Eletriptan 40 mg compared with the encapsulated conventional tablet form of sumatriptan 100 mg

Three fair-quality trials compared eletriptan 40 mg with the conventional tablet form of sumatriptan 100 mg.^24–26 In these studies, sumatriptan was put in a capsule to make it look like eletriptan so that the study could be double-blind. At 2 hours, a significantly greater proportion of patients were pain-free with eletriptan 40 mg than with the encapsulated conventional oral tablet form of sumatriptan 100 mg in 2 of 3 trials.^{24, 26} When we pooled data from all 3 trials, the combined rates were 35% (376/1063) for eletriptan 40 mg and 25% (272/1076) for the encapsulated conventional oral tablet form of sumatriptan 100 mg, with a relative risk of 1.47 (95% CI, 1.11 to 1.94) and a number needed to treat of 10. Two-hour rates of normal function were also significantly greater for eletriptan 40 mg than the encapsulated conventional tablet form of sumatriptan 100 mg in 2 of 3 trials:^{24, 26} 62% (569/913) for eletriptan 40 mg and 56% (457/819) for the encapsulated conventional tablet form of sumatriptan 100 mg, with a relative risk of 1.09 (95% CI, 0.86 to 1.38). We found rates of 24-hour sustained pain-free in only 1 trial, in which eletriptan 40 mg was superior to the encapsulated conventional tablet form of sumatriptan 100 mg (24% compared with 14%; P<0.05).²⁴ When Ferrari and colleagues¹¹ combined these data²⁴ with unpublished data for 24-hour sustained pain-free outcomes from an additional trial,²⁵ the resulting direct difference of −8 (95% CI, −14 to −3) still showed that eletriptan 40 mg was superior to the encapsulated conventional tablet form of sumatriptan 100 mg.

Findings from these trials engendered debate over whether encapsulation of the comparator triptan for blinding purposes suppressed their normal absorption rate and usual effectiveness. This concern has led to multiple studies comparing pharmacokinetic and clinical effects of the conventional tablet form of sumatriptan tablets with and without encapsulation.

In vitro and in vivo dissolution testing by the manufacturers of eletriptan and the conventional tablet form of sumatriptan have produced conflicting results.^53–55 In an in vitro dissolution study funded by the manufacturer of eletriptan,⁵⁴ no significant difference in dissolution rate (estimated as area under the curve) was found for the conventional tablet form of sumatriptan 100 mg, with or without encapsulation based on the ratio of geometric means of 0.99 (90% CI, 0.92 to 1.06). However, an in vivo study (Fuseau 2001), funded by the manufacturer of the conventional tablet form of sumatriptan, showed absorption was delayed between 0 to 2 hours after dosing (AUC₂) when the conventional tablet form of sumatriptan 50 mg was encapsulated compared to when it was not encapsulated in a sample of 26 healthy adults (geometric mean treatment ratio 0.79; 90% CI, 0.59 to 1.05) and in a sample of 30 adults during a migraine (n=30) (geometric mean treatment ratio 0.73; 90% CI, 0.52 to 1.02).⁵⁵ The Fuseau trial has been criticized by an investigator sponsored by the manufacturer of eletriptan for using twice as much magnesium stearate to encapsulate sumatriptan than was used in the original head-to-head trials of eletriptan and suggested that the greater quantity magnesium stearate could have hampered capsule dissolution and confounded absorption. Also, it is unclear why the Fuseau and colleagues evaluated only the 50 mg dose of the conventional tablet form of sumatriptan and not also the 100 mg dose or why they used a 90% confidence interval to evaluate statistical significance, rather than the more common and more stringent 95% confidence interval. Subsequently, in another study funded by the manufacturer of eletriptan involving 10 healthy volunteers, the conventional tablet form of sumatriptan 100 mg and encapsulated sumatriptan 100 mg were found to be similar in elapsed time to initial capsule disintegration (6 minutes compared with 5 minutes) and in mean time to complete disintegration (18 ± 14 minutes compared with 16 ± 7 minutes).⁵³

Meta-analyses have also been conducted to compare the 2-hour pain relief and pain-free outcomes from head-to-head trials of eletriptan and the encapsulated conventional tablet form of sumatriptan to those from all other trials of either eletriptan or the unencapsulated conventional tablet form of sumatriptan, respectively.^{11, 56, 57} But, none has conclusively found that the clinical efficacy of the conventional oral tablet form of sumatriptan 100 mg on 2-hour pain-relief or pain-free outcomes was significantly decreased in trials where it was encapsulated compared with trials where it was not encapsulated.

In their 2002 meta-analysis,¹¹ Ferrari and colleagues conducted a sensitivity analysis to examine how company sponsorship may have influenced results for sumatriptan and placebo comparators.¹¹ Because the eletriptan-encapsulated sumatriptan comparator trials were all conducted by Pfizer,^24–26 this provided an opportunity for qualitative indirect comparison of average absolute 2-hour pain-free rate for the conventional tablet form of sumatriptan 100 mg with and without encapsulation. For the outcome of 2-hour pain-free, the overall average absolute rate for sumatriptan 100 mg was 29% (95% CI, 27 to 31) and was 8% (95% CI, 7 to 9) for placebo. In the Pfizer-conducted eletriptan-sumatriptan comparator trials, however, Ferrari and colleagues found lower average absolute 2-hour pain-free rates for encapsulated sumatriptan 100 mg and for placebo, respectively. Although inconclusive, the findings of Ferrari and colleagues suggest the presence of heterogeneity between Pfizer-conducted and other company-conducted trials that could have influenced 2-hour pain-free results. However, because the pattern of non-encapsulated placebo was similar to that of encapsulated sumatriptan – lower efficacy in Pfizer-conducted trials – use of encapsulation for blinding could not be the only source of heterogeneity in these trials.

One meta-analysis compared the time course of response for the conventional tablet form of sumatriptan with and without encapsulation using model-based random-effects logistic regression techniques and data from 19 head-to-head and placebo-controlled trials.⁵⁶ No significant difference was found at any time point between 0 and 4 hours in proportion of patients who achieved pain relief for the conventional tablet form of sumatriptan with or without encapsulation.

