Abstract
Background: Hemicrania continua was originally described as a strictly unilateral, continuous headache with an absolute response to indomethacin. Recognition of an increasing number of patients with the same clinical features except for a lack of response to indomethacin has generated controversy about whether the responsive/non-responsive phenotypes belong to the same disorder.
Discussion: We suggest that the non-responsive phenotype should be differentiated from the original concept of hemicrania continua, because it probably indicates a separate type of headache of undetermined nature, i.e. hemicrania incerta. However, differentiating hemicrania incerta from hemicrania continua does not imply that the two headaches are unrelated. Both hemicranias may outline a continuum, giving rise to a broader diagnostic field.
Conclusion: There seems to be a syndrome of ‘primary continuous unilateral headache’ with at least two distinctive categories: hemicrania continua and hemicrania incerta, which are differentiated by their respective response to indomethacin. This division means plurality but adds precision, and allows a clear-cut diagnosis of some controversial cases.
Keywords
Introduction
Nowadays, hemicrania continua is a controversial condition. When first described in 1984 (1), it presented with two robust characteristics: a strict unilaterality of head pain (a moderate, fluctuating and persistent pain), and an absolute response to indomethacin. These two features were both necessary and sufficient for the declaration of the syndrome. Additional, inconsistent traits were oculofacial autonomic signs ipsilateral to the pain and/or some migraine features such as nausea, vomiting, photophobia and phonophobia, which occurred primarily during pain exacerbations (2–5). Since 1992, some atypical variants have been described, namely bilateral hemicrania continua (6,7), side-shift hemicrania continua (8,9) and, above all, hemicrania continua resistant to indomethacin (10–17). These reports have thrown doubt on the notion that a complete response to indomethacin is an intrinsic attribute of the disorder.
Diagnostic criteria for hemicrania continua (ICHD-II)
ICHD-II: International Classification of Headache Disorders, 2nd edition (18).
As originally defined (1), indomethacin response is a sine qua non attribute of hemicrania continua. Consequently, if we were to consider hemicrania continua either with or without response to indomethacin as two subsets of the same disorder, hemicrania continua would have to be defined in another way. Is this an accurate option? Moreover, is it convenient? Our aim has been to suggest a solution to this dilemma. It should be mentioned that we are dealing with a purely descriptive construct, with unknown aetiopathogenesis and no biological markers. Therefore, the debate must take place within a nominal-descriptive frame.
The nominal description of disorders
A clinical description provides certain knowledge of a disorder through the features that are sufficiently characteristic to distinguish it from other conditions. It is an incomplete definition that allows us to recognise the disorder – but not necessarily to know its essence. In such a description, everything observable is at hand, and there is nothing to explain or infer (24). Therefore, descriptive syndromes consist of those observable features by which they are standardly recognised. Investigators must select a set of the most characteristic features, thus adopting one definition of the syndrome rather than another, but carefully weighing the consequences of their decisions in terms of making them clinically valid and therapeutically useful.
Once a convincing definition for a syndrome has been adopted, the concept is settled. Because descriptions express nominal features, no classificatory mistakes are made (24). Clinically-based taxonomy is not supposed to be true or false in essence, but practical (24–26). What can induce error is to base the concept on a hidden, not yet demonstrated aetiology. There are neither biological markers nor external standards with which to judge whether one descriptive definition is more correct than another. Therefore, to debunk an already accepted nominal concept, either proof of the aetiology should be provided, or a more accurate, practical and convenient definition should be proposed.
For clinical and research purposes, operational definitions must be drawn up from descriptive definitions. Operational definitions (i.e. diagnostic criteria) identify the boundaries between a given condition and similar entities. Researchers must determine which symptoms belong to the condition and, among them, try to select the best combination of symptoms that are necessary to establish the diagnosis. In addition, they must establish which diagnostic criteria are essential and which are not, and which may be regarded as alternatives to one another.
Controversial issues
The ICHD-II diagnostic criteria for hemicrania continua (18) require the presence of a strictly unilateral headache with chronic continuous course, accompanied by at least one oculofacial autonomic feature during exacerbations and completely responsive to indomethacin (Table 1). Although an important nosologic rule is to skip atypical features, the few reported atypical cases with bilateral (6,7) or shifting side localisations (8,9), or a lack of autonomic features (1,2,20,21), could be accepted with reservations provided the remaining features are typical – and jointly sufficient to maintain the concept, i.e. ‘a primary continuous headache which is absolutely responsive to indomethacin’.
The proposals of including several unusual, atypical features and/or excluding some essential features of the clinical picture of hemicrania continua are consistent with the natural propensity of syndromes to expand. Physicians strongly interested in a given disorder accumulate experience, become self-confident, and thus make such a diagnosis more readily over time (25). As they become progressively more skilful in detecting affected patients, they may also tolerate a progressive expansion of the boundaries up to the point that the concept becomes so inclusive that it loses all meaning (25–27). For instance, if the described atypical features of hemicrania continua were to be considered all together, hemicrania continua would become a strictly unilateral, bilateral or side-shift headache, with or without autonomic accompaniments, and either responsive or not to indomethacin. Therefore, any continuous primary headache would qualify for hemicrania continua, making the concept useless.
