Abstract
We studied the duration of migraine attacks among children and adolescents who reported headache attacks lasting <2 h. Among the 720 children who attended our specialist headache clinic, over a period of 6 years, 231 children had migraine with or without aura fulfilling the International Headache Society (IHS) criteria for the diagnosis of migraine. A further 15 children reported headache attacks typical of migraine, but of duration <2 h. They were asked to fill prospective headache diaries in order to determine with accuracy the duration of their headache attacks. Ten (67%) of these children provided prospective fully analysable headache diaries and recorded a total of 120 headache attacks, 66 attacks (55%) lasting for <1 h each, 30 attacks (25%) lasting between 1 and 2 h, and 24 attacks (20%) lasting >2 h. Patient-based analysis of the headache diaries showed that only three children consistently had headache attacks lasting <1 h. Seven children had some of their headache attacks lasting for at least 1 h and four had some of their headache attacks lasting at least 2 h. This study shows that headache attacks in children with migraine can be variable, and brief attacks are rare. A combination of short (<1 h) and long (>2 h) attacks of headache can coexist in the same patient. The HIS diagnostic criteria of 1988 for migraine in children should therefore acknowledge such variation and allow reduction of the duration of migraine attacks to 1 h.
Keywords
Introduction
The causes of short-lasting headache attacks include episodic tension-type headache, chronic tension-type headache, cluster headache, idiopathic stabbing headache and possibly migraine. Headache attacks lasting for < 2 h have caused diagnostic difficulties for clinicians who treat children with migraine and also to the researchers, who like to see uniformity in approach and consistency in application of clinical criteria for the diagnosis of migraine in children and adolescents.
Since the publication of the International Headache Society's (IHS) classification and diagnostic criteria of headache in 1988 (1), there has been much debate on the appropriateness of restricting the definition of migraine attacks to only those lasting for at least 2 h. In at least one population-based study around 10% of children with migraine had headache attacks lasting between 1 and 2 h (2). Many authors proposed reducing the lower limit to 1 h and others proposed even further reduction to 30 min (3–5).
Studies of the causes and the clinical classification of the types of headache among children with short-lasting attacks are very limited. The aims of this study were to assess the occurrence of migraine among children and adolescents with short headache attacks attending our specialist clinic and to assess the applicability of the IHS criteria.
Patients and methods
This is a prospective observational study of all children who attended a specialist headache clinic over a 6-year period between October 1996 and October 2002. Demographic and clinical features of all children were recorded at time of first clinic attendance. Data included age, sex, age at onset of headache, duration and frequency of headache attacks, and the quality, site of maximal intensity and severity of pain. Detailed symptom analysis during attacks including the presence or absence of anorexia, nausea, vomiting, light intolerance, noise intolerance and aggravation by walking or physical activity was also recorded. The results of physical examination, investigations, treatment and follow-up data were also recorded. All children were given headache diaries to fill details of future headache attacks.
The diagnoses of the types of headache were made on the basis of the IHS criteria where possible (1). The authors reviewed the clinical data and confirmed or revised the diagnoses of children who were assessed by other members of the team.
Headache attacks were classified according to the duration of the reported head pain and other symptoms (until full recovery) into six groups (<5 min, 5–9 min, 10–19 min, 20–29 min, 30–59 min, 60–119 min and >120 min). Note was taken on relieving factors, if any, such as the use of painkillers, sleep and rest.
The data were stored and analysed using the statistical program SPSS for Windows v 9.0 (6).
Results
Clinic population
Seven hundred and twenty children and adolescents attended the clinic. Demographic data are presented in Table 1. Two hundred and thirty-one (32%) children had migraine with or without aura fulfilling the IHS criteria.
Demographic data of clinic patients and those reporting headache attacks lasting < 2 h
Children with short-lasting headache
Of the 720 who attended the headache clinic over a period of 6 years, 100 (13.9%) reported headache episodes lasting for < 2 h. Demographic data are presented in Table 1. Tension-type headache was the most common cause of headache, followed by migraine-like headache and idiopathic stabbing headache (Fig. 1).