In 2005, we conducted our own meta-analysis to compare the mean absolute rates of 2-hour pain relief and pain-free for eletriptan and the conventional tablet form of sumatriptan. We compared data from head-to-head trials of eletriptan 40 mg and the encapsulated conventional tablet form of sumatriptan 100 mg^24–26 with data from all other available head-to-head trials and placebo-controlled trials involving either triptan. Pooled absolute rates of 2-hour pain relief and absence of pain are shown in Table 3. For the conventional tablet form of sumatriptan 100 mg, the mean rates of 2-hour pain relief and pain-free were numerically lower when it was encapsulated compared to when it was not encapsulated, but overlapping confidence intervals suggest that the difference is not statistically significant. Unexpectedly, however, for eletriptan 40 mg, the mean rate of 2-hour pain relief and pain-free were numerically higher in trials where the comparator was the encapsulated conventional tablet form of sumatriptan compared to when the comparator was placebo or another unencapsulated triptan. But, here again, overlapping 95% confidence intervals suggest that the difference is not statistically significant.

Table 3

Pooled absolute rates of 2-hour pain-free and pain-relief (95% confidence intervals)

Overall, meta-analyses have provided suggestive evidence that sumatriptan’s usual efficacy was suppressed when it was encapsulated for blinding purposes in the Pfizer-conducted trials. However, because the pattern of lower efficacy was also seen for non-encapsulated placebo and a pattern of higher efficacy was seen for non-encapsulated eletriptan, our conclusion is that use of encapsulation cannot provide the entire explanation for the unexpected results in the Pfizer-conducted eletriptan-sumatriptan comparator trials.

Therefore, using meta-regression techniques, we explored the impact of potential sources of clinical heterogeneity including mean age, percentage of female subjects, and percentage with severe baseline pain. However, even after adjustment for those patient variables, we found that the modest differences persisted between 2-hour pain-relief and pain-free outcomes in the trials of eletriptan and the encapsulated conventional tablet form of sumatriptan 100 mg compared with those in other trials of either eletriptan or nonencapsulated sumatriptan. Other variables of interest were recruitment method, type of run-in period, type of prior migraine treatment, including whether the trial population had been previously exposed to triptans, and year the study was conducted, but the publications provided insufficient data to assess their effects. Other variables, such as the scientific group conducting the study, place of study, and sponsorship might contribute to the difference, but they are confounded with the effects of drug and were not included in the analysis.

We also explored the presence of unexplained post-randomization exclusions of treated patients as another possible explanation for the unexpected findings in the 3 head-to-head trials of eletriptan compared with the encapsulated conventional tablet form of sumatriptan 100 mg.^24–26 As in the majority of trials of triptans, the head-to-head trials of eletriptan and the encapsulated conventional tablet form of sumatriptan 100 mg excluded from their efficacy analyses an average of 16% of randomized patients who took no study medication for the primary reason that they did not have a treatable migraine during the study period. However, unlike in most other trials, an additional subset (mean=7%, range=5% to 12%) of treated patients who were not “evaluable” due to unspecified violations of the protocol were excluded from the 2-hour efficacy analyses in the head-to-head trials of eletriptan compared to the encapsulated conventional tablet form of sumatriptan 100 mg.^24–26

Using a “worst-case scenario” approach, we estimated pooled 2-hour pain-free rates for the all-treated populations which we compared for eletriptan and the encapsulated conventional tablet form of sumatriptan 100 mg based on both risk difference and relative risk meta-analyses using random-effects models (Table 4). All treated patients excluded from the eletriptan 40 mg groups were included in the “worst-case scenario” analyses as treatment failures and all treated patients excluded from the encapsulated conventional tablet form of sumatriptan 100 groups were included as if they achieved 2-hour pain-free outcomes. In contrast to published findings based on the “evaluable” populations, in our worst-case scenario analyses, the difference in rates of 2-hour pain-free between eletriptan and the encapsulated conventional tablet form of sumatriptan 100 mg was smaller and was no longer statistically significant.

Table 4

Head-to-head trials of eletriptan compared with the encapsulated conventional tablet form of sumatriptan 100mg: Comparison of 2-hour pain-free outcomes from (more...)

It is important to note that results from our “worst-case scenario” analysis are hypothetical and, without knowledge of the real reasons for the exclusion of the treated patients, it is not possible for us to assess whether such bias exists or to what degree. Therefore, meaningful interpretation of results from the head-to-head trials of eletriptan compared with the encapsulated conventional tablet form of sumatriptan 100 mg is still not possible.

Eletriptan 40 mg compared with encapsulated naratriptan 2.5 mg

We included 1 fair-quality trial of 483 adults that treated moderate to severe migraines and found eletriptan 40 mg to be superior to encapsulated naratriptan 2.5 mg in rates of 2-hour pain-free (35% compared with 14%; P<0.001), 2-hour normal function (60% compared with 52%; P=0.014), and 24-hour sustained pain-free (22% compared with 11%; P<0.05).²⁸

Eletriptan 40 mg compared with encapsulated zolmitriptan 2.5 mg

We included 1 fair-quality trial of 1337 adults that treated moderate to severe migraines and found eletriptan 40 mg to be similar to the lowest recommended dosage of zolmitriptan 2.5 mg (encapsulated) on rates of 2-hour pain-free (32% compared with 26%), 2-hour functional response (61% compared with 55%), and 24-hour sustained pain-free (20% compared with 17%).²⁷

Early intervention

The efficacy of eletriptan 40 mg administered while pain is mild has been demonstrated in 1 fair-quality placebo-controlled trial of 565 adults.⁵⁸ In this trial, patients were instructed to take trial medication as soon as they were sure that they were experiencing a migraine. Despite being encouraged to take the medication while the pain was still mild, almost half of patients reported pain that was moderate to severe upon treatment. Consequently, the investigators based analyses on only the subgroup of patients whose pain was still mild at baseline. In this subgroup, eletriptan 40 mg was superior to placebo in rates of 2-hour pain-free (68% compared with 25%; P<0.0001) and 24-hour sustained pain-free (56% compared with 18%; P<0.01). Based on our independent random-effects meta-analysis (Appendix D) for 2-hour pain-free, the relative risk was 2.72 (95% CI, 1.92 to 3.84) and the number-needed-to-treat was 2. For 24-hour pain-free, the relative risk was 3.21 (95% CI, 2.09 to 4.94) and the number-needed-to-treat was 3.

