Abstract
The aim of this study was to determine whether in migraine patients with and without aura early treatment with various triptans leads to differences in pain reduction after 1 h and in modulating cutaneous allodynia. Thirty-six patients with early manifestation of a clinically recognizable allodynia of the face and non-responders to earlier treatment with sumatriptan 100 mg were included. Patients were randomized to six triptan treatment groups. Significant pain reduction was seen only in the group receiving zolmitriptan nasal spray 5 mg with a mean visual analogue scale (VAS) score of 3.8 (S.D. 1.2) at baseline and 2.4 (S.D. 1.3; P = 0.015) at 1 h after using the triptan and was thus a predictor of a VAS score 3 within 1 h. The study results indicate that migraine headache intensity can be reduced within 1 h by using zolmitriptan 5 mg nasal spray in spite of the presence of early cutaneous allodynia.
Introduction
Animal experiments indicate that the onset of a migraine attack is first characterized by peripheral sensitization of C fibres located in the dura and in the vascular walls of the trigeminal nerve (1). In the course of the attack sensitization becomes central, which is clinically manifested by cutaneous allodynia (CA) of the face and head. In open studies in humans, Burstein et al. were able to show that patients' response to triptans depends on the stage of sensitization reached at the time of administration. Triptans are less efficacious when central sensitization and CA have already occurred (2). It has been recommended that allodynic patients take triptans no later than 20 min after the onset of migraine (3). The aim of this study was to determine whether treatment with various triptans differs in pain reduction in migraine patients with and without aura and also in changing CA.
Material and methods
This study was carried out in accordance with the Helsinki Declaration of 1975, as revised in 1983. From the database of our tertiary headache centre, migraine patients aged 18–60 years, who met the criteria of the International Headache Society Classification Committee 2nd edn for migraine with and without aura (4), had 1–12 migraine attacks each month and had ever used sumatriptan 100 mg, were invited to participate in the study in a letter requesting them to make an appointment for re-examination.
Patients who accepted the invitation underwent re-evaluation of their migraine symptoms, attack frequency and their response to sumatriptan treatment. We pre-selected those patients who historically showed insufficient response to sumatriptan 100 mg. Response was quantified by pain intensity [visual analogue scale (VAS) 0–10; 0, no pain, 1–3, mild pain, 4–7, moderate pain, 8–10, severe pain], and insufficient response was defined as migraine headache intensity after 2 h of more than VAS score > 3. Patients with somatosensory aura as a manifestation of migraine were excluded in order not to confuse allodynic symptoms and sensory aura. Also excluded were patients with chronic headache, medication overuse headache, and those using pain medications (including opioids) for reasons other than migraine.
As a first step, all patients were asked for possible signs of CA in a validated questionnaire that has high reliability in detecting allodynic patients (3). The questionnaire was translated into German and retranslated by a native speaker and again validated by five migraine patients with CA not involved in the study. Those study patients with possible CA were instructed to look for allodynic signs at the time of their next migraine attack. CA at any time during their next attack and/or constant allodynia between attacks (headache-free period) were to be recorded. As signs of CA were considered (3): patient feels pain or an unpleasant sensation when combing his hair on the headache side or pulling it back; patient is unable to wear eyeglasses, contact lenses and/or earrings; patient feels pain when touching the skin with a brush (area around the eye, frontal skin and temporal area); unpleasantness when shower water hits patient's face, when shaving his face; patient feels pain when touching the affected skin with ice; patient cannot sleep with head resting on the headache side.
Patients were further excluded when they failed to attend follow-up or violated protocol concerning inclusion criteria (e.g. did not take the particular triptan within 30 min, or took a second dose of triptan, or were on other pain medication). All patients who first treated three consecutive migraine attacks with sumatriptan 100 mg and recorded CA were randomized and allocated to one of six treatment groups, depending on their prior treatment (before randomization) with triptans. Each treatment group contained only patients naive for that particular triptan. The intention was to treat three consecutive, definite migraine attacks within 30 min after onset of the headache phase of the migraine attack.
