Abstract
Headache classification is a dynamic process through clinical testing and re-testing of current and proposed criteria. After publication of the second edition of the International Classification of Headache Disorders (ICHD-II), need arose for revisions in the classification of medication overuse headache and chronic migraine. These changes made apparent a further need for broader revisions to the standard formulation of diagnostic criteria for the secondary headaches. Currently, the fourth criterion makes impossible the definitive diagnosis of a secondary headache until the underlying cause has resolved or been cured or greatly ameliorated by therapy, at which time the headache may no longer be present. Given that the main purpose of diagnostic criteria is to enable a diagnosis at the onset of a disease in order to guide treatment, this is unhelpful in clinical practice. In the present paper we propose maintaining a standard approach to the secondary headaches using a set of four criteria A, B, C and D, but we construct these so that the requirement for resolution or successful treatment is removed. The proposal for general diagnostic criteria for the secondary headaches will be entered into the internet-based version of the appendix of ICHD-II. During 2009 the Classification Committee will apply the general criteria to all the specific types of secondary headaches. These, and other changes, will be included in a revision of the entire classification entitled ICHD-IIR, expected to be published in 2010. ICHD-IIR will be printed and posted on the website and will be the official classification of the International Headache Society. Unfortunately, it will be necessary to translate ICHD-IIR into the many languages of the world, but the good news is that no major changes to the headache classification are then foreseen for the next 10 years. Until the printing of ICHD-IIR, the printed ICHD-II criteria remain in place for all other purposes. We issue a plea to the headache community to use and study these proposed general criteria for the secondary headaches in order to provide more evidence for their utility—before their incorporation in the main body of the classification.
Introduction
The publication of the first edition of the International Classification of Headache Disorders (ICHD-I) in 1988 (1) was a revolution in the classification and diagnosis of headache disorders. The changes it introduced were so fundamental that over a decade passed before the headache community—largely unfamiliar with the science of disease classification as nothing of its sort had preceded ICHD-I—became used to, and then critical of, this new diagnostic system. Most calls for revision initially came from Headache Clinics with epidemiological studies and papers scrutinizing ICHD-I eventually revealing its shortcomings. These publications not only showed the need for a second and fully re-worked edition of the International Classification of Headache Disorders (ICHD-II) but also formed the basis on which ICHD-II was built, and then published in 2004 (2).
ICHD-II was seized upon by a headache community fully prepared for it and many studies have critically examined various aspects of it. Mistakes in ICHD-II were inevitable, despite the intense efforts of the Headache Classification Committee, and these mistakes have been quickly identified in scientific reviews and corrected in published revisions of the diagnostic criteria for medication overuse headache (MOH) (3), chronic migraine (4) and, for a second time, MOH (4). All these revisions have been incorporated in the internet-based version of ICDH-II, but the classification as a whole has not been reprinted. The originally printed version (2) continues to be the formally approved and authoritative version, not least because it has been the basis for translations into more than 20 languages.
Both ICHD-I and ICHD-II distinguish between primary and secondary headaches, the former being disorders in their own right, such as cluster headache, the latter being headaches that are caused by something else, i.e. an external factor such as cold temperature or another medical condition such as a neoplasm. Both editions of the classification define a general template criteria set for secondary headache, which is then adapted as necessary to define each of the different types of secondary headache (see Box 1 for current general criteria for secondary headache in ICHD-II). Few published studies have criticized the general criteria for secondary headaches in ICHD-II, although when they did the requirement for complete resolution after treatment was identified as a specific weakness (5), and the Committee recognized this issue most clearly in the MOH criteria.
Box 1 General diagnostic criteria for secondary headaches in ICHD-II
Headache with one (or more) of the following (listed) characteristics1,2 and fulfilling criteria C and D
Another disorder known to be able to cause headache has been demonstrated
Headache occurs in close temporal relation to the disorder and/or there is other evidence of a causal relationship
Headache is greatly reduced or resolves within 3 months (this may be shorter for some disorders) after successful treatment or spontaneous remissions of the causative disorder3
Notes:
For most secondary headaches the characteristics of the headache itself are poorly described in the scientific literature. Even for those where it is well described, there are usually few diagnostically important features. Therefore, diagnostic criterion A in the standard set of criteria is usually not very contributory to establishing causation. However, criteria B, C and D usually effectively establish causation. This makes it possible to use criterion A not only as a defining feature but also to tell as much about the headache as possible or to show how little we know of it. This is why the formulation of criterion A now allows mention of a number of features. Hopefully, this will stimulate more research into the characteristics of secondary headache so that, eventually, criterion A for most of these headaches can become much more clearly defined.
