Abstract
Background: Endpoints used to evaluate the efficacy of acute anti-migraine drugs do not measure the tolerability. Sustained pain-free response with no adverse events has been recommended as a composite endpoint which measures the efficacy and tolerability attributes that patients desire.
Methods: The aim of this study was to evaluate new composite efficacy-plus-tolerability endpoints based on a post-hoc analysis of patient-level data from a previous randomized, placebo-controlled trial of the calcitonin gene-related peptide (CGRP) receptor antagonist, telcagepant, and zolmitriptan in the acute treatment of migraine. Endpoints were 2–24-hour sustained pain freedom and no adverse events from 0–24 hours (SPF24NAE), 2–24 hour sustained pain relief and no adverse events from 0–24 hours (SPR24NAE), pain freedom at 2 hours and no adverse events from 0–24 hours (PF2NAE), and pain relief at 2 hours and no adverse events from 0–24 hours (PR2NAE).
Results: Compared with placebo, both telcagepant 300 mg and 150 mg achieved nominal superiority (p values <.05) for SPF24NAE, SPR24NAE, PF2NAE and PR2NAE. Zolmitriptan 5 mg showed nominal superiority versus placebo for SPF24NAE, SPR24NAE and PF2NAE, but not PR2NAE. Telcagepant 300 mg showed nominal superiority versus zolmitriptan for SPF24NAE, SPR24NA and PR2NAE.
Conclusion: Composite efficacy-plus-tolerability endpoints may be useful for facilitating comparisons between treatments.
Introduction
Since the introduction of the triptans (5HT1B/1D receptor agonists) almost two decades ago, the primary outcome measure that defined a responder has been the percentage of patients with headache relief two hours after dosing (headache response) (1). However, surveys from migraine sufferers and analyses of clinical trials demonstrate that the most desirable characteristics of migraine therapy are a rapid onset of complete pain relief, a long duration of activity and a low rate of side effects (2,3). Causes of dissatisfaction with current acute therapy include length of time taken to achieve pain relief, inconsistent effect, headache recurrence and side effects, the latter of which may lead to a delay in taking acute prescription or discontinuation of therapy (4).
In an effort to measure endpoints that are most desired by patients, the pain-free response at two hours is currently recommended by the clinical trials subcommittee of the International Headache Society (1). Compared with headache relief (reduction of pain to mild or none), this endpoint is considered to be more intuitive, less sensitive to placebo effects and better able to produce dose-response relationships, and is preferred by patients (1,5). Because of the importance of headache recurrence, a 24-hour sustained pain-free (SPF) measure (pain-free at 2 hours without recurrence or use of additional acute medications for 24 hours) has been proposed to encompass the effects of a single dose of an acute migraine medication during a 24-hour period (6). The SPF measure is also associated with a higher migraine-specific quality of life score (7).
Unfortunately, even the SPF outcome measure does not capture all of the attributes of acute medications that are most desirable to patients, notably the absence of adverse events. Therefore, a new composite endpoint, “SPF with no adverse events”, has been proposed (8). A recent study demonstrated that with few exceptions, there was a strong and significant relationship between triptan efficacy and tolerability, such that triptans with higher SPF rates had higher adverse event rates (8). This study was based on a meta-analysis of randomized controlled triptan trials using group or population-level data, an approach which has a number of limitations compared with analyses based on individual patient data (9).
Telcagepant (MK-0974) is a novel oral calcitonin gene-related peptide (CGRP) receptor antagonist with efficacy that appears comparable to that of triptans but with fewer overall adverse side effects (10,11). A post-hoc analysis of a large randomized, placebo-controlled trial that evaluated telcagepant and zolmitriptan (11) was conducted to evaluate new composite efficacy-plus-tolerability endpoints which are close to patients' needs and might reflect the characteristics of this new medication better than traditional endpoints. The analysis used individual patient data.
Methods
Full details of the study methods are provided in Ho et al. (11). The following section summarizes information relevant to the current analysis.
