Abstract
Introduction: This study investigated the efficacy and tolerability of the highly selective iNOS inhibitor GW274150 in prophylaxis of migraine headache.
Subjects and methods: The study was conducted in two parts, each comprising a 4-week baseline period, a 12-week, double-blind, parallel-group treatment period, and a 4-week follow-up period. The study had an adaptive design in that findings of Part 1 of the study were used to inform the conduct of Part 2. Following an interim analysis at the end of Part 1, the trial could be stopped for futility or continued in Part 2 to study the full-dose response or to increase sample size in case initial assumptions had been violated. The primary end-point in both parts of the study was the probability of the occurrence of a migraine headache day during the baseline period and the treatment period.
Results: In Part 1, adult male and female patients with migraine received GW274150 60 mg (n = 37), 120 mg (n = 37), or placebo (n = 38) once daily for 12 weeks. In Part 2, female patients with migraine received GW274150 60 mg (n = 160) or placebo (n = 154) once daily for 12 weeks. GW274150 was no more effective than placebo for the primary efficacy end-point or any secondary efficacy end-point in Part 1 or Part 2. GW274150 was generally well tolerated.
Conclusions: GW274150 at doses predicted to inhibit iNOS >80% did not differ from placebo in the prophylaxis of migraine. The results do not support a role of iNOS inhibition in migraine prevention.
Introduction
Migraine attacks are hypothesised to arise from neurally induced cranial vasodilation that produces painful inflammation of the meninges (1). The inflammation underlying migraine appears to be mediated in part by nitric oxide, a vasoactive messenger molecule synthesised from arginine via nitric oxide synthase (NOS) (2,3). Both preclinical and clinical data support a key role of nitric oxide in migraine. In trigeminal ganglia cultures, application of nitric oxide donors caused a >4-fold increase in release of calcitonin gene-related peptide (CGRP), a putative mediator of meningeal vasodilation in migraine and a contributor to sensitisation of pain pathways, over that observed in unstimulated cultures (4). Consistent with these in vitro data, activation of NOS and consequent generation of nitric oxide in animal models of pain are associated with sensitisation of central pain pathways, an effect that can be reduced with NOS inhibitors (5–11). In humans, nitric oxide donors such as glyceryl trinitrate can induce migraine-like headache in migraineurs (12,13). Conversely, inhibition of nitric oxide can alleviate migraine and other headache types (3,14,15). For example, administration of a non-selective NOS inhibitor was effective versus placebo in the acute treatment of migraine as well as the treatment of chronic tension-type headache in limited, double-blind studies (14,15) although results of a small study do not suggest efficacy of selective iNOS inhibitors in the acute treatment of migraine (16). In an open-label study of prophylaxis of migraine headache (n = 20), the nitric oxide scavenger hydroxocobalamin reduced migraine attack frequency over 3 months by ≥50% in 53% of patients and ≥30% in 63% of patients (17). The finding in migraineurs that concentrations of nitric oxide metabolites and trigeminal vasoactive peptides are increased in venous blood during a migraine attack compared with pre-attack values is also consistent with a role of nitric oxide in migraine attacks (18).
The nitric oxide synthesising enzyme NOS is expressed in three isoforms: endothelial (eNOS), neuronal (nNOS), and inducible (iNOS) (3). eNOS and nNOS are both constitutive enzymes whereas iNOS is an inducible enzyme expressed after stimulation by substances such as cytokines and endotoxins. Among the isoforms, iNOS has been most strongly implicated in mechanisms relevant to migraine (19–21); however, little is known about the clinical effects of iNOS inhibition in migraine. Whether selective inhibition of iNOS is effective in the prevention of migraine has not been reported to date.
GW274150 is a potent and selective iNOS inhibitor that produces analgesia in animal models of inflammatory pain (22,23). GW274150 has >100-fold selectivity for human iNOS over human eNOS and >80-fold selectivity for human iNOS over human nNOS (22). Unlike non-selective NOS inhibitors, which are associated with undesirable effects (e.g. changes in blood pressure and heart rate) attributed to inhibition of the constitutive isoforms eNOS and nNOS, GW274150 doses estimated to inhibit iNOS >90% (as inferred from suppression of exhaled NO measurements) were well tolerated (unpublished data and (24). The study reported herein was conducted to assess the efficacy and tolerability of the GW274150 in the prophylaxis of migraine.