If nothing is known about the headache, it is stated ‘no typical characteristics known’.
Criterion D cannot always be ascertained and some presumed causative disorders cannot be treated or do not remit. In such cases criterion D may be replaced by: ‘Other causes ruled out by appropriate investigations’.
In many cases sufficient follow-up is not available or a diagnosis has to be made before the expected time needed for remission. In most such cases the headache should be coded as Headache probably attributed to [the disorder]: a definite relationship can be established with full confidence only once criterion D is fulfilled. This is especially so in situations where a pre-existing primary headache has been made worse by another disorder. For example, the great majority of patients otherwise fulfilling the criteria for 1.5.1 Chronic migraine are overusing medication and will improve after this overuse ceases. The default rule in this case, pending withdrawal of the overused medication, is to code according to the antecedent migraine subtype (usually 1.1 Migraine without aura) plus 1.6.5 Probable chronic migraine plus 8.2.7 Probable medication-overuse headache. Following withdrawal, criterion D for 8.2 Medication-overuse headache is not fulfilled if a patient does not improve within 2 months, and this diagnosis must then be discarded in favour of 1.5.1 Chronic migraine. A similar rule applies to patients overusing medication but otherwise fulfilling the criteria for 2.3 Chronic tension-type headache.
In most cases criterion D has a time limit for improvement of the headache after cure or spontaneous remission or removal of the presumed cause. Usually this is 3 months, but is shorter for some secondary headaches. If headaches persists after 3 months (or a shorter limit) it should be questioned whether it was actually secondary to the presumed cause. Secondary headaches persisting after 3 months have often been observed, but most have not been of scientifically proven aetiology. Such cases have been included in the appendix as Chronic headache attributed to [a specified disorder].
Why new general diagnostic criteria for secondary headaches?
The idea of revising the general diagnostic criteria for secondary headache arose as a consequence of errors that were found in the definition of MOH. According to the original published version of ICHD-II (2), MOH could not be definitively diagnosed until the putative cause—medication overuse—had been discontinued. Criterion D of MOH required that this be followed by clinical improvement within 2 months. Although from a purist point of view this might have been the most valid way of defining a headache arising from medication overuse, it overlooked the principal purpose of diagnosis, which is to guide management, and if assumed that the offending condition had no medium or long-term consequences. Doctors need to diagnose patients at the time they present, not retrospectively after their treatment is successful, at which point the correct diagnoses may be merely of academic interest.
This simple truth, which quickly became evident in the context of MOH, and in that context was remedied (3, 4), applies no less to the entire range of secondary headaches. Criterion D in the definition of secondary headache states that the ‘headache is greatly reduced or resolves within 3 months (this may be shorter for some disorders) after successful treatment or spontaneous remissions of the causative disorder’. Thus, criterion D for each type of secondary headache expresses that headache must be fully relieved or greatly ameliorated following removal (through spontaneous remission or cure) of the underlying causative disorder, although in the specific text we failed to use this definition consistently. A secondary headache can be diagnosed with certainty, according to these criteria, only after it has abated, which may serve the purpose of science but not of clinical medicine. In addition, some conditions are untreatable or may leave headache as a permanent sequel. This is the main issue we want to address here.
There are further problems with the way the secondary headache criteria are constructed. For example, in order to make a diagnosis of chronic post-meningitis headache, first there must be an acute headache attributed to meningitis. According to the present system, acute headache attributed to meningitis cannot be diagnosed definitively until after its disappearance, but, if it disappears, it cannot become chronic! A similar consideration pertains to post-traumatic headache: acute post-traumatic headache after 3 months automatically becomes chronic post-traumatic headache but, strictly following the standard criteria, acute post-traumatic headache could not then have been diagnosed. In these two instances, the potential paradox was recognized by the Classification Committee and criterion D was modified for these diagnoses.