Work productivity

We included 2 placebo-controlled trials that evaluated the efficacy of eletriptan 40 mg in improving work productivity outcomes.^{59, 60} Eletriptan 40 mg reduced total time lost (4 compared with 9 hours; P not reported) and work time lost (2.5 compared with 4 hours; P=0.013) in 1 placebo-controlled trial.⁶⁰ In the other trial, improvements on the Work Productivity Questionnaire (PQ-7) were significantly greater for eletriptan 40 mg than placebo (+22.4 compared with +11.8; P<0.01).⁵⁹

Rizatriptan 10 mg compared with the conventional tablet form of sumatriptan

We included 4 fair-quality head-to-head trials comparing rizatriptan 10 mg with the conventional tablet form of sumatriptan 100 mg^{36, 37} and the conventional tablet form of sumatriptan 50 mg in patients with migraine of moderate to severe pain intensity.^{32, 33} Supplemental unpublished data for 3 of these trials was provided by the manufacturer.^{32, 33, 36}

In terms of quality, the main limitation for both trials of rizatriptan 10 mg compared with the conventional tablet form of sumatriptan 100 mg was a randomization process that did not achieve balance between treatment groups on all baseline characteristics. In the trial conducted by Tfelt-Hansen and colleagues, patients in the rizatriptan 10 mg group were significantly younger than patients in the conventional tablet form of sumatriptan 100 mg group (37 years compared with 39 years; P<0.01). The age difference was adjusted for in the analysis of the primary outcome of time to pain relief, but not for other outcomes.³⁶ In the trial by Visser and colleagues, patients in the conventional tablet form of sumatriptan 100 mg group were predominantly from tertiary referral centers in the Netherlands, and 62% had severe pain at baseline. In contrast, the rizatriptan 10 mg, 20 mg, and 30 mg and placebo groups consisted of patients from the Netherlands and the United States, with 47% to 51% having severe pain at baseline. The difference in proportion of patients with severe pain at baseline was statistically significant for only the comparison of the conventional tablet form of sumatriptan 100 mg (62%) with placebo (47%; P not reported).³⁷

Findings were mixed across these trials (Table 4) and do not demonstrate a clear advantage for rizatriptan over the conventional tablet form of sumatriptan 50 mg or 100 mg. Findings were most favorable for rizatriptan 10 mg over the conventional tablet form of sumatriptan 100 mg in the Tfelt-Hansen trial, which involved 1099 adults with migraine pain of moderate to severe intensity.³⁶ However, this trial differed from the others in one main way: Patients with prior exposure to rizatriptan were excluded, which limits the applicability of these findings to patients who are rizatriptan-naive. In the other 3 trials, patients were enrolled regardless of prior triptan use.^{32, 33, 37}

At 1 hour, rates of pain-free were generally higher in the rizatriptan 10 mg treatment groups, but only 1 difference in 1 trial reached statistical significance, a comparison with the conventional tablet form of sumatriptan 50 mg.³² At 2 hours, rates of pain-free and normal function were again generally higher in the rizatriptan 10 mg treatment groups, but the differences reached statistical significance only in the Tfelt-Hansen trial.³⁶

For the comparison of the conventional tablet form of sumatriptan 100 mg to rizatriptan 10 mg, although the difference in 2-hour pain-free reached statistical significance in only 1³⁶ of 2 individual trials,^{36, 37} when Ferrari and colleagues¹¹ pooled these trials’ data, the combined direct difference (−7) was statistically significant (95% CI, −13 to −1). For the comparison of the conventional tablet form of sumatriptan 50 mg to rizatriptan, even when Ferrari and colleagues pooled data from the 2 individual trials, the combined direct difference (−3) did not reach statistical significance for 2-hour pain-free outcomes (95% CI, −9 to +2).¹¹

At 24 hours, the rate of recurrence was similar for rizatriptan 10 mg and the conventional tablet form of sumatriptan 50 mg³² and 100 mg.^{36, 37} Data on sustained pain-free outcomes at 24 hours were not reported in the original publications. However, based on pooled direct difference estimates for 24-hour sustained pain-free outcomes that were calculated by Ferrari and colleagues using unpublished data obtained from the drugs’ manufacturers, differences between rizatriptan 10 mg and the conventional tablet form of sumatriptan 50 mg (−2; 95% CI, −7 to +3) and 100 mg (−4; 95% CI, −9 to +2) were not statistically significant.¹¹

For 24-hour quality of life, there were generally no significant differences in mean scores for the 5 domains of the Migraine-Specific Quality-of-Life Questionnaire across the trials comparing rizatriptan 10 mg with the conventional tablet form of sumatriptan 50 mg^{32, 33} or 100 mg.³⁶ The only exception was that the mean score on the Work Functioning domain was significantly greater for rizatriptan 10 mg than the conventional tablet form of sumatriptan 50 mg (12.9 compared with 12.3; P=0.029) in 1 of the 2 trials.³² Quality-of-life outcomes were not reported in the Visser trial of rizatriptan 10 mg and the conventional tablet form of sumatriptan 100 mg.