Subjects and methods
This Phase II study was conducted in two parts, each comprising a 4-week baseline period, a 12-week treatment period, and a 4-week follow-up period. Patients could participate in only one part of the study. The study had an adaptive design in that findings of Part 1 of the study were used to inform the conduct of Part 2. Part 1 investigated two doses of GW274150 (selected based on human pharmacology, tolerability, and safety) and placebo. Following a planned interim analysis at the end of Part 1, the trial could be stopped for futility or continued in Part 2 to study the full-dose response or to increase sample size in case initial assumptions had been violated. The design of each part of the study was randomised, double-blind, placebo-controlled, and parallel-group. Part 1 was conducted in 18 study centres in Belgium, Denmark, France, Germany, The Netherlands, and Norway. Part 2 was conducted in 48 study centres in Denmark, Finland, France, Germany, Italy, The Netherlands, Norway, and Spain. Ethics committees approved the study protocol for each study centre. This study was conducted in accordance with ‘good clinical practice’ (GCP) and all applicable regulatory and privacy requirements including those originating from the Declaration of Helsinki.
Patients
Males and females (Part 1) or females only (Part 2) 18–55 years old with International Headache Society (IHS) criteria-defined migraine with or without aura (25) for at least 1 year with an onset before the age of 50 years were eligible if they had ≥3 migraines but <15 migraine or non-migraine headache days per month during each of the 3 months prior to the screening visit and the baseline period and if their migraines were consistent in incidence and severity. Females were eligible if they were unable to bear children or, if able to bear children, they were not pregnant and using adequate contraception. Patients were ineligible if they had taken migraine prophylactic medication within 1 month of the screening visit or used an opiate for the first-line acute treatment of migraine; had a history of ergotamine, triptan, opioid, or combination medication intake on ≥10 days per month or simple analgesic intake for ≥15 days per month for ≥3 months; had experienced lack of efficacy of two or more ≥8-week trials of prophylaxis of migraine headache or had stopped migraine prophylactic treatment because of intolerance; or had uncontrolled hypertension (i.e. resting systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg after three readings) at the screening visit. All patients provided written informed consent.
Procedures
In Part 1, patients were randomized 1:1:1 (and stratified according to number of baseline migraine headache days [≤8 versus >8]) to receive GW274150 60 mg, GW274150 120 mg, or placebo once daily for 12 weeks. In Part 2, patients were randomized 1:1 (and stratified according to number of baseline migraine headache days [≤8 versus >8]) to receive either GW274150 60 mg or placebo once daily for 12 weeks. The values for median percentage predicted iNOS inhibition for the 60-mg and 120-mg doses of GW274150 were 82% and 90%, respectively, for the duration of the dosing interval. Use of acute migraine medication was permitted for break-through migraine attacks.
Each day during the baseline period and the treatment period in both parts of the study, patients recorded, in an electronic diary, details about use of study medication as well as the presence or absence of migraine or non-migraine headache and associated symptoms of nausea, vomiting, photophobia, and phonophobia. For break-through migraines, patients also recorded the date and time of the start and the resolution of the migraine, whether or not acute migraine medication was used, pain severity (0, none; 1, mild; 2, moderate; 3, severe) before the use of medication for break-through migraine, and maximum pain severity.
End-points
Data from Part 1 and Part 2 were analysed separately. Efficacy data were analysed for the intent-to-treat population, defined as randomised patients who received at least one dose of study medication and provided at least one post-baseline efficacy assessment. The primary end-point in both parts of the study was the probability of a migraine headache day during the baseline period and the treatment period. A migraine headache day was defined as a calendar day with any occurrence of migraine headache pain of at least 30 min in duration.
Secondary efficacy end-points, calculated for the treatment period and for each 4-week portion of the treatment period for each part of the study, included the change from baseline in the number of migraine headache days, the proportion of patients with a ≥50% reduction in the number of migraine headache days, the change from baseline in the number of migraine attacks, the proportion of patients with a ≥50% reduction in the number of migraine attacks, the change from baseline in the mean peak migraine severity, and the change from baseline in mean migraine attack duration.
