Abstract
Introduction
In a multicenter, double-blind, randomized, parallel-group, placebo-controlled, single-dose study (n = 1555), a fixed combination of acetaminophen 500 mg, acetylsalicylic acid 500 mg, and caffeine 130 mg (AAC) was compared with ibuprofen 400 mg (IB) and placebo (PLA) for acute treatment of migraine.
Subjects and methods
An exploratory post-hoc analysis compared AAC with IB and PLA in the subset of patients with severe pain at baseline (n = 660).
Results
At most time points, AAC and IB relieved the pain and associated symptoms of severe migraine significantly better than PLA (p ≤ 0.05). AAC was significantly superior to IB for pain relief at 45 minutes and at one, two, three, and four hours postdose (p < 0.04); pain intensity difference from one hour through three hours (p < 0.05); headache response at two hours (p = 0.04); functional disability reduced to little or none at three hours (p = 0.013); freedom from phonophobia at three hours (p = 0.04) and photophobia at 15 minutes postdose (p = 0.03); and use of rescue medication (p = 0.018). AAC patients also reported meaningful pain relief 16 minutes faster than IB patients (132 minutes vs 148 minutes, p = 0.026).
Conclusions
In patients with severe baseline migraine pain, AAC and IB are significantly more effective than PLA, and AAC provides significantly faster and more effective pain relief than IB.
Introduction
Migraine is a chronic neurologic disorder that causes recurrent attacks of head pain accompanied by a combination of photophobia and phonophobia and/or nausea, and/or vomiting (1). Although the frequency and severity of attacks vary, evidence suggests that the majority of migraineurs have moderate or severe pain during attacks, with slightly more than half (50.3%) reporting severe pain, and 21.8% reporting extremely severe pain (2).
Treatment guidelines developed in the United States and Europe (3–5) recommend prescription medication as first-line therapy for patients with moderate or severe attacks, but most migraineurs do not use them, preferring to self-treat their condition exclusively with nonprescription medications (2,6,7) such as acetaminophen (APAP), acetylsalicylic acid (ASA), and ibuprofen (IB), as well as fixed combinations of one or more analgesics with caffeine (CAF). In fact, data have shown that several agents widely available without a prescription—APAP (8), ASA (9), and IB (10), as well as the combinations APAP-CAF (11) and APAP-ASA-CAF (12)—can be effective in relieving migraine. Other studies have found that IB (13) and an APAP-ASA-CAF combination (12) are both significantly more effective than either APAP or ASA monotherapy for relief of migraine. Furthermore, in a study of nonprescription medications that did not exclude patients who typically vomit or require bed rest, a fixed combination containing APAP 500 mg, ASA 500 mg, and CAF 130 mg per two-tablet dose (AAC) was significantly more effective and faster at relieving migraine pain than IB (400 mg) (14). To refine understanding of the relative benefits of AAC and IB in patients with severe migraine, we extracted the subset of patients with severe baseline migraine pain from the study comparing AAC with IB (14) for a post-hoc efficacy analysis.
Subjects and methods
Subjects
The design and conduct of this double-blind, randomized, parallel-group, PLA-controlled study has been previously described (14). Subjects were diagnosed with migraine (according to the criteria of the second edition of International Classification of Headache Disorders, ICHD-2 (15)) and were at least 18 years old, in good general health, had untreated attacks of at least moderate pain intensity and had experienced at least one attack every two months, but not more than six times per month, during the previous year. No patients were excluded based on the requirement for bed rest or presence of frequent vomiting. Subjects were excluded if they had head pain that may have been caused or aggravated by recent head or neck trauma, cluster headache, specific migraine variants, or other serious nonmigraine causes of headache or used analgesic drug products for headache on more than 12 days per month.