Rizatriptan 10 mg compared with naratriptan 2.5 mg

Rizatriptan 10 mg was superior to naratriptan 2.5 mg in 1 good-quality trial (N=522).³¹ However, limitations in consistency and applicability reduced the strength of the findings from this trial. Rizatriptan 10 mg was superior to naratriptan 2.5 mg on the 2-hour outcomes of time to pain relief (hazard ratio 1.62; 95% CI, 1.26 to 2.09), rates of pain-free (45% compared with 21%; P=0.001), and normal functioning (39% compared with 23%; P<0.001). At 2-hours, overall satisfaction was also measured using a 7-point scale (1=completely satisfied and 7=completely dissatisfied) and was significantly higher for rizatriptan 10 mg (3.55; P<0.001) than naratriptan 2.5 mg (4.21). But, inconsistent with 2-hour outcomes, differences between rizatriptan 10 mg and naratriptan 2.5 mg were not statistically significant on 24-hour outcomes. At 24 hours, similar numbers of patients on rizatriptan 10 mg and naratriptan 2.5 needed additional medication (40% compared with 46%; P not reported), had recurrences (33% compared with 21%; P not reported), and had improved scores on the Migraine-Specific Quality-of-Life Questionnaire (P not reported), including Work Functioning (11.73 compared with 11.86), Social Functioning (12.16 compared with 11.92), Energy/Vitality (11.56 compared with 11.95), Migraine Symptoms (12.42 compared with 12.37), and Feelings/Concerns (11.55 compared with 11.79).⁶¹ Additionally, the applicability of this trial was potentially limited due to its exclusion of patients with prior exposure to rizatriptan or naratriptan.

Rizatriptan 10 mg compared with zolmitriptan 2.5 mg

Rizatriptan 10 mg showed an advantage over the lowest recommended dose of zolmitriptan 2.5 mg on 2-hour outcomes in a fair-quality trial of 766 adults with moderate to severe migraine pain.³⁵ Patients were eligible for enrollment regardless of their prior triptan use, but only 30% had used any triptan within the past 30 days. Compared with zolmitriptan 2.5 mg, rizatriptan had a similar rate of 1-hour pain-free (13% compared with 10%) and superior rates of 2-hour pain-free (43% compared with 36%; P<0.05) and normal function (45% compared with 37%; P<0.05). At 24 hours, rizatriptan 10 mg and zolmitriptan 2.5 mg had similar rates of recurrence (28% compared with 29%) and similar mean scores on all 5 domains of the Migraine-Specific Quality-of-Life Questionnaire.

Consistency

We found 1 fair-quality placebo-controlled trial that examined the use of rizatriptan 10 mg for treatment of 4 consecutive migraine headaches.⁶² Rizatriptan showed consistently higher 2-hour response rates than placebo during headache 1 (77% [320/246] compared with 37% [30/82]; P<0.01), headache 2 (78% [228/291] compared with 37% [27/73]; P not reported), headache 3 (80% [207/259] compared with 28% [21/75]; P not reported), and headache 4 (74% [190/255] compared with 54% [31/57]; P not reported). However, it is unclear whether differences between rizatriptan and placebo groups in the number of patients excluded from the analyses of headache 2 (9% compared with 11%), headache 3 (19% compared with 8%), and headache 4 (20% compared with 30%) may have resulted in groups compared after headache 1 being dissimilar in important patient characteristics that could have biased the analyses.

Early intervention

The efficacy of rizatriptan 10 mg administered early in a migraine, while pain is mild, has been demonstrated in 2 identically designed, good-quality placebo-controlled trials named Rizatriptan TAME1 (Treat A Migraine Early) and TAME2.⁶³ Findings from TAME1 and TAME2 were both reported in a single publication. Eligibility criteria required a history of migraines that typically started out mild. The study plan was for patients to treat their migraines while still mild in severity and present for less than 1 hour, but not spontaneously resolving. In both trials, rizatriptan was superior to placebo in rates of 2-hour pain-free and 24-hour sustained pain-free. Rates of 2-hour pain-free for rizatriptan compared with placebo in TAME1 were 57% and 31%, respectively, and in TAME2 were 59% and 31%, respectively (P not reported for pairwise comparisons). Rates of 24-hour sustained pain-free for rizatriptan compared with placebo in TAME1 were 43% and 23%, respectively, and in TAME2 were 48% and 25%, respectively (P not reported for pairwise comparisons). Based on our independent random-effects meta-analysis (Appendix D), these findings resulted in a pooled relative risk of 1.86 (95% CI, 1.57 to 2.21) and a number-needed-to-treat of 3 for 2-hour pain-free outcomes. For 24-hour sustained pain-free rates, we calculated a pooled relative risk of 3.52 (95% CI, 1.67 to 7.42) and a number-needed-to-treat of 5.

Rizatriptan orally disintegrating tablet 10 mg compared with the conventional tablet form of sumatriptan 100 mg

We found no head-to-head trials that compared rizatriptan orally disintegrating tablet 10 mg to sumatriptan 100 mg; that evaluated quality-of-life, workplace, or consistency outcomes; or that evaluated early treatment of mild migraine. Two open, fair-quality trials demonstrated rizatriptan orally disintegrating tablet 10 mg to be superior to the conventional tablet form of sumatriptan 50 mg on preference and rates of 2-hour normal function and pain-free.^{39, 41} Similar numbers of patients had recurrence of migraine within 24-hours with both rizatriptan orally disintegrating tablet 10 mg and the conventional tablet form of sumatriptan 50 mg. Only 1 of the 2 trials reported 24-hour sustained pain-free outcomes, and the rate was significantly greater for rizatriptan orally disintegrating tablet 10 mg than the conventional tablet form of sumatriptan 50 mg (41% compared with 32.3%; odds ratio 1.47; 95% CI, 1.14 to 1.90).⁴¹

Rizatriptan orally disintegrating tablet 10 mg compared with eletriptan 40 mg

We also found 1 fair-quality, open head-to-head trial primarily designed to evaluate preference for rizatriptan orally disintegrating tablet 10 mg compared with eletriptan 40 mg in 439 adults who had no prior experience with either triptan.³⁸ Greater numbers of patients expressed a preference for treatment with rizatriptan orally disintegrating tablet 10 mg (61%; 95% CI, 56 to 66) than eletriptan 40 mg (39%; 95% CI, 34 to 44), with the most common reason being “relieved my headache pain faster.” At 2 hours, similar numbers of patients in the rizatriptan and eletriptan groups were completely or very satisfied with study medication (45% compared with 40%), were pain-free (52% compared with 50%), or had any functional disability (43% compared with 47%). Rates of 24-hour sustained pain-free were also similar for rizatriptan orally disintegrating tablet 10 mg (43%) and for eletriptan 40 mg (47%).