Tolerability was assessed by recording adverse events, defined as any untoward medical occurrences regardless of their suspected cause, that were reported by a patient or noted by a clinician during the studies. For each adverse event, investigators recorded whether or not they considered it related to study medication. Tolerability measures included the incidences of adverse events regardless of suspected cause, adverse events leading to premature withdrawal from the study, and serious adverse events (defined as adverse events that resulted in death, disability, or incapacity; were life threatening; required or prolonged hospitalisation; or were a congenital anomaly or birth defect). Adverse-event data were summarised for each part of the study for the safety population, defined as randomised patients who administered at least one dose of study medication. Inferential statistics were not done on the adverse-event data.
Other measures, results for which are not reported in this manuscript, included clinical laboratory tests, vital signs, and electrocardiograms (ECGs). In addition, a total of 4–5 blood samples per patient were obtained to assess GW274150 concentrations at steady state.
Data analysis
A planned interim analysis of data from Part 1 was conducted to inform the conduct of Part 2: If the treatment effect in Part 1 was strong, defined as at least a 40% difference between active treatment and placebo in the probability of a migraine headache day, then Part 2 was to proceed with up to four doses of GW274150 to define relationships between dose and efficacy and dose and safety. If more information was needed regarding the doses in Part 1, one or more of the doses studied in Part 1 was also to be compared with placebo in Part 2, and the sample size was to be increased to match the study objective. If no effect of active treatment was evident in Part 1, the study was to be terminated.
On the basis of results of the interim analysis (described below), option 2 above was chosen for Part 2.
For analysis of the primary end-point, the null hypothesis of no difference in the probability of a migraine headache day between the active treatment groups and placebo was tested by fitting a logistic hazard model to the observed data using non-linear mixed-effect modelling analysis (26). Missing data were not imputed. Parameter estimates from the null and alternative hypothesis models and confidence intervals (CIs) for all estimated and derived parameters were summarised.
For the primary end-point, the study was designed to show superiority of active treatment over placebo using a 2-sided significance test with a type I error rate of 0.05 without adjustment for multiple comparisons. Power calculations for Part 1 revealed that a sample size of 30 patients per treatment group was required to provide 95% power to detect a difference of 40% between active treatment and placebo treatment for the primary end-point. These calculations assumed a 20% placebo response rate and interpatient variability as indexed by the coefficient of variation (CV) of 30% for migraine headache days on treatment. Power calculations for Part 2 were adjusted considering the placebo response and observed variance as well as the lower-than-expected possible effect size in Part 1 (described below). These calculations revealed that a sample size of 120 patients per treatment group was required to provide 80% power to detect a difference of 25% between active treatment and placebo treatment for the primary end-point.
Analyses of secondary efficacy endpoints were done with a last-observation-carried-forward (LOCF) dataset when appropriate. Summary statistics were calculated for each secondary efficacy end-point. Differences between each GW274150 group and placebo were tested with a rank analysis of co-variance (ANCOVA) method with adjustment for country and baseline number of events for continuous variables and with Fisher’s Exact test or a conditional logistic regression model controlled for pooled centre for categorical variables.
Adverse-event data were summarised with descriptive statistics. GW274150 exposures at steady state for each patient were estimated via non-linear mixed-effect modelling analysis (26) to verify whether the GW274150 plasma concentrations were as expected and to estimate the individual exposures for relation to efficacy or safety end-points if needed.
Results
Part 1
Patients
Patient disposition, demographics, and migraine characteristics at baseline
Demographics were similar among treatment groups (Table 1). Patients were predominantly female and white with a mean age of approximately 40 years. IHS headache type and migraine history were similar among treatment groups (Table 1).
Efficacy
Summary of the probability of a migraine headache day at day 70 (intent-to-treat population)
Exploratory analyses to elucidate the results further suggest that rescue medication use was unlikely to have affected the outcome of Part 1 and that females might have responded more robustly than males to GW274150 60 mg. In females, the probability of a migraine headache day was reduced by 31% with GW274150 60 mg compared with placebo (95% CI –15% to 60%, not significant). Data in males were inconclusive as the representation of males at the 60-mg dose was limited to 12 patients.