Causes of short headache attacks in 100 children. ETTH, Episodic tension-type headache; CTTH, chronic tension-type headache; ETTH-like, episodic tension-type-like headache; Mig-like, migraine without aura-like headache; ESH, episodic stabbing headache; FHM, familial hemiplegic migraine.
Children with migraine and migraine-like headaches
Two hundred and forty-six children had migraine including 231 children fulfilling the diagnostic criteria of the IHS and 15 children with migraine-like headache lasting < 2 h, but with clinical features that would have otherwise fulfilled the IHS criteria for the diagnosis of migraine without aura (Fig. 2).
Schematic description of patients under study.
These 15 children with migraine-like headache represent 6.1% of the 246 children with possible migraine and were studied in detail (Table 2). Ten (67%) children provided prospective fully analysable headache diaries. Five children failed to provide clinical headache diaries as two were lost to follow-up, two provided incomplete diaries that made the diagnostic assessment impossible, and one patient had no headache attacks between the two clinic visits.
Full analysis of children with migraine or migraine-like headaches lasting <2 h
M, Male; F, female; +, present; –, absent; ±, sometimes.
Analysis of headache diaries
A total of 120 headache attacks were recorded (Table 3).
Diary analysis of 10 children with migraine attacks reported to last < 2 h
Sixty-six (55%) attacks were shown to last <1 h each. Children treated 41 attacks successfully with simple analgesics (paracetamol or ibuprofen). Other attacks (25) had spontaneous remission without treatment. None of the children reported going to sleep before resolution of symptoms.
Thirty (25%) attacks lasted between 1 and 2 h. Patients treated 13 attacks successfully with simple analgesics, but the other 17 attacks resolved spontaneously without treatment. None of the patients reported going to sleep before the headache had resolved.
Twenty-four (20%) attacks lasted >2 h. Children treated 14 attacks successfully with simple analgesics, four attacks with migraleve and five attacks with nasal sumatriptan. Only one attack was not treated by medications. Children reported going to sleep before six attacks had resolved, but complete resolution was reported on waking up.
Patient-based analysis
Three children reported headache attacks consistently lasting <1 h (12.5% of all documented attacks in the diaries). Three children reported 39 attacks of which 19 (49%) lasted for between 1 and 2 h. Four children reported 66 headache attacks of which 24 attacks lasted for at least 2 h.
Discussion
The definition of the duration of migraine attacks is essential for the purpose of this study. Migraine attacks can be defined either as the total duration of symptoms from the time of onset to the time of complete resolution, or the duration can be restricted to the duration of headache only. Protocols for treatment trials in migraine put higher emphasis on pain relief and less on complete resolution of symptoms, and hence, took the view that the duration of migraine attacks were defined by the duration of headache.
For the purpose of this study we chose full recovery and complete resolution of symptoms as the end point for the duration of headache attacks. Our choice was based on the IHS's explicit definition of migraine without aura (1) as ‘idiopathic, recurring headache disorder manifesting in attacks lasting 4–72 h’. Use of the phrase ‘headache disorder’ was probably intended to emphasize the multiplicity of symptoms during the migraine attack. Furthermore, the IHS defined the attack as fulfilling the criteria B–D where ‘B’ refers to the duration of attack, ‘C’ refers to the existence of at least two other features of unilateral location, pulsating quality, moderate or severe intensity of pain and aggravation by routine physical activity and ‘D’ refers to the association of either nausea and/or vomiting or photo- and phonophobia. The footnote at the end of the same page states: ‘in children below age 15, attacks may last 2–48 h. If the patient falls asleep and wakes up without migraine, duration of attack is until time of awakening’. The IHS Committee on Clinical Trials in Migraine specifically defined the duration of migraine as the period between ‘the time of the start of attack to the time of end of attack’ (7). Other authors have also taken the same view and regarded ‘duration’ to mean the duration of the migraine attack rather than the duration of the headache as a single symptom (8). From the aforementioned definitions we took the pragmatic approach and accepted that the attack of ‘migraine’ or ‘headache disorder’ includes more than the pain, but also all other symptoms of migraine.