Changes to the general criteria proposed below make it possible and easy to add a ‘chronic post-X headache’ to all secondary headaches when appropriate because it is no longer required that headache should improve or disappear before the diagnosis of a secondary headache. Whereas, in ICHD-II, special provision was made only for meningitis and trauma, probably a similar need exists, but is not yet scientifically characterized or clinically recognized, for many of the other secondary headaches to have a ‘chronic post-X headache’. As a general rule, we now propose that headache persisting for ≥ 3 months after cure or spontaneous remission of the causative disorder should be called ‘chronic post-X headache’. However, since little evidence exists for most ‘chronic post-X headaches’, this should be mentioned in the appendix and introduced in the body of the classification only after evidence emerges for the existence of such ‘post-X headache’ for each individual cause of secondary headache. This means that we shall not distinguish between acute and chronic post-X headache but simply call secondary headaches: ‘headache attributed to X’ until 3 months after cure of the causative disorder. If headache then continues, the term ‘chronic post-X headache’ shall be applied. These changes, like the others mentioned in this paper, will be implemented and specified for each of the secondary headaches by the Classification Committee during 2009. We invite headache research groups to collect data prospectively on the secondary headaches at onset and throughout their course and provide succinct clinical descriptions. This may enable a more scientifically based classification of the secondary headaches in a future ICHD-III in 10 years' time.
The proposed changes to the general criteria
The proposed new general criteria for the secondary headaches are given in Box 2. They have been evaluated by all members of the International Headache Society Classification Committee, have undergone amendment and, finally, been accepted by the committee. The rationale for the changes is explained below:
Criterion A now requests only that headache of any type be present. In those few instances where the characteristics of the headache itself support the causal relation to the underlying disorder, this information is now applied in criterion C, which addresses causality.
Criterion B is also largely unchanged. We have changed the phrasing from ‘a disorder known to cause headache’ to ‘A disorder scientifically proven to be able to cause headache’. We want to focus attention on the need for scientific studies documenting a causal relation between a disorder and headache as also discussed in note 2. Such studies can, for example, use the disappearance of headache after treatment of the causal disorder, or advanced imaging methods that are not usually available in the routine diagnostic situation.
Criterion C has been substantially changed in order to provide more evidence of causation. We now require two out of five sub-criteria. Whereas C previously focused exclusively on the temporal relation between the causative disorder and headache at onset, C2 now adds worsening and C3 improvement in temporal relation to the causative disorder. C4 adds typical features of the headache (if any) and C5, other evidence of causation, is now an independent sub-criterion for which an example is given in the note. Many more may emerge during future work, when the general criteria presented here are to be applied to all of the many specific secondary headaches.
Criterion D, which required improvement or disappearance of headache, now requests that headache should not be better explained by another headache diagnosis. It is, for example, possible that migraine worsens along with the progression of a brain tumour, but this may be due to the mental stress of having a tumour. In such a case the headache may be better diagnosed as migraine than as headache attributed to brain tumour. Several similar examples can be given.
Box 2 Proposed general diagnostic criteria for secondary headaches
Headache of any type, fulfilling criteria C and D
Another disorder scientifically documented to be able to cause headache has been diagnosed1
Evidence of causation shown by at least two of the following:2
Headache has occurred in temporal relation to the onset of the presumed causative disorder
Headache has occurred or has significantly worsened in temporal relation to worsening of the presumed causative disorder
Headache has improved in temporal relation to improvement of the presumed causative disorder
Headache has characteristics typical of the causative disorder3
Other evidence exists of causation4
The headache is not better accounted for by another headache diagnosis
Notes:
Since headache is extremely prevalent, it can occur simultaneously with another disorder by chance and without a causal relation. Therefore, a secondary headache can be definitely diagnosed only when solid evidence exists from published scientific studies that the other diagnosed disorder is capable of causing headache. Scientific evidence can come from large clinical studies observing close temporal relationships between the disorders and headache outcomes after treatment of these disorders, or from smaller studies using advanced scanning methods, blood tests or other paraclinical tests even if these are not readily available to the diagnosing physician who will use these criteria. In other words, study methods that are not useful in routine use of the diagnostic criteria may nonetheless be useful for establishing general causal relationships as the basis of criterion B. Throughout this classification, diagnostic criteria must restrict themselves to information reasonably available to the diagnosing physician in a typical clinical situation.