Results for key secondary end-points during the last 4 weeks of the treatment period
Tolerability
Adverse events reported in ≥5% of patients in a treatment group
The number of patients who withdrew from Part 1 because of an adverse event was two with GW274150 60 mg, three with GW274150 120 mg, and two with placebo. Except for peripheral oedema, which was the reason for premature withdrawal in two patients in the group treated with GW274150 120 mg (and no patients in the other two groups), no single adverse event led to the withdrawal of more than one patient in a treatment group. Both of the occurrences of peripheral oedema that led to premature withdrawal were considered moderate in severity. One occurrence of peripheral oedema, in a 31-year-old white female, was associated with moderate drug hypersensitivity and severe weight increase. The second occurrence of peripheral oedema, in a 49-year-old white female, was associated with constipation.
No fatalities occurred during Part 1. During the treatment period of Part 1, one non-fatal serious adverse event of moderate tremor, which occurred 10 days after initiation of study medication in a 38-year-old white female being treated with GW274150 60 mg, was reported. The event resolved after 11 days. The investigator considered this event to be at least possibly related to study medication, and the patient was withdrawn from the study. The investigator also reported that the tremor could have been attributed to a localised epileptic seizure. The patient’s medical history included arteriovenous angioma that had been surgically addressed in 1983.
Implications of results of Part 1 for the conduct of Part 2
In Part 1, no statistically significant treatment advantage was found for either GW274150 60 mg or GW274150 120 mg compared with placebo. Although the higher-than-expected interpatient variability in treatment response and the larger-than-expected placebo response reduced statistical power, Part 1 was sufficiently powered to exclude the projected 40% treatment effect as a likely outcome. The sample size was increased in Part 2 based on the revised assumptions regarding the interpatient variability, placebo response, and minimal treatment effect (revised to a 25% reduction in migraine headache days versus placebo). In addition, based on exploratory analyses that suggested that females might respond more robustly to GW274150 60 mg than males, Part 2 was conducted in female subjects only and employed the 60-mg dose.
Part 2
Patients
In Part 2, the number of patients who received at least one dose of study medication (safety population) was 314 (160 in the GW274150 60-mg group and 154 in the placebo group). All of these patients provided evaluable efficacy data and, therefore, were also included in the intent-to-treat population. The majority (≥88%) of patients in each group completed Part 2 (Table 1). The most common reasons for premature withdrawal were patient’s decision and adverse events (Table 1).
Demographics were similar between treatment groups (Table 1). Patients were predominantly white with a mean age of approximately 40 years. IHS headache type and migraine history were similar between treatment groups (Table 1).
Efficacy
No statistically significant effect of study medication was found in the modelling analysis. The model best describing the data was the one that considered the active treatment to be the same as placebo. Addition of a parameter describing the drug effect did not statistically significantly improve the fit of the model. Table 2 shows treatment effects at day 70 for the full model. For key secondary efficacy end-points, as with the primary efficacy end-point, no statistically significant benefit of GW274150 60 mg versus placebo was observed (Table 3). The placebo response in Part 2 was larger than that in Part 1, particularly for key secondary efficacy end-points (Tables 2 and 3).
Tolerability
The percentage of patients with at least one adverse event during the treatment period of Part 2 was 56% with GW274150 60 mg and 53% with placebo (Table 4). The number of patients who withdrew from Part 2 because of an adverse event was five with GW274150 60 mg and four with placebo. No single adverse event led to the withdrawal of more than one patient in a treatment group.
During the treatment period of Part 2 of the study, no fatalities occurred. Serious adverse events were reported in five patients (three in the GW274150 group and two in the placebo group). The serious adverse events in the GW274150 group were tendon rupture (1 patient), musculoskeletal chest pain (1 patient), and pelvic fracture, sternal fracture, and pneumothorax (1 patient). The serious adverse events in the placebo group were appendicitis and coronary artery occlusion (1 patient each). None of these serious adverse events were considered by the investigator to be related to study medication.
