The International Classification of Headache Disorders


  • Jes Olesen MD


A set of related medical disorders that lack a proper classification system and diagnostic criteria is like a society without laws. The result is incoherence at best, chaos at worst. For this reason, the International Classification of Headache Disorders (ICHD) is arguably the single most important breakthrough in headache medicine over the last 50 years. The ICHD identifies and categorizes more than a hundred different kinds of headache in a logical, hierarchal system. Even more important, it has provided explicit diagnostic criteria for all of the headache disorders listed. The ICHD quickly became universally accepted, and criticism of the classification has been minor relative to that directed at other disease classification systems. Over the 20 years following publication of the first edition of the ICHD, headache research has rapidly accelerated despite sparse allocation of resources to that effort. In summary, the ICHD has attained widespread acceptance at the international level and has substantially facilitated both clinical research and clinical care in the field of headache medicine.


Dr. Olesen is the chief of Danish Headache Center at the Department of Neurology, Glostrup University Hospital, Copenhagen, Denmark. He is past President of the Danish Headache Society, the Scandinavian Migraine Society, and the International Headache Society, a cofounder of the European Federation of Neurological Associations, and past President of the European Federation of Neurological Societies.

Someone, perhaps Osler, has said that “there are no diseases, only patients.” If this were taken at face value, it could be taken to imply that we have no capacity for diagnosis or treatment. Proficiency in disease management depends primarily upon our ability to place patients within diagnostic groups of a size sufficient to allow systematic research. What Osler meant, of course, was that even with the best system of diagnostic classification, each individual within a diagnostic group is unique. While the caring physician should always recall that patients are individual human beings – not diagnoses – the individualistic view should not govern medical science and promote nihilism in regard to disease classification.

There are now only few “headache-ologists” who can remember the situation that existed before the advent of the ICHD in 1988.1 It was a time when headache was decidedly the poor stepchild of neurology . . . if that. Apart from the work of a few pioneers like Harold G. Wolff, John Graham, Federigo Sicuteri, Marcia Wilkinson, and – last but not least – James W. Lance, who made enormous contributions despite their small numbers and sparse resources, clinical and scientific progress in headache was limited. Headache was not accepted as a neurobiological problem but instead was considered by most a psychiatric or psychosocial problem. It was widely held before (and even years after the advent of ICHD-I) that “a headache is just a headache,” implying the absence of any major difference between the different types of headache. Common wisdom had it that headaches formed a continuum and that it would be impossible to single out specific diagnostic categories. New research results were often dismissed with the comment “How can you be sure that they really studied migraine?.” Conversely, if one wanted to study a specific biological parameter such as blood serotonin level or cerebral brain blood flow in migraine, it was not certain for which patients the findings were relevant. Thediagnostic criteria provided by an ad hoc committee of the National Institutes of Health (see Figure) were, to say the least, far from explicit; in bold are the most ambiguous words.2


Classification of migraine 1963


“Recurrent attacks of headache, widely varied in intensity, frequency, and duration. The attacks are commonly unilateral in onset; are usually associated with anorexia, and sometimes with nausea and vomiting; and are often familial”

Ad hoc classification Committee of the US National Institute of Health 193

In the early 1970s, I initiated the Copenhagen Acute Headache Clinic, and even as a young clinical scientist, I recognized the insufficiency of the National Institutes of Health (NIH) classification system. I consequently developed my own, less ambiguous diagnostic criteria for migraine, and these criteria were published together with the results of our clinical studies.3 I simultaneously began to study the classification of other neurological and psychiatric diseases and quickly came to the conclusion that headache required a comprehensive, universally accepted classification system and that this only could be achieved by a representative international body. Fortunately, the International Headache Society (IHS) had just been formed, and during the second International Headache Congress in Copenhagen in 1985, I suggested the formation of an IHS committee charged with the creation of an objective and comprehensive classification system for the headache disorders. As is typical in such circumstances, the proposer immediately was appointed to do the work. Based on my synopsis and proposal, I was delegated responsibility for the whole procedure, including the appointment of committee members, raising funds for the work and writing and publishing what would be the final classification system. I not only selected key opinion leaders but also took into account the geographical variable and individual expertise in areas relevant to the effort. Bruce Schoenberg, an outstanding neuroepidemiologist from the NIH who sadly died just before the publication of the final work, proved immensely useful as a consequence of his enormous expertise in disease classification and epidemiology. He also provided a link to the World Health Organization (WHO) and helped ensure that ICHD-1 was adopted by WHO into its 10th edition of the classification of diseases (ICD-10).4 Unfortunately, ICD-10 – with its greatly improved headache classification system – has not yet been adopted in the United States.

