Abstract
Dear sir In their recent articles in the supplement on ‘Evidence-based Migraine Therapy’, Salonen (1) and Gawel and Wiebe (2) discuss data from two clinical trials by Tfelt-Hansen et al. (3) and Goldstein et al. (4) which directly compared rizatriptan with sumatriptan. We would like to respond to some of the issues they raise.
Outcome measures
A time-to-event analysis for pain relief or pain free within 2 h was the primary endpoint in each of the comparative studies. The analysis included patients who did and did not respond within 2 h. Salonen proposes a conditional analysis including only patients who responded. The issue of whether the analysis should be based on all patients or only those who responded has been discussed in a series of letters published in the journal Headache, to which the interested reader is referred (5–8). In our opinion, a time-to-event analysis cannot ignore patients who did not respond. Perhaps the most appropriate way of describing the results from this type of analysis is to say that, over the first 2 h after dosing, patients taking rizatriptan have a greater likelihood of earlier relief.
Clinical significance of statistically significant findings: This issue is addressed to some extent by data from patient preference studies as discussed in the next paragraph. We would also point out that in the two rizatriptan vs. sumatriptan studies, more patients on rizatriptan met the new IHS recommended primary efficacy endpoint of pain free at 2 h (9) (rizatriptan 10 mg vs. sumatriptan 100 mg—40% vs. 33%, P=0.019; rizatriptan 10 mg vs. sumatriptan 50 mg—41% vs. 37%, P=0.066). This endpoint was selected because of its high clinical relevance.
Patient preferences
We agree with Salonen that patient preference data may be an approach to comparing the triptans that provides a more real life perspective. However, we should clarify what is meant by ‘preference’ because, in their discussion, Gawel and Wiebe confuse patient satisfaction and patient preference. Patient satisfaction with medication at 2 h was rated on a 7-point scale in each study. When the percentage of patients who were ‘completely satisfied with medication’ or ‘very satisfied with medication’ (a stringent outcome criteria) were compared between groups, rizatriptan was superior to sumatriptan (rizatriptan 10 mg vs. sumatriptan 100 mg—33% vs. 26%, P=0.038; rizatriptan 10 mg vs. sumatriptan 50 mg—44% vs. 31%, P=0.003). Patient preferences for one treatment over another can only be assessed in situations where the same patient receives both treatments. This was not the case in the Tfelt-Hansen study. In the Goldstein et al. study (4), patients received both treatments but preference was not assessed. We have recently completed two preference studies (submitted for publication) which showed that more patients preferred rizatriptan 10 mg wafer to sumatriptan 50 mg tablets (57% vs. 43%, P=0.009, in Study 1; 64% vs. 36%, P< 0.001, in Study 2). Interestingly, the main reason patients gave for preferring one treatment over another was speed of action. This suggests that the advantage seen for rizatriptan in the time-to-event analysis in the randomized, blinded, comparative studies translates to real life clinical benefits for patients.
Potential bias from prior exposure to one of the study drugs: We agree with Salonen that if any bias were present it could operate in either direction. Ideally, comparative studies would include only those patients who were triptan naïve. However, this ideal will become increasingly impractical as the triptans gain widespread acceptance. The patient preference studies described above used patients who were naïve to both rizatriptan and sumatriptan.
Insufficient information regarding the number of patients randomized but not included in the analysis in the Tfelt-Hansen et al. study: 1268 patients were randomized and 1099 took study medication. The main reasons why 169 patients did not take study medication were that they did not experience a moderate or severe migraine (n = 103) or that they decided to withdraw from the study (n = 32).
We fully support an evidence-based approach to migraine therapy. In our opinion, the available evidence from direct comparative trials suggests that rizatriptan 10 mg is the most effective of the triptans and that this difference is clinically relevant (see also 10–14).