Zolmitriptan 5 mg compared with the conventional tablet form of sumatriptan

One fair-quality trial compared zolmitriptan 5 mg to the conventional tablet form of sumatriptan 100 mg in 1058 adults who had never been treated with either triptan.⁴⁵ Zolmitriptan 5 mg and the conventional tablet form of sumatriptan 100 mg had similar rates of pain-free at 1 hour (8% compared with 10%; rate ratio 0.70; 95% CI, 0.47 to 1.04)⁶⁵ and 2 hours (29% compared with 30%; rate ratio 0.98; 95% CI, 0.81 to 1.18),⁶⁵ no activity impairment at 2 hours (data not reported), recurrence at 24 hours (26% compared with 28%), and complete response at 24 hours (39% compared with 38%). In the Ferrari meta-analysis of unpublished data provided by manufacturers, the conventional tablet form of sumatriptan 100 mg and zolmitriptan 5 mg also had similar rates of 24-hour pain-free (direct difference −1; 95% CI, −5 to +6).¹²

For the comparison of zolmitriptan 5 mg to the conventional tablet form of sumatriptan 50 mg, 2-hour and 24-hour pain-free rates were published for only 1 of the 2 trials for 1522 (90%) of participants who treated at least 2 attacks.⁴⁶ Using those data and unpublished data for the other trial,⁴⁴ Ferrari and colleagues calculated pooled direct differences for 2-hour pain-free (0%; 95% CI, −4 to +4) and 24-hour sustained pain-free (−1%; 95% CI, −5 to +3), suggesting that zolmitriptan 5 mg and the conventional tablet form of sumatriptan 50 mg have similar effects on these outcomes.¹²

The 2 head-to-head trials comparing zolmitriptan 5 mg to the conventional tablet form of sumatriptan 50 mg also provided the best data on consistency. The first of these, conducted in the United States, compared zolmitriptan 2.5 mg and 5 mg to sumatriptan 25 mg and 50 mg.^{44, 66} Over 6 months, each patient was treated for up to 6 consecutive headaches. Patients were recruited from primary care, neurology, and research clinics. Of 1445 patients enrolled, 1212 treated at least 2 migraine headaches and 1043 completed the study. However, this trial has been criticized because it did not exclude patients who had previously taken sumatriptan.⁶⁷ There may have been a selection bias favoring zolmitriptan, since patients who responded inconsistently to sumatriptan in the past may be more likely to enroll in an experimental trial of a newer triptan. To assess consistency, the authors calculated the proportion of patients who responded in 2 hours in 80% to 100% of headaches (Table 6). The results indicate that the 2-hour response is not a reliable indicator of consistency across multiple migraine headaches.

Table 6

Consistency of response in Gallagher 2000

A good-quality trial of similar design was conducted in Europe.⁴⁶ In that trial, there were essentially no differences in efficacy among zolmitriptan 2.5 mg, zolmitriptan 5 mg, and sumatriptan 50 mg. The 3 treatments also had similar consistency across attacks: about 40% of patients in each group reported a 2-hour response in 80% or more of their headaches.

Zolmitriptan 2.5 mg compared with naratriptan 2.5 mg

An unpublished trial comparing zolmitriptan 2.5 mg with naratriptan 2.5 mg consisted of 2 parts. In Part 1, 553 adults were randomized to treat 1 headache with zolmitriptan 2.5 mg, naratriptan 2.5 mg, or placebo. The 438 who treated a headache and provided efficacy data were re-randomized to either zolmitriptan 2.5 mg or naratriptan to treat up to 3 more headaches in Part 2. According to the trial’s brief summary report, a higher proportion of patients in the zolmitriptan groups had headaches of severe intensity at baseline in both Parts 1 and 2. However, we could not examine the magnitude of these differences or any other baseline characteristics as their details were not provided in the trial summary report. It was noted that the baseline difference was more marked in Part 1 and was adjusted for in the analysis of 2-hour pain-relief data. The adjusted 2-hour pain-relief rate was similar for zolmitriptan 2.5 mg and naratriptan 2.5 mg (54% compared with 47%). Although the trial summary did not report 2-hour or 24-hour pain-free outcomes, Chen and colleagues obtained these data from the manufacturer and estimated risk ratios of 1.73 (95% CI, 1.10 to 2.72) and 1.04 (95% CI, 0.74 to 1.47), respectively.⁶⁵ However, as these risk ratios do not appear to have been adjusted for the above-described baseline differences in headache intensity, we interpret these risk ratios with caution.

Zolmitriptan orally disintegrating tablet compared with the conventional tablet form of sumatriptan 50 mg

In 1 head-to-head trial, 218 adults were randomized to open treatment with either zolmitriptan orally disintegrating tablet or the conventional tablet form of sumatriptan and were then crossed over to treat a second migraine with the alternative trial medication.⁴⁹ Results were reported for only the combined treatment periods. Patients with prior use of either trial medication within the past 3 months were excluded. The trial was designed to measure patient preference. The standard pain, associated migraine symptom, and functional capacity outcomes were not reported. Preference data were unavailable for 18 (10%) of patients. Because of these flaws, this trial was rated poor quality and its results will not be discussed here.

Comparisons of different zolmitriptan formulations

One good-quality, randomized trial (N=1372) compared double-blinded, double-dummy treatment with zolmitriptan nasal spray 0.5 mg, 1.0 mg, 2.5 mg, and 5.0 mg and oral zolmitriptan 2.5 mg.⁴⁷ Another trial used a crossover design to compare patient preference among zolmitriptan orally disintegrating tablet 2.5 mg, zolmitriptan standard oral tablet 2.5 mg, and zolmitriptan nasal spray 5 mg, but it was rated poor quality due to lack of blinding, presence of high attrition, and lack of separately reported results from the first treatment period.⁴⁸

The good-quality trial found zolmitriptan nasal spray 5 mg to be superior to zolmitriptan standard oral tablet 2.5 mg on rate of pain-free at 30 minutes (7% compared with 2%; P<0.05) and 45 minutes (10% compared with 5%; P<0.05) and on rate of resumption of normal activities at all time points (53% compared with 45%; P not reported). Zolmitriptan nasal spray 5 mg and zolmitriptan standard oral tablet 2.5 mg were similar on rate of 2-hour pain-free (38% compared with 37%) and rate of recurrence at 24 hours (26% for both). Zolmitriptan nasal spray 2.5 mg was similar to zolmitriptan standard oral tablet 2.5 mg in rate of pain-free at timepoints between 30 minutes and 1 hour, but was inferior at 2 hours (26% compared with 37%; P<0.05) and 4 hours (43% compared with 54%; P<0.05).

