Main result of the present study is that in the present sample of migraine patients, 42.5% had failed ≥ 1 triptan, 13.1% had failed ≥ 2 triptans and 3.9% had failed ≥ 3 triptans. Subtracting patients who found another effective and tolerable acute medication (e.g. another triptan or non-opioid analgesic) reduced these numbers to 21.5%, 7.1% and 2.3%, respectively. It further resulted that migraine patients who failed one or several triptans were significantly more severely affected by their migraine than triptan users without non-response. On a single substance/formulation level, largest overall responder rates were found for nasal and oral zolmitriptan, eletriptan and subcutaneous sumatriptan. It is important to note that the definition of response used here requires the triptan to be both effective and tolerable.

Present data show that 42.5% of the patients had failed ≥ 1 triptan and 13.1% had failed ≥ 2 triptans. It must be considered that these percentages are with respect to the total population, that includes 29.7% of patients never having tried a triptan. These results are slightly higher compared to a pooled analysis of rimegepant studies that showed insufficient response to ≥ 1 triptan and to ≥ 2 triptans in 35.2% and 9.3% of the study population, respectively [8]. It has been emphasized that failure of one triptan does not mean failure of every triptan [1], and the EHF definition accounts for this fact by demanding non-response to at least two triptans for ‘triptan resistance’. In view of the availability of new and effective oral drugs, requiring failure of 2 triptans before switching to a new drug class such as ditans or gepants (as proposed by the EHF [10]) seems reasonable.

However, in the present study, a significant number of patients with failure of two triptans responded to another triptan. Patients having tried 3 triptans and failed 2 of them had a 38.9% probability to respond to the third. However, it has to be noted that only 49% of the patients with failure of 2 triptans even tried a third triptan, and there might be a reason for this (e.g. partial response might prompt additional trials) so that the real response rates to a third triptan after failure of 2 might be lower. Nonetheless, the overall percentage of patients having failed 2 triptans (13.1%) was reduced to 10.4% when subtracting those who found an effective triptan during additional trials. Therefore, trying another triptan, especially one with a high response rate, can be an alternative to switching drug class for part of the patients. In the present study, highest response rates were found for zolmitriptan (oral and nasal), eletriptan (oral) and sumatriptan (subcutaneous) (see below). Obviously, it is also important to optimize treatment with a specific triptan (use early in the attack, appropriate dosing, treatment of several attacks) before declaring failure of this triptan [1, 2]. Finally, part of the patients with failure of 2 triptans respond to non-opioid or combination analgesics, so making sure that an adequate trial with these substances has been made is important.

Although it has the largest efficacy according to clinical studies [2], and is also among the triptans with the highest responder rates within the present study, requiring a trial with subcutaneous sumatriptan before switching to a novel oral drug class does not seem reasonable. Many patients prefer oral medication, and subcutaneous sumatriptan can be more expensive than the novel drugs. E.g., in Germany, at the time of this publication, a sumatriptan s.c. 6 mg dose is 12.5× the cost of an oral triptan dose, and 1.6× the cost of a lasmiditan 100 mg dose. Gepants are not yet available in Germany. Indeed, present data show that subcutaneous sumatriptan was prescribed only in a minority of patients (3.1% total and 13.9% of those having tried ≥ 3 triptans).

Clearly, the present results have to be interpreted in the context of specialized headache care, as DMKG Headache Registry data stem from headache centers and practices with a special interest in headache. This is also reflected by the large proportion of chronic migraine diagnoses (25.7%), the high average number of headache days per months (12.3 ± 8.2) and MIDAS score (39.6 ± 47.1; a score above 20 indicates severe migraine-related disability [14]). As failure of acute medication is one reason for referring patients to secondary/tertiary headache care, proportions of triptan resistant patients likely are smaller in primary care.

Present data (Supplementary Fig. 1) also show that a small proportion of patients switch to a second or even third triptan without failure of the first triptan. This might be due to economic considerations, availability, physician or patient preferences, the hope to achieve even better efficacy and tolerability, or use of more than one effective and tolerable triptan (e.g. an oral triptan for most attacks and a parenteral formulation for escalation therapy).