Steady-state plasma exposures of GW274150 were as expected in the majority of patients.
Discussion
GW274150 administered for prophylaxis of migraine headache at doses predicted to inhibit iNOS by >80% was not more effective than placebo at reducing the probability of having a migraine headache day in this study. Steady-state plasma exposures of GW274150 were as expected in the majority of patients, a finding that suggests that the anticipated levels of iNOS inhibition were achieved. The study’s adaptive design was intended to maximise the ability to detect a treatment effect by using the results of Part 1 to inform the conduct of Part 2. Like Part 1, Part 2 showed no statistically significant treatment benefit with GW274150 versus placebo for the primary end-point. Likewise, no advantage of GW274150 versus placebo was observed for secondary efficacy end-points in either part of the study. This fairly rigorous test of a selective iNOS inhibitor in prophylaxis of migraine headache does not support a role of iNOS inhibition in migraine prevention.
The placebo response rates were larger than expected in Parts 1 and 2 of this study. In a previous meta-analysis of placebo response in the prophylaxis of migraine, the pooled estimate of the placebo response (i.e. the percentage of placebo-treated patients with >50% reduction in migraine attacks at end-point) was 21% (27). While this placebo rate was smaller than those observed in the current study (32% in Part 1 and 53% in Part 2 for proportion of patients with ≥50% reduction from baseline in migraine attacks), the placebo response rates observed in the current study lay within the rather large range observed in previous placebo-controlled studies (6–55%) (27).
While the results cast doubt on a role of iNOS in migraine pathophysiology, they do not disprove the hypothesis that nitric oxide could be an important mediator of migraine. Nitric oxide is generated and released not only by iNOS but also by eNOS, which is found in the endothelium, and nNOS, found in neurons (3). Activity of these constitutive enzymes could be responsible for generating nitric oxide that contributes to migraine. Evidence consistent with this possibility was reported in studies using intravital microscopy to investigate the effects of NOS inhibitors on neurogenic dural vasodilation and CGRP-induced dilation in the rat (28). Neurogenic dural vasodilation was inhibited by a non-specific NOS inhibitor and an nNOS inhibitor whereas CGRP-induced dilation was inhibited by a non-specific NOS inhibitor and an eNOS inhibitor. An iNOS inhibitor was ineffective at inhibiting both neurogenic inflammation and CGRP-induced dilation in this model. (While these results in rats were known at the time the current study was conducted, the effects in humans were unknown and were believed to warrant further investigation.) While the results of the current study do not refute the hypothesis that nitric oxide could be an important mediator of migraine, they are consistent with the possibility that NO and its metabolites might not be critically important in the typical migraine attacks seen in clinical practice (versus experimentally induced attacks).
GW274150 was generally well tolerated in this study, a result consistent with previous findings (23). This result suggests that GW274150 is tolerable up to a dose of at least 120 mg, which is associated with a median predicted iNOS inhibition of approximately 90%. The adverse-event data from this study reveal no safety signal that should preclude additional exploration of GW274150 for other nitric-oxide-associated inflammatory conditions.
Conclusions
GW274150 at doses predicted to inhibit iNOS by >80% did not differ from placebo in the prophylaxis of migraine. The results do not support a role of iNOS inhibition in migraine prevention.