Early intervention

The efficacy of zolmitriptan standard oral tablet 2.5 mg administered while pain is mild has been demonstrated in 1 fair-quality placebo-controlled trial.⁶⁸ In this trial, 280 patients were instructed to administer treatment when pain was still mild and within 4 hours of onset. Zolmitriptan was superior to placebo in rates of 2-hour pain-free (43% compared with 18%; P<0.001) and 2-hour normal function (68% compared with 51%; P<0.01). The only 24-hour outcome reported was need for further medication, which was significantly lower after zolmitriptan 2.5 mg (46%) than placebo (71%; P<0.0001). Based on our independent random-effects meta-analysis (Appendix D), these findings correspond to a pooled relative risk of 2.41 (95% CI, 1.81 to 3.20) and a number-needed-to-treat of 4 for 2-hour pain-free outcomes.

Almotriptan 12.5 mg compared with the conventional tablet form of sumatriptan

Both trials comparing almotriptan 12.5 mg with the conventional tablet form of sumatriptan were rated fair quality due to differences between comparison groups at baseline, and both provided data on 2-hour pain-free and 24-hour recurrence outcomes.^{20, 69} Rate of 2-hour pain-free was consistently lower for almotriptan 12.5 mg in both trials. Compared with the conventional tablet form of sumatriptan 50 mg (25%), significantly fewer patients were pain-free at 2 hours after taking almotriptan 12.5 mg (18%; P=0.005). It is unknown, however, whether the higher mean body weight in the almotriptan group (74.5 kg compared with 72.3 kg; P=0.003) may have disadvantaged those patients’ treatment response. Compared with the conventional tablet form of sumatriptan 100 mg, fewer patients on almotriptan 12.5 mg were pain-free at 2 hours (28% compared with 33%), but this difference was not statistically significant.²⁰ At 24 hours, rates of recurrence for almotriptan 12.5 mg were slightly higher than for the conventional tablet form of sumatriptan 50 mg (27% compared with 24%)⁶⁹ and slightly lower than for the conventional tablet form of sumatriptan 100 mg (18% compared with 25%).²⁰ Differences in 24-hour recurrence rates were nonsignificant in both trials.

Sustained 24-hour pain-free, functional disability, and quality-of-life outcomes were not reported in either of the original trials comparing almotriptan 12.5 mg with the conventional tablet form of sumatriptan. Based on findings from a more recent review of almotriptan trials, however,⁷⁰ similar rates of patients had sustained 24-hour pain-free outcomes with almotriptan 12.5 mg and the conventional tablet form of sumatriptan 100 mg (rate ratio 0.86; 95% CI, 0.62 to 1.21).

Almotriptan 12.5 mg compared with zolmitriptan 2.5 mg

One good-quality trial provided evidence that almotriptan 12.5 and zolmitriptan 2.5 mg were similar on 2-hour and 24-hour efficacy outcomes in patients who were enrolled regardless of prior triptan use.²³ Both almotriptan and zolmitriptan tablets were encapsulated for blinding purposes. At 2-hours, almotriptan 12.5 mg and zolmitriptan 2.5 mg were similar in rates of pain-free (43% compared with 48%) and no functional impairment (47% compared with 49%). Almotriptan 12.5 mg and zolmitriptan 2.5 mg were also similar in rates of “excellent” satisfaction (16% compared with 15%) and 24-hour sustained pain-free plus no adverse events (29% compared with 32%).

Almotriptan 12.5 mg compared with rizatriptan 10 mg

One fair-quality trial was designed primarily to compare patient preference for open almotriptan 12.5 mg against open rizatriptan 10 mg in patients from Germany, Italy, and Spain who had never been treated with either triptan.²² Among the 255 of 327 patients in the 2-attack intention-to-treat population who recorded a preference for one triptan over another, half preferred almotriptan (n=128) and the other half preferred rizatriptan (n=127). Among the secondary efficacy variables analyzed (e.g., 2-hour pain-free; 2-hour pain-relief; sustained pain-free; sustained pain-free plus no adverse events; use of rescue medications; recurrence between 2–24 hours; recurrence between 24–48 hours), the only significant difference found indicated an advantage for rizatriptan 10 mg over almotriptan 12.5 mg on 2-hour pain-free outcomes (58% compared with 52%; P=0.03). This trial did not report quality-of-life or functional disability outcomes.

Indirect comparison of almotriptan with the conventional tablet form of sumatriptan 100 mg for 24-hour pain-free

In their meta-analysis of 53 triptan trials, Ferrari and colleagues included data from 3 abstracts of placebo-controlled trials of almotriptan 12.5 mg.^75–77 Using pooled data from the almotriptan 12.5 arms of these trials, they calculated a mean absolute rate of sustained pain-free, which they compared to the mean for the conventional tablet form of sumatriptan. The actual mean value and 95% confidence interval was not provided for almotriptan but it was described as being higher than for the conventional tablet form of sumatriptan 100 mg. However, this comparison did not assess or adjust for potential clinical or methodological heterogeneity across trials. Therefore, we suggest that this finding be interpreted with caution.

Consistency

We found 1 fair-quality, placebo-controlled trial that examined the use of almotriptan 12.5 mg for treatment of 3 consecutive headaches.⁷¹ The results of this trial demonstrated that a significantly greater number of patients achieved 2-hour pain-free outcomes in 3 of 3 headaches with almotriptan 12.5 mg than placebo (18% compared with 5%; P<0.05).