In our group of patients, the attacks of migraine were short and well defined from the onset to the end, and therefore, the recorded duration periods of attacks are very likely to be accurate. It is also very likely that short migraine attacks are different from longer attacks. Short migraine attacks may not be associated with post-dromal symptoms and patients return to normal soon after the cessation of pain.
Symptoms during attacks were largely consistent and reproducible. Issues relating to the fulfilment of other diagnostic criteria arose in younger children who were at times unable to describe the quality of their head pain or other associated symptoms. At first glance, patient number 7 may not seem to have enough symptoms to fulfil the criteria for the diagnosis of migraine. However, he had more than five attacks, his pain was severe enough to prohibit physical activities (although he was unable to describe aggravation of pain on walking) and, understandably, he was unable to describe the quality of pain. He consistently reported photophobia and phonophobia. Therefore, we felt that the diagnosis of migraine is acceptable. Patient number 8 presented at 18 months of age with episodes of benign paroxysmal torticollis of infancy. Her family also reported episodes of headache that were very brief in duration, with intense symptoms lasting only a few minutes.
The data collection and diary recording have confirmed the occurrence, though rare, of short attacks of migraine without aura in children and adolescents. Longer attacks were invariably treated with painkillers and the roles of treatment and sleep were recognized as factors to estimate the attack duration.
The role of acute treatment in the management of short migraine attacks was not clear, as treated and untreated migraine attacks followed a similar pattern. This similarity may offer some explanation for the large placebo effect noted in many drug trials in the treatment of childhood migraine.
We studied children reporting headache attacks of <2 h duration due to the uncertainty that this may cause to the clinician and researcher. Diary recording has shown that at least one in five attacks in this group of children lasted for >2 h, confirming that estimation of duration of attacks by children and their parents can be inaccurate. Because children with reported headache attacks lasting >2 h are less problematic from the diagnosis and classification point of view, we did not review and report on the diaries of those children. It is conceivable that some of their headache attacks may be <2 h and the true occurrence of short migraine attacks may be underestimated.
Full analysis of other causes of short headache attacks in children and adolescents will be reported separately. Preliminary data from this study show that tension-type headache (episodic or chronic tension headache) is the most common cause, followed by idiopathic stabbing headache.
Our data suggest that migraine is a rare cause of short headache attacks. Based on the analysis of our data, the application of the IHS criteria (1988) to the diagnosis of migraine without aura would have led to misclassification of only 15 children out of 720 patients. The application of modified criteria to the diagnosis of migraine without aura by reducing the duration of headache attacks to 1 h would have misclassified only three (2.1%) patients.
From the clinical practice point of view, the misclassification of the three patients may not disadvantage the children concerned, as no diagnosis of underlying sinister cause was or should be missed and the management should not be much influenced by the lack of a diagnostic label. From the clinical research point of view, recognition of the existence of short migraine attacks is necessary to understand the spectrum of the migraine syndrome in childhood and its pathophysiological processes.
Diagnostic clinical criteria that have 100% sensitivity and specificity are practically unachievable. In order to achieve uniformity of approach, it would be reasonable for acceptable criteria to include at least 95% (2 SD from the mean) of the total target population. Thus, it would be reasonable to expect 2.5% of patients to lie outside either side of the limits.
If the IHS criteria (1988) were strictly applied in this study, 15 (6.1%) out of the possible total of 246 patients with migraine would have been missed. Therefore, the 2-h minimum limit may be very restrictive and may lead to a higher level of attrition than the acceptable 2.5%. However, the application of the modified criteria with 1 h as the lower limit of the duration of migraine attack would have allowed only three children to be misdiagnosed out of the possible total of 246 (1.2%). This level of attrition is lower than the acceptable 2.5%. It can therefore be suggested that a change in the diagnostic criteria in order to include patients with headache attacks lasting for at least 1 h would be justifiable.
The diagnosis of migraine in children and adolescents reporting short headache attacks should be based on a long period of observation and accurate recording of headache diaries. The criteria for the diagnosis of migraine in children should take into account the variability of headache attack duration and may state that children should have at least five attacks lasting at least 1 h, rather than all attacks should last for at least 1 h.