Hemicrania continua responds to indomethacin ergo exists
We face a dilemma, namely whether the clinical phenotypes ‘hemicrania responsive to indomethacin’ and ‘hemicrania without indomethacin response’ are two variants of the same syndrome or two independent disorders (28). In medicine, different disorders may have similar or even identical clinical pictures, and this situation may also apply to hemicrania continua and the controversial issue of its pharmacological response. Several authors have deliberated on the pathogenic nature of the group of patients who met the clinical phenotype of hemicrania continua but did not respond to indomethacin. Some have postulated that ‘indomethacin-resistant hemicrania continua’ may be a subset of hemicrania continua (e.g. 19,22,29,30), while others (e.g. 20,21,28,31) have considered that the nature of this condition may qualify as a different headache category.
The prodigious effect of indomethacin in hemicrania continua indicates that this drug jugulates the symptoms in an antidote-like fashion, thus pointing to a rather specific interruption of the pathogenic process. This superb drug effect is so distinctive that it has been included in the syndrome definition (1,18). In operative terms, indomethacin responsiveness is a compulsory criterion for the diagnosis of hemicrania continua (1,18, Table 1) that theoretically excludes any headache without such response. Accordingly, the nature of a headache similar to hemicrania continua but not responsive to indomethacin remains unclear, and it may not belong to the hemicrania continua concept at all (20,21,28).
In hemicrania continua, autonomic features are not constant: they may be present only during exacerbations (1,2,18,19). Thus, such a diagnostic criterion may either be lacking when the diagnosis is to be made, or simply the autonomic features may be so meagre that they go unnoticed or under-reported (1,2,20,21). In fact, the first reported patient (1) would not have been recognised if the presence of autonomic phenomena had been considered critical for diagnosis at the time (23). Because a substantial number of patients with hemicrania continua may lack prominent autonomic features (1,2,19-21), the response to indomethacin becomes index suis of the syndrome, thus confirming or denying the concept.
Primary continuous unilateral headache: a conceptual nosologic model
Patients with a headache similar to hemicrania continua but non-responsive to indomethacin certainly exist. In 2001 we proposed (20,21) that the typical clinical picture of hemicrania continua, including a complete response to indomethacin, be termed hemicrania continua vera, thus stressing the original and genuine concept of hemicrania continua. On the other hand, we suggested that analogous but indomethacin-resistant headaches could be termed hemicrania generis incerti, i.e. of undetermined nature. However, for the sake of simplicity, both phenotypes may be called from now onwards hemicrania continua – the one with an absolute response to indomethacin – and hemicrania incerta – the one with no indomethacin responsiveness.
This classification brings up a dichotomy, a simple but practical way of classification based on the presence/absence of a particular property which empirically defines the considered domain. Other dichotomous classifications, such as the categorisation of bacteria into Gram+ and Gram–, have proved to be worthwhile. Otherwise, differentiating hemicrania incerta from hemicrania continua does not imply that the two headaches are unrelated. Both hemicranias may outline a continuum, giving rise to a broader diagnostic field that could be termed ‘primary continuous unilateral headache’. The division hereby proposed means plurality but adds precision. It allows clear-cut diagnosis of most patients presenting with a continuous unilateral headache, thus reducing the number of controversial cases.
Our proposal faces the nosologic challenge raised by those patients presenting with a phenotype similar to hemicrania continua but resistant to indomethacin, preserving the original description and permitting further development. These arguments are consistent with the rationale applied to other similar nosologic problems. For instance, when paroxysmal hemicrania was classified (32,33), it was neither considered as a cluster headache subset nor was cluster headache then revisited as an ‘indomethacin-resistant paroxysmal hemicrania’. Moreover, splitting the syndrome into hemicrania continua and hemicrania incerta is reminiscent of what happened with the cluster headache syndrome, which was divided into cluster headache and paroxysmal hemicrania – two headaches with similar clinical features, but sharply distinguished in terms of their differential response to indomethacin (33). At present, both headaches are classified independently, and at the same level, in Group 3 of the ICHD-II (18).
Nosology
Nowadays, there is a claim to code hemicrania continua in Group 3 of the ICHD, along with cluster headache and other trigeminal autonomic cephalalgias. This is a reasonable proposal, but where would the non-responsive to indomethacin phenotype be classified? Regarding the response to indomethacin, hemicrania continua is as different from hemicrania incerta as cluster headache is from paroxysmal hemicrania (Figure 1). Within the same ICHD Group 3, the nosologic rules have to be consistent, i.e. the rationale used for the classification of cluster headache and paroxysmal hemicrania cannot change, even oppositely, when hemicrania continua and hemicrania incerta are brought on stage (Table 2). Let us suppose that paroxysmal hemicrania had been described prior to cluster headache. Would cluster headache be considered as an indomethacin-resistant paroxysmal hemicrania, or as a putative or possible paroxysmal hemicrania?
