Abstract
In order to investigate the plausible association of migraine recurrence with anxiety and depressive symptoms, a multicentre, randomized, double-blind, placebo-controlled, crossover clinical trial was conducted using sumatriptan as a vehicle drug. Migraineurs were randomly assigned to receive either 50 mg sumatriptan or placebo for three consecutive migraine attacks, and then cross over to the other treatment for three more migraine attacks. The primary measurements were the observed rate of migraine recurrence in relation to (i) patient's mood condition, measured by the Hamilton rating scales for depression and anxiety and (ii) patient's general health and functioning measured by the Symptom Checklist (SCL)-90-R. Migraine recurrence was defined as any migrainous headache that occurred within 24 h post treatment, only when pain free at 2 h was achieved. The analysis of efficacy was performed on 376 migraine attacks treated with sumatriptan and 373 attacks treated with placebo. Recurrence ratio was 14.1% and 5.1%, respectively (P = 0.045). The number needed to treat for pain free at 2 h post dose was 5.4. Recurrence was not affected by Hamilton scores for depression or anxiety, SCL-90-R scores or treatment. Apparently, depressive or anxiety symptoms do not influence headache recurrence in acute pharmaceutical migraine treatment, but further investigation is required.
Introduction
In acute treatment of migraine attacks, recurrence is defined as any headache relapse after initial pain relief post treatment. It was first reported in migraineurs in the carefully designed and monitored clinical trials investigating efficacy and safety of sumatriptan (1–3). In these trials migraine recurrence following initial relief occurred in about 35% of successfully treated attacks (4,5). Over time, recurrence has been documented for all subsequent triptans, in percentages varying from 13 to 53% (5–7). Additionally, migraine recurrence is a significant reason for patients' treatment dissatisfaction. In a telephone survey of 688 migraineurs by Lipton and Stewart, patients expressed a strong preference for a migraine medication that could provide complete pain relief, which was considered important or very important by 87% of the participants, followed by no recurrence by 86% (8). Thus, migraine recurrence is a significant treatment variant that may be responsible for secondary treatment failure (9). Early interpretation associated migraine recurrence with drug pharmacokinetics and pharmacodynamics (10,11). Indeed, long plasma half-life triptans display lower recurrence, but the phenomenon persists, as it does in triptans that show a greater bioavailability than sumatriptan or penetrate the blood–brain barrier (BBB) and reach the neurons (12,13), suggesting that other co-factors may also attenuate headache relapse.
Migraine often co-exists with anxiety and depression (14). It has been suggested that the lifetime association of these two conditions may be explained by a model of bi-directional influence between migraine and major depression (14,15). On the other hand, anxiety and depression amplify pain perception (16). We have previously shown that patients suffering from medication overuse headache revealed higher scores in Hamilton scales for anxiety and depression than healthy subjects (17). Migraine is the most common primary headache type complicating drug overuse (18,19). Since headache recurrence is a significant reason for drug overuse, it is likely that the combination of migraine with anxiety or depression may enhance and magnify migraine recurrence, which in turn results in drug overuse.
To investigate this hypothesis, or at least part of it, we designed a cross-over, placebo-controlled study, using sumatriptan as vehicle drug, in order to explore if depressive and anxiety symptoms influence migraine recurrence.
Patients and methods
Patients
Male and female migraineurs, 18–65 years old, with a current history of migraine with or without aura, were eligible to participate. The diagnosis of migraine was made according to the International Headache Society (IHS) criteria (20) after direct interview. Patients had to have a history of 1–10 migraine attacks per month of any severity for the last 2 months prior to enrolment. Pregnant or lactating women as well as women of childbearing potential who refused to use a valid method of contraception were excluded. Subjects with a history of basilar, hemiplegic or ophthalmoplegic migraine, cardiovascular disease, Raynaud's disease, epilepsy or structural brain lesions, impaired renal or hepatic function were also excluded. Before undergoing any trial-related examination or other procedure, patients had to return a signed and dated informed consent form.
Study design
The study followed a randomized, double-blind, placebo-controlled, crossover design. Each patient was scheduled to undergo three study visits within a maximum period of 7 months. Following evaluation of the inclusion/exclusion criteria, all eligible patients that agreed to participate were randomized to receive either 50 mg sumatriptan or identically matching placebo tablets and a diary card. Neither the patient nor the investigator was aware of the treatment allocation. Patients were instructed how to use the diary cards at each migraine attack in order to record treatment, migraine severity (0 = no pain; 1 = mild pain; 2 = moderate pain; 3 = severe pain), date and time of onset of migraine symptoms, relief, recurrence and use of study medication. Patients were asked to use treatment for three consecutive migraine attacks and document all treatment-relevant information in the diary card. Subsequently, patients returned to the study centre for the second visit, where the diary card and any other study-related material was collected and each patient was crossed to the other treatment group, according to the randomization plan. Participants were advised to use the study drug and the diary card for three consecutive migraine attacks, as in the first study period, and then return for the final visit, where all study materials were collected.
