Oral rizatriptan for acute migraine

Clinical bottom line:

Oral rizatriptan 10 mg has been tested in seven randomised trials with about 2700 patients. The NNT for two hour headache response was 2.7 (2.4 to 2.9). For pain free response at two hours compared with placebo, the NNT 3.1 (2.9 to 3.4) in 2770 patients. For sustained headache response the NNT was 5.6 (4.5 to 7.4) in 1677 patients in three trials.

Oral rizatriptan 5 mg has been tested in four randomised trials with about 1900 patients. The NNT for two hour headache response was 3.9 (3.3 to 4.7). For pain free response at two hours compared with placebo, the NNT 4.7 (4.0 to 5.7) in 1646 patients. For sustained headache response the NNT was 8.3 (6 to 14) in 1450 patients in three trials.

Reference

Oldman AD, Smith LA, McQuay HJ, Moore RA. Rizatriptan for acute migraine (Cochrane Review). In: The Cochrane Library, Issue 3, 2001. Oxford: Update Software.
AD Oldman, LA Smith, HJ McQuay, RA Moore. A systematic review of treatments for acute migraine. Pain 2002 97: 247-257.
MD Ferrari et al . Meta-analysis of rizatriptan efficacy in randomized controlled clinical trials. Cephalalgia 2001 21: 129-136.
MJ Gawel et al. A systematic review of the use of triptans in acute migraine. Canadian Journal of Neurological Science 2001 28: 30-41.

Date review completed: September 2001
Number of trials included: Seven randomised controlled trials for 10 mg rizatriptan and four for 5 mg rizatriptan
Bumber of patients: 2770 and 1646
Control group: Active and placebo
Main outcomes: Headache intensity and duration

Inclusion criteria were: treatment of acute migraine; randomised allocation to treatment groups; double-blind design; adult population and headache outcomes.

Search

Comprehensive searches of the following databases were made: MEDLINE (1966 - July 2001), EMBASE (1980 - June 2001), Cochrane Library (Issue 3, 2001), PubMed (to September 2001) and the Oxford Pain Relief Database (1950 - 1994). A series of free text searches were undertaken, using generic and trade names. There was no restriction to language. Retrieved reports were searched for additional trials. Neither individual authors nor pharmaceutical companies were contacted for unpublished data.

Findings

Tables with all included trials and excluded trials are available.

Oral rizatriptan 10 mg

The results for the seven trials for oral rizatriptan 10 mg were consistent for all outcomes, and those for two hour headache response are shown in Figure 1.

Figure 1: Headache response at two hours for oral rizatriptan 10 mg

Table 1: Results for oral rizatriptan 10 mg

		Number/total with outcome (%)
Outcome	Number of trials	Treatment	Placebo	NNT (95%CI)
Two hour headache response	7	1219/1783 (68	303/987 (31)	2.7 (2.4 to 2.9)
Two hour pain free	7	720/1783 (40)	83/987 (8)	3.1 (2.9 to 3.5)
Headache response at 2 hours sustained to 24 hours	3	420/1030 (41)	148/647 (23)	5.6 (4.5 to 7.4)

For oral rizatriptan 10 mg in seven trials, 1219/1783 patients (68%) had a headache response at two hours compared with 303/987 (31%) with placebo. The NNT was 2.7 (2.4 to 2.9). For pain free response at two hours compared with placebo, the NNT 3.1 (2.9 to 3.4) in 2770 patients. For sustained headache response the NNT was 5.6 (4.5 to 7.4) in 1677 patients in three trials.

Oral rizatriptan 5 mg

Figure 2: Headache response at two hours for oral rizatriptan 5 mg

Table 2: Results for oral rizatriptan 5 mg

		Number/total with outcome (%)
Outcome	Number of trials	Treatment	Placebo	NNT (95%CI)
Two hour headache response	4	548/933 (59)	234/713 (33)	3.9 (3.3 to 4.7)
Two hour pain free	4	284/933 (30)	65/713 (9)	4.7 (4.0 to 5.7)
Headache response at 2 hours sustained to 24 hours	3	280/803 (35)	148/647 (23)	8.3 (6 to 14)

For oral rizatriptan 5 mg in four trials, 548/933 patients (59%) had a headache response at two hours compared with 234/713 (33%) with placebo. The NNT was 3.9 (3.3 to 4.7). For pain free response at two hours compared with placebo, the NNT 4.7 (4.0 to 5.7) in 1646 patients. For sustained headache response the NNT was 8.3 (6 to 14) in 1450 patients in three trials.

Other outcomes

The Ferrari review also reports other outcomes.

For patients who were pain free at two hours, and then had no headache recurrence (moderate or severe) and no additional analgesics for 22 hours the NNT was 5.5 (4.9 to 6.4), occurring in 25% on rizatriptan 10 mg and 7% on placebo.
At two hours, the number needed to treat with rizatriptan 10 mg compared to placebo to achieve one patient who was not nauseated was 4.8.
At two hours, the number needed to treat with rizatriptan 10 mg compared with placebo to achieve one patient who did not have photophobia was 3.6.
At two hours, the number needed to treat with rizatriptan 10 mg compared with placebo to achieve one patient who did not have phonophobia was 3.8.
At two hours, the number needed to treat with rizatriptan 10 mg compared with placebo to achieve one patient who did not have functional disability was 4.0.

Adverse effects

Adverse effects are not reported in any way that makes obvious sense because efficacy was measured over 24 hours and adverse events for up to 10 days.

Comment

Four systematic reviews (one as a dual publication including a Cochrane review) for one treatment (rizatriptan 10 mg) is unusual. It is worth asking the obvious question of whether they all give the same result, especially two reviews were from published information, and one from a single-patient meta-analysis, and that all took different definitions of intention to treat. Table 3 shows the results for all three:

Table 3: Comparison of three reviews for rizatriptan 10 mg

Number needed to treat (95% confidence interval) for rizatriptan 10 mg
	At 2 hours		Over 24 hours
	Headache response	Pain free	Sustained response	Sustained pain free
Gawel et al	2.8 (2.6 to 3.2)	3.2 (2.9 to 3.5)
Ferrari et al	3.0 (2.8 to 3.4)	3.2 (3.0 to 3.5)	5.3 (4.6 to 6.2)	5.5 (4.9 to 6.4)
Oldman et al	2.7 (2.4 to 2.9)	3.1 (2.9 to 3.4)	5.6 (4.5 to 7.4)
Notes: Gawel review used only tablet data from 6 trials. Ferrari review used data from tablets and wafers in 7 trials, but had individual data from trial records. Oldman review combined tablet and wafer from 7 trials, but probably used a different definition of intention to treat as numbers of patients differed.

The results for two hour and sustained outcomes were essentially the same.