Patients were eligible for the study if they were ≥18 years of age, had a history of migraine for at least one year and in the two months prior to the screening visit had experienced one to eight moderate or severe migraine attacks per month with or without aura (12) that typically lasted 4–72 hours untreated. Patients taking migraine prevention medication were allowed to enter the study provided that their prescribed daily dose had not changed during the three months prior to screening. The study was approved by the appropriate ethical review committee for each site and each patient provided written informed consent.
This was a randomized, double-blind, placebo- and active-controlled, parallel-group, outpatient study. Patients were allocated in equal ratios to telcagepant 150 mg, telcagepant 300 mg, zolmitriptan 5 mg or placebo. Patients were instructed to take study medication when they experienced a moderate or severe migraine attack. Patients had the option of taking a blinded optional second dose of study medication if they still had a moderate or severe migraine attack two hours after dosing or if they experienced a return of headache within 48 hours post–initial dose. Patients who were randomized to either zolmitriptan or placebo as their initial treatment were allocated to receive placebo for their optional second dose, whereas those initially randomized to telcagepant were allocated to receive either a second dose of telcagepant or placebo in equal randomization ratio.
Following the initial dose of study medication, patients recorded subjective assessments of migraine symptoms as well as the use of any rescue medication in a paper diary. Headache severity was recorded using a 4-grade scale (no pain, mild pain, moderate pain, severe pain) at baseline (0 hours – time of taking study medication) and at 0.5, 1, 1.5, 2, 2.5, 3, 4, 6, 8 and 24 hours postdose. For those patients who had pain relief (reduction of pain to mild or none) or pain freedom (no pain) at two hours, presence or absence of headache worsening (recurrence) within 2–24 hours was recorded. Patients were also instructed to record information about any adverse events that occurred up to the time they returned to the clinic. Patients were instructed to return to the study site within approximately seven days after treatment.
Statistical analysis
The present analysis was performed on a post-hoc basis and was intended to be hypothesis generating. The following post-hoc endpoints were evaluated: (i) SPF and no adverse events over 24 hours (SPF24NAE), defined as 2–24 hour SPF (pain-free from 2–24 hours without use of a second study dose or rescue medication) and no adverse events from 0–24 hours; (ii) sustained pain relief and no adverse events over 24 hours (SPR24NAE), defined as 2–24-hour sustained pain relief (reduction of pain to mild or none from 2–24 hours without use of a second study dose or rescue medication) and no adverse events from 0–24 hours; (iii) pain-free at 2 hours and no adverse events over 0–24 hours (PF2NAE); (iv) pain relief at 2 hours and no adverse events over 0–24 hours (PR2NAE).
Analyses of the above endpoints were conducted using the same methods as those used to analyze the other pre-specified binary endpoints in the trial (11). The full-analysis-set (FAS) was the primary population for assessing efficacy. For each endpoint, the FAS included all treated patients who had a baseline headache severity score and at least one postdose measurement occurring at, or prior to, two hours postdose. Non-baseline missing data were imputed using a last-observation-carried-forward approach (no baseline values were carried forward). Patients were counted in the treatment group to which they were randomized. Model-based p values for treatment comparisons were computed using a logistic model adjusting for geographic region (USA/ex-USA), baseline migraine severity (moderate/severe) and age (continuous). p values are referred to as “nominal” because of the post-hoc nature of the analysis and the lack of adjustment for multiplicity.
Results
Patient demographics
Characteristics of patients and the treated migraine attack at baseline: values are number (%) of patients, except for age where the mean is given
SD = standard deviation. NSAID = nonsteroidal anti-inflammatory drug.
N = number of treated patients; sample sizes differed slightly from this for some characteristics due to missing data. Likewise, some percentages do not add up to 100% (e.g. the different categories of baseline function) due to missing data.
Endpoints
Number of patients/number evaluable (% [95% CI]) meeting each endpoint in each treatment group
CI = confidence interval. Data shown are observed values. p values are from a model adjusting for geographic region, baseline migraine severity and age.
