receive a diagnosis of post-traumatic headache, only of TTH. This problem has been solved by allowing (according to the circumstances) one or two diagnoses: the primary headache diagnosis—in this case infrequent episodic TTH, with or without a secondary headache diagnosis—in this case chronic post-traumatic headache. The following circumstances support the addition of the secondary headache diagnosis: (i) a close temporal relation exists between the suspected cause and the apparent aggravation of the previously present headache; (ii) the suspected cause is known to be able to cause headache such as is now present; (iii) the aggravation of the previously present headache is very marked; (iv) the headache greatly improves (returning to its previous pattern) or disappears within three months after cure or remission of the suspected causative disorder.

All chapters on the secondary headaches have benefited from much more careful review, and include better descriptions and are better referenced than in the first edition. Two new chapters have been added. One deals with headache attributed to disorder of homoeostasis and includes headaches due to systemic disorders such as high blood pressure, hypoxia, and hypercarbia as well as hormonal, fluid, and other disturbances. The other new chapter recognises headache attributed to psychiatric disorder. In the 1988 edition, psychiatric diseases were recognised only for their ability to be a cause of TTH. Now they are placed in line with other causative diseases. Unfortunately, very few research studies have focused on headache in psychiatric patients, and the chapter contains only two entities that were considered to be proved beyond doubt to cause (rather than be comorbid with) headache. Yet, in the opinion of the experts, psychiatric disorders can quite frequently be the cause of headache and a number of these have been set out, with carefully proposed diagnostic criteria, in the appendix. Hopefully, these will encourage more epidemiological and nosological research in this field.

A clinically very important entity, new in ICHD-II, is medication-overuse headache (MOH). Previously this was inadequately covered under headache associated with chronic use of a substance. It is now well documented that frequent and regular use over time of acute antimigraine medication and/or analgesics by people with migraine or TTH risks aggravation of the primary headache; this is the most common cause of a chronic migraine-like syndrome. The new term avoids the pejorative abuse or misuse since, in their traditional senses, these do not apply to the great majority of patients. Overused medications are often prescribed by a doctor who is sadly unaware of the risks of over-frequent use, and used within limits specified in the prescription. The limits indicated are for triptans, ergotamine, opioids, or combination analgesics, use on 10 or more days per month and, for simple analgesics, use on 15 or more days a month. Note that the quantity of medicine taken per month is no longer regarded as the main criterion of overuse.


IHCD-II is a major advance in the classification and diagnosis of headache disorders. Together with a growing body of evidence showing that different headache disorders often require quite different treatments, it should stimulate general practitioners (who manage most headaches) as well as neurologists and other specialists to make detailed as well as accurate headache diagnoses and tailor treatments accordingly. As in the first edition, the rule is still to diagnose and code all the distinct types of headache that a patient has. This multiple-diagnosis system has proved valuable in pointing out the various problems and thereby leading to appropriately differentiated treatments. We hope that the IHCD-II, like its predecessor, will have substantial impact upon clinical diagnosis of headache disorders and thereby improve patient management.

Better classification of headache disorders enables better headache research, understanding of headache, communication, and, ultimately, management of a set of disabling neurological disorders


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.