Nosologic arrangement of hemicrania continua and hemicrania incerta in parallelism with cluster headache and paroxysmal hemicrania (modified from reference 21; the term hemicrania continua vera has been replaced by hemicrania continua; hemicrania generis incerti is now termed hemicrania incerta). On the left: the two short-lasting, excruciatingly severe headaches associated with prominent autonomic features. On the right: the two relatively long-lasting, moderately severe headaches, with non-constant and modest autonomic features. At the top: the two headaches non-responsive to indomethacin. At the bottom: the two headaches absolutely responsive to indomethacin. Proposed classification of the trigeminal autonomic cephalalgias (Group 3 of the International Classification of Headache Disorders, ICHD) SUNCT: Short-lasting Unilateral Neuralgiform headache attacks with Conjunctival injection and Tearing.
So far, hemicrania continua and paroxysmal hemicrania are the only headaches showing an absolute response to indomethacin. Accordingly, both headaches may share essential pathogenic mechanisms. Recent experimental research has demonstrated that indomethacin reduces nitric oxide (NO) induced dural vasodilation, while other common non-steroidal anti-inflammatory drugs (namely, naproxen and ibuprofen) do not have the same effect (34). This particularity might explain why indomethacin is so effective in both types of headache. The strict unilaterality and qualitative autonomic features also strengthens the bond between them, and so does their clear female preponderance. Actually, the links between hemicrania continua and paroxysmal hemicrania may be stronger than those connecting paroxysmal hemicrania and cluster headache.
A proposal based on the natural determination of biological phenomena
In theory, the dichotomy response/non-response to indomethacin may divide the domain of primary continuous unilateral headaches into two separate categories, i.e. hemicrania continua and hemicrania incerta. However, biological phenomena do not change so abruptly in practice, but show transitions from one variety to the other through a series of intermediate forms. Accordingly, hemicrania continua and hemicrania incerta are subject to gradual transitions that may accommodate some rare variations. For instance, hemicrania continua might present with topographic varieties (6–9), several degrees of autonomic features (1,2,20,21), and variations in the response to indomethacin with regard to velocity of response (35), optimal effective dosage (9,36), drug tolerance (36), and concomitant response to other drugs (3,16,37–47). Within Group 3 of the ICHD, such a gradation may also explain why some cases of cluster headache respond to indomethacin (48), and why some paroxysmal hemicranias respond to sumatriptan (49,50).
A natural distribution of all the patients suffering from hemicrania continua and hemicrania incerta will doubtless display a gradual distinction. Therefore, in comparing both clinical phenotypes, the distribution of patients will include not only the ‘white’ and the ‘black’, but also the ‘grey’ (25). Supposedly, the immense majority of hemicrania continua patients will exhibit continuous, strictly unilateral headache, with absolute response to indomethacin, whereas the vast majority of patients with hemicrania incerta will have continuous, strictly unilateral headache unresponsive to indomethacin. Rarely, both disorders may overlap, giving rise to mixed forms: hemicrania continua with different degrees of suboptimal – not absolute – response to indomethacin, and hemicrania incerta with various degrees of substantial – not yet absolute – response to indomethacin (Figure 2C). In this conceptual model the scarce mixed forms represent the natural boundary between both phenotypes.
Three possible models for the differentiation of hemicrania continua into two clinical phenotypes (with/without response to indomethacin). A. The two headache phenotypes theoretically considered as two subsets of the same disorder. B. The two headache phenotypes theoretically considered as two independent disorders (dichotomous concept). There is no apparent relationship between the two disorders. C. The two headache phenotypes considered as two independent disorders, but with some mixed forms (gradual concept). There is a connection between both disorders.
A natural presentation of events with most cases being typical and only a few being atypical will apply to both disorders, giving rise to a natural distribution with a point of rarity centred on a few mixed cases (Figure 2C). Because the number of patients with fully fledged typical features will, by and large, be more frequent than the few patients exhibiting one atypical feature, the boundaries between hemicrania continua and hemicrania incerta are guaranteed. This proposal will allow us to diagnose almost all cases belonging to either clinical phenotype, and will substantially reduce the number of controversial cases to only a few mixed forms.
In conclusion, because response to indomethacin is a sine qua non diagnostic criterion of hemicrania continua, encountering clinical phenotypes similar to hemicrania continua but non-responsive to indomethacin raises a controversy that needs to be solved. There seems to be a syndrome of primary continuous unilateral headache with at least two distinctive categories: hemicrania continua and hemicrania incerta, which are differentiated by their respective response to indomethacin. Distinguishing both phenotypes adds precision both for research purposes and for the best outcome of the patients. Otherwise, a gradual distinction between both clinical phenotypes is in accordance with the natural distribution of biological phenomena.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.