Drug treatments
Patients were instructed to receive the first study dose (sumatriptan or placebo) as early as possible, following the onset of the migraine attack. A second tablet was allowed only in case of recurrence of a migraine and not for inadequate relief. ‘Recurrence’ was defined as the reappearance of a migraine of any grade (3/2/1) within 2–24 h, following initial complete pain relief at 2 h. If a migraine attack did not respond to treatment until 4 h post dose, the patient was allowed to use their regular medication as ‘rescue’. A 24-h pain-free interval was required between two different attacks. Patients were allowed to continue any prophylactic medication they used, as long as it did not contain ergotamine or its derivatives and non-steroidal anti-inflammatory drugs and as long as the dose was stable for at least 3 months prior to study entry and the participant agreed to keep using the stable dose throughout the study. The use of antidepressive drugs was disallowed during the trial for all subjects.
Scales for anxiety and depressive symptoms and for quality of life
In order to evaluate depressive or anxiety symptoms, the Hamilton scales for anxiety and depression (HA and HD, respectively) (21,22) were administered by the investigator in all three study visits. The HA rating scale is a valid and reliable tool, one of the first to attempt to score the clinical status of patients in ‘neurotic anxiety state’, and currently one of the most widely used symptom rating scales in the world. The HD rating scale is also a well-known questionnaire with almost universal use as the ‘gold standard’ for the measurement of depression. The Greek versions of both scales were used for the needs of this trial, as previously (17). In addition, participants filled out the Symptom Checklist (SCL)-90-R questionnaire (23), a questionnaire for self-report of quality of life and psychological distress, widely used as part of the evaluation of chronic pain patients and other non-psychiatric populations. In the final visit, patients were asked for their satisfaction and treatment preference between the two treatment groups. All investigators participated in a teaching course focused on these structured interviews and self-reported questionnaires.
The protocol was approved by the National Organization for Medicines, the National Ethics Committee and the scientific committees of all participating health facilities. The study was conducted in seven centres in Greece, according to all applicable local laws and regulations for clinical trials in human subjects, the principles of the Declaration of Helsinki and the International Conference on Harmonization/Good Clinical Practice guidelines.
Statistical analysis
Responses to sumatriptan or placebo intake and migraine recurrence were analysed by logistic regression analysis with treatment and treatment sequence as fixed factors. Baseline values of HA, HD and SCL-90 scores were fitted as covariates. The interaction of HA, HD and SCL-90 scores with treatment were also considered in the models. All tests were considered to be significant at a 0.05 level.
One hundred and fifty patients were required to be enrolled into the trial to ensure that at least 125 patients completed the study, with an estimated drop-out rate of 15%, under the assumption of 6.1 mean difference of Hamilton scale between treatment groups with common standard deviation of 12.2 (24,25).
Efficacy measurements
The primary measurements were the observed rate of migraine recurrence in relation to (i) patient's anxiety and depressive symptoms, as measured by use of Hamilton rating scales for depression and anxiety, and (ii) patient's general health and functioning measured by the SCL-90-R. The secondary measurements of the study included (i) the percentage of headache-free (grade 0) patients at 2 h post dose, (ii) the percentage of patients who used rescue medications, and (iii) patients' satisfaction and treatment preference. To measure this, participants were asked to evaluate in diary cards their satisfaction of the treatment they received by using a four-point scale (1 = satisfied, 4 = not satisfied).
Safety measurements
The safety end-point of the trial was the incidence of any adverse events (AEs) throughout the study. Any abnormal measurement of vital signs (i.e. heart rate, blood pressure) that was clinically significant in the investigator's opinion was also reported as an AE.
Results
Patients
A total of 160 patients were enrolled and randomized. All participants were consecutive patients seen at headache centres. Five of the enrolled patients withdrew their consent and 24 were lost to follow-up, hence end-points were assessed on 131 evaluable patients, who reported a total number of 749 migraine attacks documented in diary cards. Analysis was performed on a per-protocol basis. Eighteen of the evaluable patients were male (13.7%) and 113 female (86.3%), with a mean age of 40 years (range 19–63 years). No major protocol violations were reported.
