Abstract
In the absence of a biological marker and expert consensus on the best approach to classify chronic migraine (CM), recent revised criteria for this disease has been proposed by the Headache Classification Committee of the International Headache Society. This revised criteria for CM is now presented in the Appendix. Herein we field test the revised criteria for CM. We included individuals with transformed migraine with or without medication overuse (TM+ and TM-), according to the criteria proposed by Silberstein and Lipton, since this criterion has been largely used before the Second Edition of the International Classification of the Headache Disorders (ICHD-2). We assessed the proportion of subjects that fulfilled ICHD-2 criteria for CM or probable chronic migraine with probable medication overuse (CM+), as well as the revised ICHD-2 (ICHD-2R) criteria for CM (≥15 days of headache, ≥8 days of migraine or migraine-specific acute medication use—ergotamine or triptans). We also tested the ICHD-2R vs. three proposals. In proposal 1, CM/CM+ would require at least 15 days of migraine or probable migraine per month. Proposal 2 required ≥15 days of headache per month and at least 50% of these days were migraine or probable migraine. Proposal 3 required ≥15 days of headache and at least 8 days of migraine or probable migraine per month. Of the 158 patients with TM-, just 5.6% met ICHD-2 criteria for CM. According to the ICHD-2R, a total of 92.4% met criteria for CM (P < 0.001 vs. ICHD-2). The ICHD-2R criterion performed better than proposal 1 (47.8% of agreement, P < 0.01) and was not statistically different from proposals 2 (87.9%) and 3 (94.9%). Subjects with TM+ should be classified as medication overuse headache (MOH), and not CM+, according to the ICHD-2R. Nonetheless, we assessed the proportion of them who had ≥8 days of migraine per month. Of the 399 individuals with TM+, just 10.2% could be classified as CM+ in the ICHD-2. However, most (349, 86.9%) had ≥8 days of migraine per month and could be classified as MOH and probable CM in the ICHD-2R (P < 0.001 vs. ICHD-2). We conclude that the ICHD-2R addresses most of the criticism towards the ICHD-2 and should be adopted in clinical practice and research. In the population where use of specific acute migraine medications is less common, the agreement between ICHD-2R CM and TM may be less robust.
Introduction
The appropriate diagnostic criteria for chronic daily headaches (CDH) related to migraine remain controversial (1–3). In its first edition, the International Classification of Headache Disorders (ICHD-1) did not provide criteria for chronic migraine (CM) or transformed migraine (TM) (4). To fill this vacuum, Silberstein and Lipton proposed criteria for TM (5). TM was subdivided into TM with and without medication overuse.
The Second Edition of the International Classification of Headache Disorders (ICHD-2) provided criteria for CM (6). These criteria required ≥15 days per month of migraine without aura, for at least three consecutive months, in subjects not overusing acute medication. Patients who would otherwise meet criteria for CM but overused medication were to be classified as probable CM with probable medication overuse (referred to herein as CM+).
Although in the absence of a gold standard or of a biological marker it is not possible to say which criteria better correspond to biology, it has been demonstrated that the criteria for CM and CM+ exclude the majority of patients with TM (2, 7, 8). The small degree of overlap is problematic for several reasons. Diagnostic criteria which exclude most patients who might be treated in clinical practice are unsuitable for clinical trials. Alternatively, the results of a study that use restrictive criteria may well not be applicable to clinical practice. A more parsimonious approach might facilitate diagnosis, enhance the utility of trials and better correspond to biology, although in the absence of biological markers that permit the definition of distinct diseases, this may obviously be not true, and multiple diagnoses may better correspond to biology.
For these reasons, we recently tested three alternative proposals for the classification of CM using a subspecialty care sample of persons with CDH who kept daily diaries for at least 3 months (8). All three proposals were less restrictive than ICHD-2 criteria, including from 48.7% to 94.9% of the patients with TM.
As a consequence of the research showing that ICHD-2 criteria for CM were restrictive, the International Headache Society has recently added revised criteria for CM to its Appendix (9). The Appendix criteria are intended for further testing. Accordingly, herein we expand previous research and conduct a field-test study assessing the ICHD-2 revised (ICHD-2R) criteria for CM.
Methods
We reviewed the clinical records and headache diaries of patients seen from 1990 to 2001 in a headache centre, with a diagnosis of TM according to the criteria proposed by Silberstein and Lipton (5).
As detailed elsewhere, during the entire time period reviewed, the clinic used a uniform clinical intake form and standardized headache calendars (8). At the time of the first visit a spreadsheet was completed by a nurse and subsequently checked by the headache specialist containing information regarding the features of pain. Relevant information was transferred to a standardized form which included operational criteria for the ICHD-2 and the ICHD-2R.
Table 1 summarizes the three sets of criteria. In brief, the ICHD-2 classifies CM as a complication of migraine in those cases where the headaches fulfil criteria for migraine in >15 days a month, and medication overuse is not present. If medication overuse was present, CM+ is the diagnosis. The ICHD-2R classifies CM in individuals with at least five previous migraine attacks, who currently have ≥15 days of headache and ≥8 days of migraine and/or headaches that respond to specific acute migraine medications (compounds containing ergotamine, or triptans). However, if medication overuse is present, the individuals should not be classified as CM+, but as having medication overuse headache (MOH). Herein we compare the proportion of individuals with TM that have an ICHD-2 and ICHD-2R diagnosis of CM. Although individuals overusing medication should be classified in the ICHD-2R as MOH, not CM+, we also assessed the proportion of TM+ who had MOH and ≥8 days of migraine according to the ICHD-2R.
