Published and not fully published double-blind, randomised, controlled trials with oral naratriptan in the treatment of migraine: a review based on the GSK Trial Register

A recent review on reporting bias in clinical trials [25] concluded that “the prevalence of incomplete outcome reporting is high”. In addition, when all randomised controlled trials (RCTs) are not published, it results in publication bias, which can be considerable and can distort available evidence [26].

Publication of well-conducted RCTs is primarily the responsibility of the clinical investigators. However, all the RCTs on naratriptan were multicenter trials (range 10–152 centers) and the pharmaceutical company, GSK was most likely fully in control of both the conduct and the publication of the RCTs. Seven pivotal RCTs [11, 14, 15, 17, 19, 21, 23] were fully published, but one was not [21] (see Table 1). The reasons for not publishing the 10 other RCT in full remains unknown.

Notably, the 2,729 patients included in the fully published RCTs represent less than half (45%) of the 6,112 patients participating in the naratriptan trial program. When migraine patients were recruited to an RCT, which may include placebo or a less effective drug, for up to 4 h, it should be an obligation to publish the results in a peer-reviewed journal. The conduct of RCTs solely for registration purposes, without full publication, should be avoided. The data presented in this review should be and are now in the public domain. They confirm, as expected, that naratriptan is superior to placebo as was also found in meta-analyses [1‐3]. It should be noted, however, that headache relief (a decrease in headache from moderate or severe to none or mild) was first measured after 4 h because naratriptan is generally held to be a slow-acting triptan [5]. The “slow onset of action” (a delay in onset) of naratriptan was apparently confirmed in one RCT (see Fig. 1) [21]. After 2 h, the headache relief for patients taking sumatriptan 100 mg and for naratriptan 2.5 mg were 59 and 50%, respectively. After 2 h, the curves for headache relief were parallel, with the endpoint headache relief at 4 h for sumatriptan 100 mg at 76% and naratriptan 2.5 mg at 66%. Thus, it is difficult to judge the speed of onset of action of the two drugs when the final endpoints are different [27]. In contrast, in an RCT using the higher 10-mg dose of naratriptan versus 100 mg sumatriptan, the endpoint of 4-h headache relief was similar, at 80%, for both groups [24]. As shown in Fig. 2, there is no difference in the “onset of action” when equipotent doses of the two drugs are used. Thus, naratriptan is not a slow-acting drug per se.

Naratriptan 1 mg b.i.d was superior to placebo for short-term prophylaxis of MAM in 3 RCTs, and naratriptan halved the MAM in one RCT [19], whereas in 2 RCTs 36% were without MAM versus 27% for placebo [12, 13]; these results suggest a modest effect of short-term prophylaxis with naratriptan in MAM.

Naratriptan was evaluated in 3 RCT versus the NSAID, naproxen sodium [8‐10] (see Table 1). In two RCTs, naratriptan was not superior to naproxen sodium 375 mg for migraine-related quality of life [8, 9]. In one RCT, more patients (34%) preferred naratriptan than naproxen sodium 500 mg (25%) [10]. It is unclear why naratriptan 2.5 mg is similar to naproxen 375 mg but better than the higher 500-mg dose of naproxen sodium. It may be due to the design of the studies.

In conclusion, the investigators are obligated to the migraine patients participating in the trials, who often endured placebo administration for 4 h, to ensure that the results of well-conducted, randomised trials are published in peer-reviewed journals. This can be difficult when there are multicenter trials with many investigators, and I suggest that investigators choose a publication committee among themselves, e.g. one from each country or each region.

In this review, all double-blind RCTs with oral naratriptan in the treatment of migraine are presented [8‐24]. This does not change the overall picture of an effective and well-tolerated triptan [3]. Whereas naratriptan is superior to placebo, the 2.5-mg dose chosen is less effective than sumatriptan 100 mg (2 and 4 h) (see Table 1), and rizatriptan 10 mg (2 h) [28]. The inferiority of naratriptan versus most other triptans at 2 h has been shown in meta-analyses [1, 2, 4], and this has been ascribed to a slow action of naratriptan [5]. The apparent slow onset of naratriptan 2.5 mg is, however, a matter of dose (see above).

When naratriptan 2.5 mg became an OTC drug, the question of its efficacy compared with other OTC drugs and NSAIDs became relevant. Naratriptan was not compared with paracetamol or aspirin, which in an effervescent form (52% headache relief) was found to have similar efficacy as sumatriptan 50 mg (46% headache relief) in a meta-analysis [28]. There are 2 RCTs comparing naratriptan 2.5 mg with naproxen sodium 275 mg (Table 1). Unfortunately, no placebo-control was used in the RCTs, but no differences between the active drugs were observed. Patients’ preference for naratriptan (34%) was marginally superior to naproxen sodium (26%) (Table 1) in patients dissatisfied with simple analgesics [10]. What is missing is a comparative RCT of naratriptan 2.5 mg (50% headache relief Table 1), and lysine acetylsalicylate and metoclopramide, which showed headache relief of 56% in 2 RCTs [29, 30]. Until such comparative RCTs become available, one cannot, based on the available evidence, recommend what migraine patients should try next when they are dissatisfied with simple analgesics. Naratriptan is a reasonable OTC choice as second or third choice on the step care ladder.