Design
Trained interviewers screened potential subjects, and those selected to participate provided written informed consent and a complete medical history and underwent physical and neurologic exams (including urine pregnancy testing for women). Subjects were randomized (3:3:1 ratio) to receive double-blinded study medication containing a single, four-tablet dose of either (1) two unbranded AAC tablets (APAP 500 mg, ASA 500 mg, caffeine 130 mg) and two dummy IB tablets; (2) two unbranded IB tablets (400 mg IB) and two dummy AAC tablets; or (3) four dummy tablets. They were instructed to treat a single acute migraine attack of at least moderate intensity and, if possible, to avoid rescue medication for at least two hours.
Efficacy measurements
Subject-rated efficacy measurements.
From the above-recorded efficacy outcomes, the following additional efficacy measures were derived: pain intensity difference from baseline, headache pain intensity reduced to mild or none (headache response, Y/N), headache pain intensity reduced to none (pain free, Y/N), total of pain relief scores (TOTPAR), and use of rescue medication (Y/N).
The prospectively defined primary efficacy endpoint was TOTPAR through two hours postdose.
Safety assessments
Subjects recorded the occurrence of any unusual symptoms in a self-report treatment diary, which was reviewed by the investigator at follow-up to confirm and document whether any adverse events (AEs)—unusual symptoms or side effects—occurred during the treatment phase. Clinical laboratory data were not collected.
Sample size calculation
The original primary efficacy population was the intent-to-treat (ITT) population, which included all subjects who medicated and supplied any efficacy assessments. The planned size of this population was calculated to achieve a pre-planned level of statistical power based on the efficacy parameter TOTPAR4, as described elsewhere (14). The size of the cohort selected for the present post-hoc analysis was not based on pre-planned statistical power considerations.
Statistical analysis
All results of this post-hoc analysis must be considered exploratory in nature and reported p values must be interpreted as informal measures of the strength of the evidence. Consequently, no attempt has been made to adjust for multiplicity. Statistical significance was declared if p ≤ 0.05. When statistical tests are reported for multiple timepoints with a summary p-value (e.g. p ≤ 0.041 in Figure 2), it should be understood that the tests done over the timepoints are not statistically independent.
The subset of subjects who rated their baseline pain intensity as severe was extracted for post-hoc efficacy analysis. The chi-square test was used to compare the characteristics of the treated attack of this subset with the remainder of the original ITT population.
Any pain intensity assessments made more than 15 minutes after taking rescue medication were set to the more severe of either the baseline score or the score made immediately before rescue; the associated pain relief assessments were set to zero. If results were missing for an evaluation before the final evaluation, results were interpolated linearly from the evaluations adjacent to the missed evaluation. If results were missing because a subject did not complete the study, all missing scores were extrapolated from the last non-missing assessment through the end of the study.
Treatments were compared on time to meaningful pain relief with the log rank test, stratified by investigator, and Kaplan-Meier curves were generated for visual comparison. Pain intensity difference, pain relief, and TOTPAR were analyzed at each time point using analysis of covariance, with treatment group and investigator as main factors and baseline pain intensity as the covariate. On binary (Y/N) outcomes—proportion of subjects who were pain free at a fixed time point; had headache response at a fixed time point; had nausea, photophobia, and phonophobia at a fixed time point; took rescue medication within four hours—treatments were compared with the Cochran-Mantel-Haenszel test, stratified by investigator and baseline pain intensity.
Results
Study population
There were 660 subjects (AAC = 286, IB = 291, PLA = 83) in the post-hoc ITT population (Figure 1), and the demographic profiles and migraine histories of the treatment groups were similar (Table 2). The mean age was 38 years, 80% were female, and 68% were white. Most subjects (89%) had a history of severe pain with untreated attacks.
Patient disposition. AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo. Demographics and migraine history: subjects with severe pain at baseline. AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Characteristics of the treated attack: subjects with and without severe pain at baseline.
%: Percentage menstruating relative to number of females. AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Pain relief
Total pain relief
On the pre-specified primary endpoint, TOTPAR2, both AAC and IB were significantly superior to PLA (p ≤ 0.026), and AAC was significantly superior to IB (p = 0.037).