Diagnostic criteria for transformed migraine according to the Silberstein and Lipton criteria, and for chronic migraine, according to the ICHD-2 and ICHD-2R
Since we have recently tested three proposals for the classification of CM, we also compared the ICHD-2R with these proposals (8). Proposal 1 required at least 15 days of migraine or probable migraine days per month. The second proposal suggested that CM/C
Data were summarized using frequency tables and descriptive statistics. The proposals for revision of criteria and the ICHD-2 were compared with the McNemar test for proportions adjusted for repeated measurements.
Results
We assessed 557 individuals with TM. The majority of them were female (70.9%) and ages ranged from 18 to 75 years. The mean age was 39.8 years (SD 12.3). Among participants, 399 (62.5%) were overusing medication, while 158 (37.5%) were not.
Testing the ICHD-2R criteria for chronic migraine vs. the ICHD-2
Table 2 compares the ICHD-2 and ICHD-2R classifications. Of the 158 patients with TM without overuse, just nine (5.6%) met ICHD-2 criteria for CM. The others had 15 days of headache and a link with migraine, but <15 days of migraine. According to the ICHD-2R, a total of 146 (92.4%) met criteria for CM (P < 0.001 vs. ICHD-2).
Classification of individuals with transformed migraine, according the Silberstein–Lipton criteria, using the ICHD-1, ICHD-2, and ICHD-2R
M, Migraine; ETTH, episodic tension-type headache; CTTH, chronic tension-type headache; MD, migrainous disorder; HI, headache induced by medication overuse; PM, probable migraine; MOH, medication overuse headache; CM+, probable chronic migraine with probable medication overuse.
CM+ reflects the patients with MOH with eight or more migraine attacks per month.
We emphasize that individuals with medication overuse should be classified as MOH, and not CM+, according to the ICHD-2R. Nonetheless, we assessed the proportion of them who had ≥8 days of migraine per month. Of the 399 individuals, just 41 (10.2%) could be classified as CM+ in the ICHD-2. Most individuals could be classified in the ICHD-1 and ICHD-2, but required a combination of diagnoses. However, most (349, 86.9%) had ≥8 days of migraine per month and could be classified MOH and probable CM in the ICHD-2R (P < 0.001 vs. ICHD-2).
Testing the ICHD-2R criteria vs. other proposed criteria
In individuals without medication overuse, the ICHD-2R criteria (92.4% of agreement with TM) performed better than proposal 1 (47.8% of agreement, P < 0.01) and were not statistically different from proposals 2 (87.9%) and 3 (94.9%) (Fig. 1).

Proportion of individuals with transformed migraine without medication overuse that are classified as chronic migraine in the ICHD-2 and ICHD-2R.
Similarly, for individuals with medication overuse, agreement was higher for individuals with an ICHD-2R diagnosis of CM+ (86.9%) than what was seen in proposal 1 (37.1%), and was not significantly different from proposal 2 (80.9%) and proposal 3 (91%) (Fig. 2).

Proportion of individuals with transformed migraine with medication overuse that are classified as probable chronic migraine with medication overuse in the ICHD-2 and ICHD-2R.
Discussion
Migraine seems to be a progressive disease in some individuals (10–13). The result of migraine chronification is CM. In the absence of a true ‘gold standard’, it is difficult to develop an optimal definition for CM or to map clinical definitions onto disease biology. Nonetheless, several studies showed that, clinically, transformation is characterized by increased frequency of decreased intensity headaches (14–17). The process of transformation frequently ends as a pattern of daily or nearly daily headaches, where just occasional typical attacks of migraine are superimposed. This clinical hypothesis was recently tested in a study, where we showed that the proportion of TM subjects with ≥15 days of migraine per month decreased with age and with the time since the onset of CDH (18). Further support came from a study of adolescents, where we have shown that early in the process of transformation, migraine is more frequent, and that as CDH evolves, fewer typical attacks of migraine occur (19). Therefore, when the ICHD-2 system required that CM sufferers had ≥15 migraine days, it failed to consider the main clinical feature of the disease, the reduced intensity of the CDH over time.
Herein we have shown that the ICHD-2R significantly improves the criteria for CM. Caution is required when interpreting our data. First, this study was retrospective. However, since, to be included in this study, at least three consecutive monthly headache calendars had to be available for review, we retrospectively assigned diagnoses using prospectively collected data. Second, we have previously shown that the number of migraine days in individuals with TM declines with time since onset of TM and age. Since most individuals in a headache clinic have had CDH for many years or decades, our data may reflect an overrepresentation of TM sufferers with few migraine attacks. Third, and most important, this study was done in a headache clinic, where most patients use a triptan, which may increase the sensitivity of the criteria, since use of migraine medications is one of the criteria. The proposed criteria for CM may not perform as well in the population.
We conclude that the ICHD-2R represents an important improvement in the definition of CM, although it is still not possible to say if it maps onto disease biology. It clearly addresses most of the criticism towards the ICHD-2 and should be adopted in clinical practice and research. Further discussion should consider whether CM should be divided into with or without medication.