Treatment group I consisted of patients who were encouraged to treat three consecutive migraine attacks with zolmitriptan 2.5 mg, treatment group II of patients who were encouraged to treat three consecutive migraine attacks with eletriptan 80 mg, treatment group III of patients who were encouraged to treat three consecutive migraine attacks with naratriptan 2.5 mg, treatment group IV of patients who were encouraged to treat three consecutive migraine attacks with frovatriptan 2.5 mg, treatment group V of patients who were encouraged to treat three consecutive migraine attacks with zolmitriptan nasal spray 5 mg, and treatment group VI of patients who were encouraged to treat three consecutive migraine attacks with eletriptan 40 mg. Pain intensity (VAS 0–10; 0, no pain, 1–3, mild pain, 4–7, moderate pain, 8–10, severe pain) was measured before treatment, in the early treatment phase (after 1 h), and 2 h and 4 h after taking triptan. CA was measured with a small brush, cold allodynia with a small ice tube. Patients were instructed to touch the affected skin area every 15 min. Intensity of CA was measured according to the patient's rating of unpleasantness (0–100%) at every such touch.
Overall tolerability assessments included the incidence of adverse events (AEs), abnormal findings on physical examination and vital signs. After each hour the mean of all ratings was entered in the statistics.
Data were analysed statistically with SPSS 13.0 software (SPSS Inc., Chicago, IL, USA) and the MedCalc 8.0.0.0 package (MedCalc Software, Mariakerke, Belgium). Data are presented as mean [standard deviation (
Results
Description of the study sample
In our database we identified 756 patients who had ever used sumatriptan 100 mg since 1993 and asked them to re-evaluate their migraine attacks; 137 patients responded to our request and were further evaluated. Finally, after three consecutive attacks had been treated with sumatriptan 100 mg without success, 36 patients were randomized and allocated to one of the six treatment groups. Figure 1 shows the comprehensive trial profile; Table 1 details demographic information on the study participants for each treatment group.
Flow of participants through the study.
Demographic characteristics of the study population according to treatment group (TG)
Expressed as mean [standard deviation (
Time course of VAS scores
The individual time course of the VAS scores for each patient as a function of the treatment group is shown in Fig. 2. In treatment group I (zolmitriptan 2.5 mg) the mean VAS score was 4.7 (
(a–f) Time course of individual visual analogue scale (VAS) scores in each treatment group.
Prediction of VAS score ≤ 3 within 1 h
In addition to the data detailed in the previous paragraph, we aimed to assess whether in a multivariable analysis a particular treatment group might perform better in achieving a VAS score ≤ 3 within 1 h. For this purpose we performed a logistic regression analysis using a stepwise forward approach, the six treatment groups as independent variables and a VAS score ≤ 3 as a dependent variable. As a result, of all six treatment groups used in the present study, the group receiving zolmitriptan nasal spray 5 mg was the only treatment group entered in the statistical model and thus a predictor of a VAS score ≤ 3 within 1 h (regression coefficient B, 2.30; standard error, 0.99; Wald, 5.36; d.f. = 1; P = 0.021; exp B, 10.0; 95% confidence interval exp B, 1.4, 70.3).
Change in CA (Fig. 3)
Consistently, allodynia to cold stimuli occurred not at onset of the attack, but within 2 h of commencement of the migraine attack, whereas response to brushing started with onset of the headache attack. Response to brushing did not decrease in treatment groups I, II, III or IV during the study period. Treatment group II showed an initial increase followed by a slight decrease after 2 h. Only treatment group V showed a slight decrease in response to brushing. Response to cold stimulus increased in treatment groups I, III and IV, but decreased in treatment groups II and V after 1 h. In none of the six treatment groups did allodynic signs decrease to 0. The degree of response to brush and cold stimuli was comparable in all six treatment groups. In none of the treatment interictal was CA observed.