Each committee member chaired one or more of the subcommittees for the separate chapters. In these subcommittees, we included as many headache experts as possible, partly in order to produce the best possible result but also to ensure that the final product would prove acceptable and so be implemented quickly and extensively.

All went smoothly, and with the contributions by committee members and 50% of my own working capacity devoted to the project over a period of 2 to 3 years, we were able to publish ICHD-1 in 1988. ICHD-1 was accepted almost immediately throughout the world and was translated into more than 20 different languages. More that 100,000 copies of the English version were distributed.

It has been shown in many studies that using International Classification of Headache Disorders (ICHD) criteria, the reliability of diagnosis by 2 different observers is good and prospective studies have shown that individuals in the general population can be classified appropriately using these same criteria.

It originally was assumed that a revision would become necessary rather quickly, but the experts had done their work so well that more than 10 years elapsed before the need for revision and expansion became obvious. The second International Headache Classification Committee again was appointed by me and operated according to the same principles as the first committee. The main structure of ICHD-I was preserved, but along with a large number of minor improvements, some major changes were made.5 Perhaps the single most important new addition was that medication overuse headache was defined as separate diagnostic entity, thus focusing attention on a huge iatrogenic problem. Another major change was the inclusion of headache attributed to psychiatric disorder; this brought psychiatric causes of headache into focus, but much research is still needed in this area. The severely afflicted chronic headache patients seen so commonly in tertiary headache referral centers and heretofore variably called transformed migraine, chronic daily headache, or chronic mixed headache remained an area of controversy after ICHD-2. The category of chronic migraine was created to accommodate these patients, but the diagnostic criteria proved too restrictive, and we later had to revise the criteria and, specifically, clarify that the diagnosis should be applied only when no medication overuse exists.

Another problem that became apparent after the publication of ICHD-2 concerned the diagnostic criteria for secondary headaches. A need for revision was identified, and this work is currently ongoing. Even so, these relatively small problems fortunately have not hampered the acceptance or dissemination of ICHD-2.

What about the validity, reliability, exhaustiveness, sensitivity, and specificity of ICHD-1 and ICHD-2? Lacking an external “gold standard” for headache diagnosis, validity is difficult to judge. Migraine diagnosed according to ICHD criteria responds in 80% to 90% of treatments to injected triptans, drugs that can claim a highly specific mode of action. This proves that clinical diagnosis according to ICHD has been able to identify a group of patients who share a reasonably uniform response to pharmacologic intervention and presumably then share a common pathophysiological pathway. Because the treatment response rates did not vary significantly from country to country, results from the triptan drug trials also suggested that patients were diagnosed in a similar fashion throughout the world.

It has been shown in many studies that using ICHD criteria, the reliability of diagnosis by 2 different observers is good and prospective studies have shown that individuals in the general population can be classified appropriately using these same criteria. While specificity seems to be good (as reflected in the above-mentioned trials), the sensitivity of the criteria is perhaps less well established. Are we underdiagnosing migraine patients? This is certainly the case in North America in general clinical practice, but is it also the case in the hands of experts using ICHD criteria appropriately? Probably not, but it is an issue to be addressed in future research.

In summary, from a humble beginning, headache has progressed to be one of the best classified and defined areas of neurologic disease. Where ambiguity once reigned, we now are blessed with valid, reliable, and generally accepted diagnostic criteria to provide a foundation for further advances in basic and clinical science and in patient care.