Primary measurements
Recurrence rate is independent of treatment, depression status, anxiety status and patient's general health and function. (logistic regression analysis modelling migraine recurrence)
HD, Hamilton scale for depression; HA, Hamilton scale for anxiety; SCL-90-R, Symptom Checklist 90-R.
Hamilton Anxiety and Depression rating scores
All evaluable patients (n = 131) had a baseline HA assessment. The mean HA score at baseline was 9.98 (median = 8). Using as a threshold for mild/severe anxiety symptoms a score of ≥ 14 points, 81.7% of migraineurs were not suffering from mild or severe anxiety symptoms. The mean HA scores at visits 2 and 3 were 8.95 and 8.79, respectively. The mean reduction (± S.E.M.) from baseline was 1.0 (± 0.5) points after the use of sumatriptan and 1.2 (± 0.5) points after placebo. One hundred and thirty patients had a baseline HD assessment, demonstrating a mean score of 9.78 points (median = 8). The majority of enrolled patients (79.2%) had a baseline HD score < 14, signifying that these patients were free of mild or severe depressive symptoms. The mean HD scores of patients who received sumatriptan and placebo were 8.18 (± 0.5) and 8.12 (± 0.5), respectively, and there was a mean reduction of 1.6 (± 0.) points following treatment with either sumatriptan or placebo.
SCL-90-R
The mean score at baseline for all evaluable patients was 64.68 (± 4.7) points and, following the use of sumatriptan and placebo, it was significantly reduced to 48.63 (± 4.1) and 49.60 (± 4), respectively (reduction of 15.4 points in both treatments, P < 0.01), but no significant difference between sumatriptan- and placebo-treated groups was observed (P > 0.1).
Secondary measurements
As expected, response was significantly higher in the sumatriptan-treated migraine attacks. Patients reported that they were pain free at 2 h post dose in 39.6% of migraine attacks treated with sumatriptan vs. 21.2% with placebo (P < 0.0001). The number needed to treat (NNT) for pain free at 2 h post dose was 5.4. In 66/385 cases (19.2%) of sumatriptan-treated migraine attacks, the use of a rescue medication was deemed necessary, while rescue treatment was used in 74/383 cases (21.4%) of placebo-treated migraines. No statistically significant difference was noted between the two groups. Patients who received treatment with sumatriptan were significantly more satisfied compared with those treated with placebo (37.3% vs. 19.3% P = 0.003).
Safety
A total number of four AEs were reported during the trial. Three of the AEs were reported when the patient was under treatment with placebo (i.e. status migrainosus, back pain and vomiting), and the fourth (vomiting) after sumatriptan treatment. The investigators judged back pain and vomiting not to be drug-related events. None of the AEs was characterized as serious; study treatment discontinuation was not deemed necessary and all AEs resolved.
Discussion
In this randomized, double-blind, placebo-controlled, crossover trial aimed to show the relation of migraine recurrence after 50 mg sumatriptan to anxiety and depressive symptoms, the Hamilton rating scales for anxiety and depression, as well as the SCL-90-R scores were not related to migraine recurrence. NNT for pain free at 2 h after oral treatment with 50 mg sumatriptan was 5.4. The use of rescue medication did not differ between treatments. Migraineurs were more satisfied when treated with sumatriptan than with placebo. No serious AEs were reported.