Early intervention

The efficacy of almotriptan 12.5 mg administered early in a migraine, while pain is mild, has been demonstrated in 2 fair-quality placebo-controlled trials named Act when Mild (‘AwM’)⁷³ and Axert^® Early Migraine Intervention Study (‘AEGIS’).⁷⁴ The ‘AwM’ trial was designed to compare early and non-early intervention and involved 4 treatment groups. For the purposes of this review, our interest was in the 2 treatment groups in which patients were randomized to administer treatment with almotriptan or placebo when pain was still mild and within 1 hour of onset. Results from the other 2 treatment groups, in which patients were randomized to administer treatment with almotriptan or placebo when pain was moderate to severe, were reported separately and will not be discussed here. In the Axert^® Early Migraine Intervention Study, patients were allowed to treat pain of any intensity, as long as it was within 1 hour of onset, but outcomes for mild and moderate-to-severe headaches were reported separately. In both trials, almotriptan was superior to placebo in rates of 2-hour pain-free and 24-hour sustained pain-free. Rate of 2-hour pain-free in ‘AwM’ was 49% for almotriptan and 25% for placebo (odds ratio 2.93; 95% CI, 1.62 to 5.31; P=0.0004), and in ‘AEGIS’ were and 37% and 24%, respectively (P=0.01). Rate of 24-hour sustained pain-free was 46% for almotriptan and 16% for placebo in ‘AwM’, and in the ‘AEGIS’ trial was 25% and 16%, respectively (P=0.040). Based on our independent random-effects meta-analysis (Appendix D), these findings correspond to a pooled relative risk of 1.71 (95% CI, 1.32 to 2.21) and a number-needed-to-treat of 6 for 2-hour pain-free outcomes. For 24-hour sustained pain-free rates, we calculated a pooled relative risk of 2.08 (95% CI, 1.12 to 3.86) and a number-needed-to-treat of 6. Functional disability and quality-of-life outcomes were also reported in a secondary publication of the ‘AEGIS’ trial.⁷² At 2 hours, mean functional disability scores showed that significantly more patients functioned normally with almotriptan than placebo (54% compared with 38%; P=0.007). At 24 hours, scores in all 5 domains of the Migraine Quality-of-life Questionnaire were consistently better for almotriptan than placebo.

Early intervention

The efficacy of reformulated sumatriptan 100 mg administered early in a migraine, while pain is mild, was demonstrated in a fair-quality trial of 432 adults who were instructed to administer treatment when pain was still mild and within 1 hour of onset.^{78, 79} Rate of 2-hour pain-free was 66% for reformulated sumatriptan 100 mg and 20% for placebo (P<0.001). At 24 hours, rate of sustained pain-free also was significantly greater for reformulated sumatriptan 100 mg than placebo (40% compared with 10%; P<0.001). From these data, we calculated a relative risk of 3.38 (95% CI, 2.65 to 4.30) and a number needed to treat of 2 for 2-hour pain-free and a relative risk of 4.09 (95% CI, 2.83 to 5.92) and a number needed to treat of 3 for 24-hour sustained pain-free.

Function and productivity outcomes from this trial were reported.⁷⁸ Compared with placebo, rate of normal function was significantly greater for reformulated sumatriptan 100 mg at 45 minutes (29% compared with 18%; P<0.05), 1 hour (50% compared with 25%; P<0.001), and 2 hours (60% compared with 28%; P<0.001). At 24 hours, significantly less time was lost on activities other than paid work for reformulated sumatriptan 100 mg (2.0 hours) than placebo (3.6 hours; P<0.05). However, lost time in paid work was similar for reformulated sumatriptan 100 mg and placebo (2.5 and 1.9 hours, respectively).

Indirect comparison of reformulated with the conventional tablet form of sumatriptan

In the absence of head-to-head trials that directly compared reformulated and the conventional tablet form of sumatriptan, we explored indirect comparisons between formulations using data from placebo-controlled trials. Data from placebo-controlled trials of reformulated sumatriptan⁸⁰ and the conventional tablet form of sumatriptan^{36, 37, 45, 81–85} were pooled, and combined relative risks and numbers needed to treat were generated for each triptan for 2-hour pain-free rates (Table 6). Estimates of relative risk were similar for the conventional tablet form of sumatriptan and reformulated sumatriptan and the large overlap of 95% confidence intervals did not suggest a clear advantage for either formulation over the other. However, the somewhat higher rate of 2-hour pain-free rates in the placebo group of the reformulated sumatriptan trial compared with those of the conventional tablet form of sumatriptan trials suggests the presence of at least some heterogeneity between the 2 sets of trials, likely in patient population or outcome assessment. Therefore, we caution against drawing firm conclusions about the comparison of reformulated and the conventional tablet form of sumatriptan until results from adjusted, quantitative, indirect comparisons, or head-to-head trials become available.

We also sought results on 24-hour sustained pain-free outcomes from placebo-controlled trials of reformulated and the conventional tablet form of sumatriptan, but insufficient data were available from trials of conventional sumatriptan.

Indirect comparisons of subcutaneous sumatriptan to oral formulations of other triptans

Sumatriptan is the only triptan approved in the United States and Canada in an injectable form. Given the lack of fair-quality or good-quality head-to-head trials involving subcutaneous sumatriptan 6 mg, we examined findings of a good-quality systematic review that qualitatively evaluated indirect comparisons between subcutaneous sumatriptan 6 mg and other triptans on the basis of unadjusted estimates of relative risk calculated for each triptan using pooled data from placebo-controlled trials.⁵² The main advantage of subcutaneous sumatriptan 6 mg over oral triptans is that it could potentially provide earlier pain relief. In 12 trials,^{86–89 90–96} pooled rates of 1-hour pain relief were significantly greater for subcutaneous sumatriptan 6 mg than placebo (70% compared with 22%), which resulted in the largest relative benefit estimate (3.2; 95% CI, 2.8 to 3.6) and a number needed to treat of 2.⁵² Benefits relative to placebo calculated for other triptans were lower, ranging from 1.6 (95% CI, 1.3 to 1.9) for oral the conventional tablet form of sumatriptan 100 mg to 2.3 (95% CI, 1.9 to 2.8) for eletriptan 40 mg.