(a–f) Time course of mean pain scores [, visual analogue score; signs of cutaneous allodynia (CA) (brush 
; cold ▪)] in each treatment group.
Migraine history: duration and number of migraine attacks
Mean headache history in all six treatment groups was 18 years. Compared with the various treatment groups, migraine history was longer in treatment group II than in the others (19 years).
Mean number of headache attacks was higher in patients in treatment group I (six attacks/month) and treatment group II (seven attacks/month).
Side-effects
The incidence of overall AEs was low for all treatment groups. The overall rate of AEs was 18%. When drug-related AEs were examined, the incidence of AEs was seen to decrease further. Only In treatment group II did two patients suffer from slight to moderate paraesthesia and chest pressure.
Discussion
The study results indicate that migraine pain and central sensitization, reflected as allodynia, can be reduced within 1 h if treated early (within 30 min) with zolmitriptan 5 mg nasal spray instead of sumatriptan. There is also a favourable tendency to influence pain and allodynia after 2 and 4 h by using zolmitriptan 5 mg nasal spray and eletriptan 80 mg within 30 min after migraine onset.
Historically, triptans have been used in a treatment paradigm that requires patients to wait until their migraine headache pain is moderate or severe. When data from triptans used in an early treatment paradigm (mild pain) are considered in aggregate, triptan therapy may be optimized when used early in the mild pain phase of a migraine (5–7). Furthermore, it has also been argued that an early intervention treatment strategy may result in the treatment of headaches that would not progress to moderate or severe migraine (8).
In a series of studies, Burstein and colleagues have shown that central sensitization and CA exhibit two distinct phases: a developmental phase, during which the activity of the central neurons that mediate the allodynia depends on a continuous flow of incoming pain signals from the meninges; and an established phase, during which the activity of the central neurons is independent of the incoming pain signals (9). In their open-labelled study, they showed that early but not late administration of sumatriptan can abort migraine pain and reverse CA. Using animal models, they showed that triptans disrupt communication between peripheral and central trigeminovascular neurons by inserting a presynaptic blockade in the dorsal horn. Given that in the present study patients were treated early (i.e. within 30 min of onset of attack), it was surprising that none of the presented triptans aborted the pain and reversed the allodynia. The fact that zolmitriptan 5 mg nasal spray and eletriptan 80 mg influenced the course of the migraine attack and allodynia better than did the others would have been predictable with the Burstein model. The fact that all the other triptans failed to achieve such an outcome is at odds therewith, however.
Nevertheless, specifically, neurobiological evidence indicates that the progression of migraine pain may result from activation and sensitization of peripheral nociceptors followed by central nociceptor sensitization. As a result, Burstein et al. have suggested that early migraine treatment targeted at the sensitization of peripheral nociceptors could prevent the development of central sensitization. Progressive sensitization to the level of the central nociceptors may result in development of the full-blown migraine complex, complete with severe pain and additional associated symptoms such as allodynia (10, 11). This symptom should be identified by means of special questionnaires, because other methods are not amenable to patient self-assessment at the time of an attack. It was postulated that in the presence of allodynic signs triptans are adequately effective only when given early (within 20 min) in the stage of an acute migraine attack (1, 12). If given too late, after the establishment of central sensitization, triptan therapy can still mitigate throbbing, but not migraine pain or CA.
The recently published TAME study (rizatriptan 10 mg vs. placebo) has reported similar results. Rizatriptan 10 mg was able to reduce pain und allodynia within 2 h when given early in an attack, regardless of the presence or absence of symptoms suggesting CA (13). The presence of CA at 2 h after treatment, however, was correlated with lack of response to acute therapy in terms of pain freedom. These results partially correlate with the results of our study concerning the fact that CA cannot be sufficiently reduced within 2 h if it is present at baseline. The difference between the TAME study and our study is that the former treated patients within 1 h, whereas in our study population patients were treated 30 min after the onset of the headache phase of the migraine attack. Another difference is the fact that other headache diagnoses were not excluded. For example, chronic tension-type headache, chronic daily headache and episodic tension-type headaches were all permitted, as long as the patient could identify accurately the migraine headache as such. The authors concluded that they believed it to be more likely that this response reflects the action of the class of triptans than that being a unique response to rizatriptan. However, it cannot be ruled out that the clinical resolution of these symptoms is due to the dose amount of 10 mg rizatriptan, which would support our hypothesis that doubling the triptan dose produces a better effect. It would be interesting to perform the same study with 5 mg of rizatriptan.