Footnotes
Acknowledgements
GlaxoSmithKline funded this study. The authors acknowledge Jane Saiers PhD (The Write Medicine, Inc.) for assistance with writing this manuscript. Dr Saiers’ work was funded by GlaxoSmithKline. The authors also acknowledge Andreas Krause of Pharsight Corporation (now with Actelion) for the assistance with the analysis of Part 2. The following investigators are acknowledged for their participation in the study. Belgium: L. Herroelen, UZ Leuven, Leuven; P. Louis, Eeuwfeestkliniek Antwerpen; J. Schoenen. CHR La Citadelle, Liège. Denmark: J. Beier. Sankt Anne Plads 2, Odense; H. Jespersen. Bryggertorvet 4C, Ølstykke; P. Tfelt Hansen, T. Wienecke. Glostrup Universitetshospital Neurologisk Afdeling, Glostrup. Finland: M. Ilmavirta. Suomen Terveystalo Oyj, Jyvaskyla; M. Kallela, M. Farkkila. Helsingin Paansarkykeskus, Helsinki; J. Liukkonen, R. Anttinen, H. Merivuori. Mikkeli Neurologipalvelu, Mikkeli; E. Sako, J-P. Eralinna. Turun Paansarkykeskus, Turku; M-L. Sumelahti. Tammer-Tutka, Tampere. France: H. Belhadj-Tahar. Groupe Santé Recherche SA, Toulouse; P. Boyer. Cabinet Médical, Luynes; A. Duplan, V. Terracol. Cabinet Médical, Montbrison; M. Hérent. Cabinet Médical, Anzin; D. Lejay. Cabinet Médical, Vieux Conde; C. Lucas. CHU – Hôpital Roger Salengro, Lille; G. Mick, Centre Hospitalier Pierre Bazin, Voiron; D. Saillard. Cabinet médical, Tours; A. Serrero. Cabinet Médical, Chilly Mazarin; F. Spilthooren. Cabinet Médical, Evreux. Germany: K. Bachus-Banaschak, R. van Heys. Praxis Mehringdamm, Berlin; H.C. Diener, C.F. Schorn, P. Stude, A. Gendolla, M. Küper. Universitätsklinikum Essen, Essen; S. Evers, K. Biehl, M. Marziniak. Universitätsklinikum Münster, Münster; H. Göbel, A. Heinze. Schmerzklinik Kiel, Kiel; J.P. Jansen, S. Saschowa, S. Bas. Praxis Schönhauser Allee, Berlin; B. Möckesch, M. Möckesch. Praxis Bahnhofstr, Weinheim; T. Nolte, U. Kern, H. Braun, L. Tarau. Facharztzentrum Medicum, Wiesbaden; R. Schellenberg, V. Schellenberg. Institut für ganzheitliche Medizin und Wissenschaft GmbH, Hüttenberg. Italy: P. Barbanti. I.R.C.C.S. San Raffaele Pisana, Roma; E. Del Bene. Azienda Ospedaliero-Universitaria Careggi, Firenze; F. Devetag. Ospedale Civile di Feltre – U.L.S.S. 2 Feltre, Feltre; G. L. Mancardi. Azienda Ospedaliera Universitaria San Martino, Genova; B. Panascia. A.O. Universitaria Vittorio Emanuele, Catania; L. Pini. Policlinico di Modena – A.O. Universitararia-Modena; P. Tanganelli. Presidio Osp. GE-Ponente Osp. Padre Antero Micone, Genova. The Netherlands: M.E.J.M. Baggen. Huisartspraktijk, Grubbbenvorst; L.J.M. Boonman-de Winter, E. Kristensen. SHL, Etten-Leur; J.A. Carpay. Tergooiziekenhuizen, Blaricum; A. de Jong. Huisartspraktijk, Hoogwoud; P.J.J. Koehler. Atrium Medisch Centrum, Heerlen; K.J. Moes. Huisartspraktijk, Geldermalsen; W.M. Mulleners. Canisius-Wilhelmina Ziekenhuis, Nijmegen; P.H.M. Pop. VieCuri Medisch Centrum, Venray; T.J. Tacke, J.J.W. Prick. Streekziekenhuis MiddenTwente, Hengelo; F.A.A.M. Vermetten. Huisartspraktijk, Spijkenisse. Norway: E. Åserud, M. Kleiven. Cort Piil Helsesenter, Bergen; S. Elle. Elverum legesenter AS, Elverum; M. Grepp. Nøste legekontor, Lier; J.O. Gronert. Flattum legesenter, Hønefoss; H.O. Høivik, K. Høye. Olav Medi 3 Innlandet AS, Hamar; A. Nome Dueland. Sandvika nevrologpraksis, Sandvika; A.C. Poole, I. Monstad. Sjølyst medisinske senter AS, Oslo. Spain: J.M. Laínez, A.P. Balbastre. Hospital Clínico de Valencia, Valencia; M. Sanchez Del Rio. Hospital Ruber Internacional, Madrid; J. Pascual. Hospital Del Mar, Barcelona.