Functional capacity, work productivity, and quality of life

Numerous fair-quality, placebo-controlled studies of subcutaneous sumatriptan reported on functional capacity, work productivity, and quality of life.^{86–90, 92–106} Subcutaneous sumatriptan consistently reduced time to return to work,^{86, 89, 90, 94–96, 103} degree of clinical disability,^{87, 88, 93, 98, 99, 102, 105, 106} and time to emergency room discharge⁹⁸ and improved quality of life-related symptoms (contentment and vitality dimensions of the Minor Symptom Evaluation Profile).¹⁰²

Indirect comparisons of frovatriptan to other oral triptans

Two-hour pain-free data from placebo-controlled trials were pooled and a combined risk difference for frovatriptan 2.5 mg and for the conventional tablet form of sumatriptan 100 mg were qualitatively compared. For the conventional tablet form of sumatriptan 100 mg, we conducted a risk difference meta-analysis of 8 placebo-controlled trials.^{36, 37, 45, 81–85} Compared with placebo (8%, 57/696), rates of 2-hour pain-free were 20% higher (95% CI, 0.16 to 0.25) for the conventional tablet form of sumatriptan 100 mg (30%, 437/1478), with a number needed to treat of 4. For frovatriptan 2.5 mg, we obtained the risk difference estimate for 2-hour pain-free rates from a good-quality systematic review that pooled data from 5 placebo-controlled trials involving a total of 2866 patients.¹⁰⁸ Results of their risk difference meta-analysis indicate that rates of 2-hour pain-free were only 9% higher (95% CI, 0.07 to 0.10; number needed to treat of 12) for frovatriptan 2.5 mg (12%) compared with placebo (3%), indicating frovatriptan is probably inferior to the conventional tablet form of sumatriptan 100 mg.

Early intervention

One fair-quality, placebo-controlled, crossover trial of frovatriptan 2.5 mg reported results from 137 adults who took study medication in the early stage of their migraine.¹⁰⁹ Rate of 2-hour pain-free was better with frovatriptan 2.5 mg than placebo (28% compared with 20%; P=0.04), with a relative risk of 1.40 (95% CI, 1.11 to 1.76) and a number needed to treat of 12. Results of the comparison between frovatriptan 2.5 mg and placebo for rate of 24-hour sustained pain-free were not reported.

Early intervention

Treximet^® is the most well-studied triptan for early treatment of mild migraine. The efficacy of Treximet^® (rapid-release sumatriptan RT 85 mg/naproxen 500 mg) administered early in a migraine while the pain is still mild has been demonstrated in 6 trials (GlaxoSmithKline Protocols TRX101998, TRX101999, TRX103632, TRX103635, TRX106571, and TRX106573), enrolling a total of over 2700 adults. Methods and results for 2 pairs of protocols (TRX101998 and TRX101999; TRX103632 and TRX103635) are fully published in 2 journal articles, respectively.^{116, 117} Methods and results for protocols TRX106571 and TRX106573 had not yet been published at the time of this report, but were accessed from the summary reports available on the manufacturer’s clinical trial registry website (http://www.gsk-clinicalstudyregister.com). Protocols TRX101998 and TRX101999 used parallel designs and were rated good quality. Protocols TRX106571 and TRX106573 used crossover designs to specifically evaluate efficacy and harms in adults with a history of poor response or intolerance to previous triptan treatment. Protocols TRX106571 and TRX106573 were rated fair-quality mainly because the summary report only provided combined results for both crossover periods, which did not appear to be assessed or adjusted for potential order effects. Protocols TRX103632 and TRX103635 used 4-period crossover designs to evaluate consistency across 3 attacks.¹¹⁷ Patients were randomized to 1 of 5 treatment sequences, 4 of which contained 1 interspersed placebo treatment period. One sequence that contained 4 consecutive treatment periods of Treximet^® was included for comparison in order to assess period effects and within-subject consistency. Results for protocols TRX103632 and TRX103635 were reported separately for the first period only and were rated good quality.

Patients in all 6 trials were instructed to take trial medication within 1 hour of migraine onset and while the pain remained mild. In all 6 trials, Treximet^® was superior to placebo on rates of 2-hour pain-free and 24-hour sustained pain-free. We calculated separate pooled relative risk estimates for the subgroup of 4 trials (TRX101998, TRX101999, TRX103632, TRX103635; N=1537) that enrolled patients regardless of their triptan treatment history and for the subgroup of 2 trials, which required prior poor response or intolerance (TRX106571 and TRX106573; N=535). For 2-hour pain-free outcomes, compared to the combined estimate of benefit from the 4 trials that enrolled patients regardless of their prior triptan treatment history (relative risk, 3.12; 95% CI, 2.64 to 3.69), the benefit of Treximet^® over placebo was somewhat smaller in the 2 trials which required prior poor response or intolerance to triptans (relative risk, 2.62; 95% CI; 1.92 to 3.58). For 24-hour sustained pain-free outcomes, however, compared with the combined estimate of benefit from the 4 trials of patients with an unspecified triptan treatment history (relative risk, 3.21; 95% CI, 2.63 to 3.91), the benefit of Treximet^® over placebo was somewhat larger in patients with a prior history of poor response or intolerance to triptans (relative risk 3.77, 95% CI, 2.38 to 5.99).

Protocols TRX103632 and TRX103635 also evaluated within-subject consistency of 2-hour pain-free and 24-hour sustained pain-free outcomes in 973 of 1135 (86%) patients who treated at least 3 attacks with Treximet^®.¹¹⁷ The rate of patients who were pain-free at 2 hours postdose in at least 2 of the first 3 attacks treated with Treximet^® was 52% to 55% across both trials. The rates of patients with a sustained pain-free response through 24 hours postdose in at least 2 of the first 3 attacks treated with Treximet^® ranged from 14% to 15% across the 2 trials. Subgroup analyses of the patients randomized to the sequence with no interspersed placebo treatment found similar rates of 2-hour pain-free and 24-hour sustained pain-free, which suggests against significant period effects. In patients randomized to the sequence that contained 4 consecutive treatment periods of Treximet^®, 21% (18/84) in TRX103635 and 28% (27/95) in TRX103632 had 2-hour pain-free outcomes in all 4 attacks.