Previous studies have reported that triptan non-responders have more severe migraine (higher frequency and intensity, more accompanying symptoms, higher disability) than triptan responders [1, 15], and a previous DMKG Registry analysis showed that migraine patients with higher headache frequency have lower acute medication efficacy [16]. The present data expand these results, showing that migraine severity (headache and severe headache frequency, a chronic migraine diagnosis, headache intensity) and associated disability (MIDAS) further increased with increasing level of triptan failure (Fig. 2). While headache frequency increased by a factor of 1.4 from the ‘current use, no failure’ group to the ‘failure of ≥ 3 triptans’ group, MIDAS scores increased by 2.4. Thus, triptan failure disproportionately increased migraine-related disability, likely because insufficiently treated migraine attacks cause more disability than sufficiently treated attacks. In addition, insufficient acute migraine treatment increases the risk for migraine chronification [17]. Together, these results show that patients with failure of one or several triptans need our special attention. Acute treatment optimization strategies may include: treating early during the attack, using a different dose or formulation, switching to a different triptan, combining triptans with non-steroidal anti-inflammatory drugs (NSAIDs, e.g. naproxen) or switching to a different class of acute medication [1, 2]. These patients also need close follow-up to further adjust treatment if necessary. In addition, starting a migraine preventive medication may improve efficacy of the acute medication [11, 18]. In the present study, the use of migraine preventive medication was high, and even higher in patients with non-response to ≥ 2 triptans (Fig. 2). This shows that this strategy was often used in the present setting of specialized migraine care, but that the migraine burden of these patients nonetheless remained high. In addition, biobehavioural migraine preventive treatments (e.g. relaxation techniques, physical activity, biofeedback) should be implemented in all patients needing a migraine preventive medication. In the present data, triptan non-responders also were slightly younger than patients without triptan failure. It seems unlikely that this is a sign of better triptan response with increasing age. A possible explanation might be that patients with triptan non-response are referred to specialized care earlier than patients with a good response.

Consistent with German health insurance data [3], oral formulations of sumatriptan, rizatriptan, naratriptan and zolmitriptan were most frequently used as the first triptan (Fig. 3). In patients trying a third triptan, the pattern shifted towards rizatriptan, eletriptan, zolmitriptan (oral) and naratriptan. Use of rizatriptan or eletriptan after first triptan failure has been reported before [8]. As rizatriptan and eletriptan according to clinical studies are the most effective oral triptans and naratriptan is among the most tolerable triptans, this indicates reasonable choice of treatments [19, 20].

Regarding responder rates and reasons for non-response (Fig. 4), our results are generally consistent with the results of clinical trials [2, 19, 20]. Several points merit further discussion. Oral sumatriptan, the most frequently used triptan, showed a relatively low proportion of responders and a high proportion of tolerability failures. In clinical studies, sumatriptan ranged among the triptans with medium efficacy, which may partly explain these results [19]. Alternatively, failure of sumatriptan, the most frequently used triptan in Germany, may prompt referral to specialized headache care, resulting in patients with a poor response to sumatriptan being overrepresented in the present study population. Second, nasal sumatriptan had the largest number of efficacy failures. This result might not have been expected from early clinical trial data [21] but later studies indeed showed limited efficacy [22]. It might also be worth mentioning that some oral triptans (eletriptan and oral zolmitriptan) achieved responder rates very similar to the strong and fast acting parenteral triptans (nasal zolmitriptan and subcutaneous sumatriptan). It has been emphasized before that patient efficacy ratings comprise more than 2 h pain-free rates [23]. In addition, response in the present study encompassed both efficacy and tolerability. Finally, > 50% of patients discontinued frovatriptan for “other” reasons, likely because frovatriptan is expensive and (different from other triptans) only partially reimbursed by German health care. In addition, there have been availability issues. This makes frovatriptan data difficult to interpret.

Determination of triptan failure under real world conditions might easily be biased in both directions. A specialized headache care population (as represented by the DMKG Headache Registry) likely has higher rates of triptan resistance, both because triptan resistance might be more frequent in this population and because patients have higher odds to be offered several triptans in the first place. There could also be a bias towards patients having failed the most frequently used triptans (sumatriptan, rizatriptan, naratriptan) that might have affected the response rates of these substances. In contrast, true rates of failure of ≥ 2 triptans likely would be underestimated in a primary care setting where some patients may not be offered a second triptan after failure of the first. Even in the population analyzed here, 19.7% of the patients had failed their first triptan and had not (yet) tried a second one. The number of visits within the DMKG Headache Registry varied between 1 and 14 (2.7 ± 2.1), therefore not all patients had already undergone acute treatment optimization in specialized care. Thus, the present results have to be taken as a snapshot giving a coarse estimation of how frequent triptan resistance is in a specialized headache care setting. Second, it must be considered that triptan failure in the present study was based on retrospective patient self-report. Especially, past medications used before the first visit within the registry were assessed retrospectively at the first visit, possibly leading to recall bias and to incomplete reporting. Also, there is no way to know if patients observed the rules of early dosing and treating several attacks before declaring failure of a specific triptan. Regarding dosages, our previous results show that triptan underdosing is very rare in the present sample [16]. Nonetheless, choice of a low vs. high dose within the range of recommended dosages might affect triptan efficacy and tolerability. This was not analyzed here because numbers in some of the dosage subgroups were too small.