From Ho et al. (11).
*p < .05, **p < .01, ***p ≤ .001 for comparison of active treatment versus placebo.
#p < .05, ##p < .01, ###p ≤ .001 for comparison of telcagepant versus zolmitriptan.
Note: No comparisons were performed between the two telcagepant doses.
Discussion
The challenge for acute migraine treatment is to provide patients with a treatment that offers a rapid onset of complete and sustained pain relief with minimal or no adverse effects, thus restoring the patient’s ability to function normally. This large randomized, controlled trial provided the opportunity to evaluate new, clinically meaningful composite efficacy-plus-tolerability endpoints for telcagepant, a new oral CGRP receptor antagonist, as well as for an established triptan. This post-hoc analysis, which was meant to be hypothesis generating (i.e. to serve as a preliminary evaluation to potentially guide further research), suggested that telcagepant 300 mg was nominally superior to zolmitriptan 5 mg for the endpoints of SPF24NAE, SPR24NAE and PR2NAE. Although the response rates for these new measures using current acute medications are low, they encompass many of the medication attributes that are important to patients, and might be considered for inclusion in future trials comparing acute migraine treatments.
Telcagepant, a new oral CGRP receptor antagonist, appeared to have efficacy comparable to that of established triptans, but with a tolerability profile comparable to placebo in two randomized studies (10,11). Adverse events associated with current acute therapies have been shown to result in delayed use by patients during acute attacks, which leads to inferior efficacy and higher levels of attack-related disability. The generally favorable tolerability of telcagepant could potentially help reduce this barrier to early treatment while pain is mild, a strategy that has been demonstrated to enhance outcomes and reduce disability with triptans (13–18).
An advantage of the present analysis, compared to the previous analysis evaluating SPF with no adverse events as an endpoint among different triptans (8), was the use of individual patient data rather than aggregate population-level data (9). In the latter case, unverified assumptions about the interdependence of efficacy and tolerability were required to perform the calculations (9). In addition, the present analysis was a head-to-head comparison of data within a study versus the inherently difficult to make cross-study comparisons.
The limitations of this study are several. First, the analysis is post hoc and therefore only hypothesis generating in purpose. These results need to be confirmed in an appropriately powered study in which the composite endpoints are prespecified and multiplicity adequately addressed. Secondly, the zolmitriptan dose used in the study was 5 mg. In many countries, the approved starting dose is 2.5 mg. The 2.5 mg dose may be better tolerated while maintaining similar efficacy to the 5 mg dose (5). Finally, the evaluation of side effects did not take into account type, severity and duration. Adverse events range from mild and short-lived to severe and longer-lasting. In addition, certain side effects (e.g. chest pressure) may be more significant to patients than others (e.g. flushing). Mild, transient side effects may not be troublesome to patients, in which case efficacy considerations might predominate. Alternatively, some side effects (e.g. raised levels of laboratory measures) may be asymptomatic to patients in the short term but critical for their long-term health, or side effects may occur rarely but be important when they do occur. Future studies that evaluate composite efficacy-plus-tolerability endpoints could take into account those side effects that are deemed by the patient or physician to be moderate or severe or long-lasting. However, it should be recognized that, while useful for facilitating comparisons between treatments at a group level, no single composite measure will substitute for a full assessment of a drug’s efficacy and safety profile.
Footnotes
Disclosure
Merck & Co. provided financial support for the study. DWD has received honoraria from and has had consulting agreements with Allergan, Pfizer, Merck, Endo, OrthoMcNeil, Coherex, MAP, Neuralieve, Addex, Solvay, Eli Lilly, Neuraxon, Minster, HS Lundbeck, and Kowa; has research grants from AstraZeneca, Medtronic, St Jude, and Advanced Neurostimulation Systems; and has independent support from that National Institute of Neurological Diseases and Stroke (USA) and the Mayo Clinic. JK, CA, CL and TWH are employees of Merck & Co., and own stock and/or stock options in the company.