Pain relief
Both active treatments provided superior pain relief compared with PLA at all time points. AAC subjects had significantly higher pain relief scores than PLA subjects from 45 minutes through four hours postdose (p ≤ 0.004), and IB was significantly superior from 60 minutes through three hours postdose (p ≤ 0.041). AAC was superior to IB at all time points from 45 minutes through four hours postdose and, as Figure 2 shows, the difference between them was significant at 45 minutes and at one, two, three, and four hours postdose (p ≤ 0.039).
Pain relief: Subjects with severe pain at baseline (N = 660). AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Pain intensity
Headache response
More AAC subjects than PLA subjects had a headache response at all time points, and the difference between AAC and PLA was significant from 90 minutes through four hours postdose (p ≤ 0.013, Figure 3). A greater proportion of IB subjects than PLA subjects had a headache response from 30 minutes through four hours postdose, and the difference between IB and PLA was significant at three and four hours postdose (p ≤ 0.050). At two hours postdose, headache response for AAC was significantly higher than for IB (62% vs 54%, p = 0.036).
Headache response: Subjects with severe pain at baseline (N = 660). AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Pain free
Compared with PLA-treated subjects, a greater proportion of AAC-treated subjects were pain free from 30 minutes through four hours postdose, and there was a significant separation between AAC and PLA at three and four hours postdose (p ≤ 0.003). IB was significantly superior to PLA for pain free at four hours postdose (p = 0.025). More AAC subjects than IB subjects were pain free at 45 minutes and two, three, and four hours postdose, but the differences between AAC and IB were not statistically significant.
Pain intensity difference from baseline
Mean pain intensity difference scores for AAC subjects were higher than for PLA subjects from 30 minutes through four hours postdose, and they were significantly higher than PLA subjects from 45 minutes through four hours postdose (p ≤ 0.006). IB significantly outperformed PLA from 60 minutes through four hours postdose (p ≤ 0.039). As shown in Figure 4, AAC was superior to IB at all time points from 45 minutes through four hours postdose for pain intensity difference, with significant superiority over IB from 60 minutes through three hours postdose (p ≤ 0.046).
Pain intensity difference: Subjects with severe pain at baseline (N = 660). AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Time to event outcomes
Meaningful pain relief
Median meaningful pain relief was 48 minutes earlier for AAC subjects than for PLA subjects (p = 0.010), and 32 minutes earlier for IB subjects than PLA subjects (p = 0.250). As Figure 5 shows, median meaningful pain relief was significantly earlier for AAC subjects than IB subjects (16 minutes, p = 0.026).
Time to meaningful pain relief: Subjects with severe pain at baseline (N = 660). AAC: acetaminophen 500 mg, acetylsalicylic acid 500 mg, caffeine 130 mg per two-tablet dose; IB: ibuprofen 400 mg per two-tablet dose; PLA: placebo.
Rescue
Over the four-hour study period, 3.8% of AAC subjects, 8.6% of IB subjects, and 15.7% of PLA subjects took rescue medication. Use of rescue medication was significantly less frequent in AAC subjects than either PLA subjects (p < 0.001) or IB subjects (p = 0.018). There was no significant difference in use of rescue medication between IB and PLA.
Associated symptoms and functional disability
Significantly more AAC subjects than PLA subjects were nausea free at two hours postdose (p = 0.049); photophobia free at two, three, and four hours postdose (p ≤ 0.036); and phonophobia free from 90 minutes through four hours postdose (p ≤ 0.007). The differences between IB and PLA in incidence of nausea free were not statistically significant at any postdose time point. However, IB was significantly superior to PLA for photophobia free at two hours postdose (p = 0.043) and phonophobia free from 90 minutes through four hours postdose (p ≤ 0.044). The differences between AAC and IB in incidence of nausea free were not statistically significant at any postdose time point. However, AAC was significantly superior to IB for photophobia free at 15 minutes postdose (p = 0.028) and phonophobia free at three hours postdose (p = 0.036). AAC and IB were significantly superior to PLA from 60 minutes through four hours postdose for reduction of functional disability to little or none (p ≤ 0.005 for AAC and p ≤ 0.031 for IB). AAC was significantly superior to IB at three hours postdose (p = 0.013).