Our results also partially confirm the results of Burstein and colleagues. On the one hand, we cannot confirm that even in the case of early treatment (within 30 min) with zolmitriptan 2.5 mg, naratriptan 2.5 mg, frovatriptan 2.5 mg, or eletriptan 40 or 80 mg pain or CA was adequately influenced. However, despite the presence of CA, headache intensity was significantly reduced within 1 h by administering zolmitriptan 5 mg nasal spray. On the other hand, we were able to show that in sumatriptan non-responders and in the presence of early CA, early treatment with zolmitriptan 5 mg nasal spray and eletriptan 80 mg tended to reduce migraine pain and ongoing CA after 2 and 4 h. Our explanation is that in migraine attacks associated with CA, the central neurons eventually become sensitized and develop their own autonomous activity. Hypothetically, at this stage, the same presynaptic inhibition produced by triptans becomes more effective by doubling the dosage, to render central trigeminovascular neurons quiescent, thus providing consistent pain relief. We speculate that at higher concentrations zolmitriptan 5 mg nasal spray and eletriptan 80 mg may be available to the presynaptic sites in the dorsal horn and that the selection of sumatriptan-resistant patients changes the synaptic environment along the trigeminovascular pathway. However, the favourable characteristics of nasal spray may also be a pharmacokinetic phenomenon. Delayed gastric emptying causes erratic gastrointestinal absorption of any triptan taken p.o. (14).
The strength of our study is supported by the fact that we used only triptan-naive patients in each of our six treatment groups and that they were a negatively selected group of migraine patients (sumatriptan non-responders). The study aimed to show the clinical impact of treatment strategies and assist practitioners in finding the right treatment for their patients.
One limitation of the study is the small number of patients in each treatment group. Therefore, from a statistical point of view our results can be interpreted as more or less explorative. Because of the multiple testing problems in the study, the P-values reported for the time course of VAS scores for each treatment group can only be considered descriptive. Nevertheless, the final multivariable statistical model revealed that the treatment group using zolmitriptan 5 mg nasal spray was the only treatment group able to reduce the VAS score to ≤ 3 within 1 h (statistically significant). Furthermore, we have to acknowledge the fact that we did not employ quantitative sensory testing in each patient to examine CA. However, we used a reliable screening test for identifying patients with CA (3) and we performed a trial close to real-life conditions. Moreover, it should be noted that it seems difficult to draw any conclusions regarding early/late strategies from studies comparing treatment administered in different time frames and pain intensities, since the natural course of untreated attacks is unknown. There is no clear relationship between the point of time and the intensity of pain. Furthermore, the four-point scale could be biased towards better outcome when pain is mild, since a smaller change is needed to meet the definition of response.
Our clinical data may support the hypothesis that in patients with migraine and early allodynia it is important how quickly a specific concentration of 5-HT1B/D agonist is achieved. That reduction of central nociceptor sensitization results in enhanced efficacy and fewer associated allodynic symptoms. Further clinical and basic science research is needed to examine these hypotheses.
Acknowledgements
We gratefully acknowledge Professor Rami Burstein (Department of Anesthesia and Critical Care, Beth Israel Deaconess Medical Center, Boston, MA, USA) for his comments and suggestions on the manuscript. Furthermore, we would like to thank Dr Thomas Müller (Department of Laboratory Medicine, Konventhospital Barmherzige Brüder) for his important work and thoughtful discussion on statistics.