Methodological issues
We conducted a double-blind, placebo controlled, crossover, multicentre clinical trial to investigate the possible association of anxiety and depression-like symptoms with migraine recurrence rate, using sumatriptan as a vehicle drug. We had no specific reasons to choose this particular drug, apart from a medium recurrence ratio reported in previous trials in comparison with other triptans (4,5). We decided to cross over treatments and investigate migraine recurrence within individuals when treated with placebo or the active drug, instead of following the two-parallel design, subsequently divide migraineurs into depressed and non-depressed, in an uncontrolled trial. We used the HA and HD scales to scan only for anxiety and depressive symptoms, not to make a psychiatric diagnosis. Previous experience with the HA and HD scales was good (17), and they are a stable part of the initial medical interview taken in Athens Naval Hospital, Headache Clinic. In this trial, all interviewers were trained by a psychiatrist (A.T.) in order to rate all participants accordingly in the same manner. Sample size and other methodology issues were in line with IHS suggestions for symptomatic migraine clinical trials (26). In order to investigate whether frequency of migraine relates to recurrence or not, participants should have 1–10 migraine attacks per month during the previous 2 months. We did not find any relation. Because we wished to replicate daily clinical practice, we allowed current prophylactic treatments (with the exception of antidepressants), although this could decrease migraine relapses. Our analysis of data excluded any treatment influence. Some could argue, however, that patients with comorbid anxiety and depression-like disorders were excluded by excluding patients with current antidepressive treatment. Antidepressants, on the other hand, reduce anxiety and depressive symptoms that were scanned by HA and HD for correlations with migraine recurrence. Because we wanted to avoid the worse outcome (bias the primary end-point of the study), we decided to bar antidepressants from prophylaxis. We used a cut-off threshold of 14 points in the score of HA and HD to define mild to severe anxiety or depressive symptoms. Although this may lack evidence-based documentation, it gives a picture of the patients' mood condition together with the mean and median values of HA and HD scores (17). We performed a logistic regression analysis, however, to see if HA, HD and SCL-90-R scores were associated with migraine recurrence. After both treatments, all scales significantly decreased, although no change in prophylaxis was made. Since the SCL-90-R scale is also useful in measuring patient progress or treatment outcomes, the observed improvement in all scales may reflect the healthcare the participants enjoyed during the study.
Migraine recurrence
In this study population migraine recurrence was not associated with anxiety or depressive symptoms as measured by HA and HD, or with general health or daily functioning as measured by SCL-90-R. According to the IHS, migraine recurrence or relapse occurs when a patient is pain free within 2 h after treatment and headache of any severity returns within 48 h (26). However, in most triptan clinical trials migraine recurrence was assessed as any headache relapse within 24 h post treatment, after initial pain free or relief, which is why in our study we used the 24-h time window. To avoid any overlap between attacks and/or recurrence, a 24-h pain-free period within the attacks was required. Differences in definition of recurrence (e.g. pain relief vs. pain free before any headache relapse), in addition to current preventive antimigraine treatment may account for the potential low recurrence ratio in our study. Recurrence ratio was significantly higher in sumatriptan than in placebo because the pain-free ratio was higher in sumatriptan-treated attacks.
Migraine is often comorbid with anxiety and depression-like disorders, as has been shown in migraineurs attending headache centres (17) and in the general population (27). We found no high prevalence of these conditions in our patients, probably because we excluded patients treated with antidepressants for methodological reasons. Secondary parameters evaluated the accuracy of the study, since they confirm the sumatriptan results from previous studies (5–7,28,29). The drug was well tolerated and no short-term or permanent drug discontinuation due to an AE was deemed necessary. Others have investigated migraine recurrence extensively in previous studies. They found that migraine relapsed more frequently in patients with more severe attacks and longer untreated attack duration, as well as in women with menstruation-related migraine. Patients with a sensation of an ongoing attack, despite headache relief after sumatriptan treatment, also reported more frequent migraine recurrence. They concluded that migraine recurrence is largely patient dependent (i.e. a consistent feature of each attack in a patient) rather than attack related (i.e. may vary per attack within a patient) (11), which is close to our opinion. Although some studies have shown that headache recurrence after sumatriptan is not related to pharmacokinetic or pharmacodynamic differences between patients (10), other triptans with longer half-life display a better recurrence ratio in clinical trials (5). In addition, penetration of BBB may influence recurrence, but centrally acting triptans like eletriptan also show high recurrence ratios (5). Bioavailability may also be important, but clinical data do not show this. Undoubtedly, pharmacodynamics and pharmacokinetics play a very significant role in post-treatment migraine recurrence. Recently, it has been speculated that neurogenic inflammation might also be involved (30). Whether or not neurogenic inflammation is indeed involved, recurrence of migraine remains a clinical problem to be resolved, as demanded by migraineurs. Repeated treatment seems to be effective (31), but patients still ask for a better solution (8). The aim of this study was exactly this, to uncover specific patients' characteristics that enhance recurrence. Unfortunately, the results do not support the hypothesis that anxiety/depressive symptoms, or general health and functioning could influence recurrence. If this was the case, then special care and treatment could avoid recurrence in these patients.
Conclusions
Apart from pharmacokinetic and pharmacodynamic factors, no other hypotheses have been raised or investigated as plausible aetiological co-factors in migraine recurrence. Although further investigation is needed, the results of this trial do no support our hypothesis that anxiety or depressive symptoms could influence migraine recurrence. A positive relation would have significant clinical importance.
Footnotes
Acknowledgments
This study was supported by a grant provided by GlaxoSmithKline Greece.