Safety
The most common AEs for AAC and IB (nausea, dizziness) were comparable to those seen in the larger study population: mild, transient, and similar in incidence and severity to PLA (14).
Discussion
In this double-blind, randomized, parallel-group, PLA-controlled study, subjects with moderate or severe migraine pain were treated with AAC, IB, or PLA. The exploratory post-hoc analysis of the patient subset with severe migraine pain at baseline showed that AAC and IB were significantly more effective than PLA, and AAC was significantly faster acting and more effective than IB. AAC was significantly superior to IB on the primary endpoint, TOTPAR2, as well as on numerous clinically important measures of pain intensity, pain relief, functional disability, and migraine-associated symptoms. AAC also provided meaningful pain relief significantly faster than IB and with significantly less use of rescue medication. These results indicate that AAC is superior to IB, not only in the overall study population (14), but also in the population of migraineurs who treated an attack specifically featuring the most severe baseline migraine pain.
Unlike many studies of over-the-counter (OTC) medications in migraine, including those with a fixed combination similar to AAC (12,16), subjects in this study had ICHD-2 migraine; they did not represent a “nonprescription” population, nor was the severity of symptoms in previous attacks (i.e. vomiting or the need for bed rest) a criterion for exclusion. In the subset population assessed in this post-hoc analysis, subjects were significantly more disabled and affected by associated symptoms than subjects without severe pain at baseline, possibly because of the greater intensity of their pain. Nonetheless, AAC provided meaningful relief.
This study followed subjects for only four hours postdose, which limits insight into the sustained performance of these compounds in patients with severe baseline pain. Still, based on these findings, AAC could be an appropriate treatment for the full spectrum of migraine pain. Perhaps more important, they suggest that the recommendations for AAC in evidence-based guidelines (i.e. first-line only for mild or moderate attacks) should be reconsidered so that migraineurs with attacks featuring severe pain can also benefit from this effective treatment. In this study, AAC provided stronger and faster relief of baseline migraine pain than IB in the most painful episodes of migraine, highlighting the unique value of this medication as a particularly effective OTC treatment for the symptoms of migraine.
Conclusions
In this post-hoc analysis, AAC was a more effective analgesic than IB for the treatment of severe baseline migraine pain, as demonstrated by clinically meaningful and statistically significantly higher pain intensity difference and pain relief scores. Furthermore, AAC achieved significantly faster onset of meaningful relief than IB. Finally, AAC was associated with clinically meaningful and statistically significant relief of the symptoms associated with migraine, including nausea, photophobia, phonophobia, and functional disability.
Clinical implications
In a multicenter, double-blind, randomized, parallel-group, placebo-controlled, single-dose study (N = 1555), a fixed combination of acetaminophen 500 mg, acetylsalicylic acid 500 mg, and caffeine 130 mg (AAC) was compared with ibuprofen 400 mg (IB) and placebo (PLA) for the acute treatment of migraine. A post-hoc exploratory analysis compared AAC with IB and PLA in the subset of patients with severe pain at baseline (n = 660). AAC was a more effective analgesic product than IB for the treatment of severe baseline migraine pain, as demonstrated by clinically meaningful and statistically significantly higher pain intensity difference and pain relief scores. AAC achieved significantly faster onset of meaningful relief than IB. AAC was associated with clinically meaningful and statistically significant relief of the symptoms associated with migraine, including nausea, photophobia, phonophobia, and functional disability.
Footnotes
Funding
This work was supported by Novartis Consumer Health Inc, which funded this post-hoc analysis and the medical writing work of Christopher Caiazza.
Conflicts of interest
Novartis Consumer Health Inc paid an honorarium to Jerome Goldstein for performance of the clinical trial, as well as for preparation and presentation of the poster and manuscript, and employs Martina Hagen and Morris Gold.