The general criteria require two separate evidential features to be present in all cases, and allow up to five types of evidence as set out. Not all of these five types are appropriate for all disorders, and not all four need form part of the specific criteria for a particular secondary headache when this is so. Worsening or improvement of the causative disorder can be clinical or determined by imaging or other laboratory tests.
Examples are orthostatic headache after lumbar puncture and very sudden onset headache in subarachnoid hemorrhage. The characteristics (if any) must be specified for each secondary headache.
This is normally to be specified for each of the secondary headaches. An example of this kind of evidence is accordance between the site of the headache and the location of a presumed causative disorder.
The future of headache classification
As a standard procedure, new proposals are placed in the appendix of ICHD-II for testing, and this is happening with the new general criteria here proposed for the secondary headaches. Normally, ≥ 2 years should go by so that sufficient testing precedes incorporation into the main body of the classification.
However, a situation has developed in which a number of amendments to ICHD-II have been made in separate publications (3, 4), and this is not acceptable in the long term. At present, only the internet-based version is up-to-date, while thousands of the originally printed versions and probably tens of thousands of translated versions are circulating, all of them not quite up-to-date—and, in all likelihood, not identical regarding MOH and chronic migraine. Now there are new proposals affecting all secondary headaches. Meanwhile, the revised criteria for chronic migraine and MOH have already been tested and found to be valid, so they ought to be moved soon into the main body of the classification.
Therefore, we have decided to begin work on a limited revision of ICHD-II, which will become ICHD-IIR. It will incorporate the new criteria for MOH and chronic migraine as well as these proposals for the secondary headaches (subject to any changes that might result from field testing) and their implementation in each of the chapters on secondary headaches. To show how the new general criteria proposed here can be implemented, we suggest revised criteria in Box 3 for 7.4.2 Headache attributed directly to intracranial neoplasm.
Box 3 Suggested revised criteria for headache attributed directly to intracranial neoplasm
7.4.2 Headache attributed directly to intracranial neoplasm
Diagnostic criteria:
Headache of any type, fulfilling criteria C and D
Intracranial neoplasm shown by imaging
Evidence of causation shown by at least two of the following:
Headache has occurred in temporal relation to the onset of the neoplasm
Headache has worsened in temporal relation to growth of the neoplasm
Headache has improved in temporal relation to improvement of the neoplasm
Headache is progressive, or worse in the morning or after daytime napping, or made worse by coughing or bending over
Headache is localized corresponding to the site of the neoplasm
The headache is not better accounted for by another headache diagnosis
A number of other small inconsistencies in ICHD-II have been revealed in published papers, and corrections to these will also be included. There has been some criticism of the criteria for the primary headaches largely limited to rarer forms where large series had not been available, such as for paroxysmal hemicrania (6) and short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (7). Although in essence headache classification must be conservative, as is the case in any disease area, this need should be balanced against a mechanism for testing and incorporating changes arising from appropriate and published data. So much of our routine clinical work depends on the diagnostic criteria for primary headaches that they cannot—unless it is absolutely necessary—change at intervals shorter than one or two decades.
The plan is, therefore, to publish ICHD-IIR in 2010, as a supplement to Cephalalgia. From then on, there will no longer be differences between the latest printed and internet-based versions. It will, unfortunately, be necessary to translate ICHD-IIR into the many languages of the world, and re-publish these translations. Thereafter it is expected that no further revision of the classification will be necessary for another reasonable period, probably a decade, at which time a totally revised edition, ICHD-III, should be produced.
In the meantime we issue a plea to the headache community to use and study these proposed general criteria for the secondary headaches in order to provide more evidence for their utility—before their incorporation into the main body of the classification in ICHD-IIR.