Sustained response to onabotulinumtoxin A in patients with chronic migraine: real-life data

Headache disorders are largely prevalent in the general population and constitute an important health concern [1], not only in developed but also in developing countries [2]. Among those disorders, chronic migraine (CM) is responsible for the highest levels of disability. CM is defined as the occurrence of ≥15 monthly headache days, of which ≥8 monthly migraine days, for ≥3 months [3]. The abovementioned definition remains valid even if some authors have recently proposed to consider ≥8 monthly migraine days for ≥3 months as sufficient to diagnose CM [4]. CM has a prevalence in the general population of about 2% [5] and imposes a significant burden on the patients in terms of social exclusion, isolation, anxiety, and depression, significantly lowering their quality of life [6, 7]. Onabotulinumtoxin A (BT-A) is a safe and effective preventive treatment for CM, as shown in the Phase 3 Research Evaluating Migraine Prophylaxis Therapy (PREEMPT) trials [8‐12]. Several real-life studies confirmed its safety and efficacy in clinical practice [13‐22] and showed that a shorter disease duration, some characteristics of headache (ocular, imploding), allodynia, and the absence of medication overuse and depressive symptoms are predictors of clinical response [23‐33]. However, it is common experience to observe fluctuations in the clinical response to BT-A as to other preventive treatments. Assessing sustained response to BT-A is important to reliably quantify the benefit of the treatment on the patients’ disability and quality of life. It is especially important nowadays, because alternative treatments are available for CM, namely monoclonal antibodies acting on the calcitonin gene-related peptide (CGRP) or its receptor [34, 35].

The present study aimed to assess proportion of patients who achieve sustained response to BT-A and to establish baseline headache characteristics which predict sustained response.

During the study period, 115 consecutive patients were treated with BT-A and provided consent to be followed up for 15 months. Among them, 87 (75.7%) completed 5 BT-A treatment cycles; 9 (7.8%) refused to continue BT-A after the third or fourth treatment cycle for clinical benefit (all had > 75% reduction in monthly headache days); 19 additional patients stopped BT-A because of treatment failure (6; 5.2%), adverse events (2 with allergic reaction; 1.7%), or lack of adherence (11; 9.6%), (Fig. 1). All patients discontinuing treatment filled out diaries and questionnaires until the end of the study; no other patient was lost to follow-up. Patients who stopped BT-A for clinical benefit did not restart BT-A by the end of the study, nor started any other preventive treatment.

The baseline characteristics of the included patients are reported in Table 1. Most patients (84.3%) were women, while the median age was 50 years (interquartile range [IQR] 44.5–54). Eighty-nine (77.4%) patients had medication overuse and 61 (53.0%) were on concurrent preventive treatments. The median migraine duration was 30 years, while the median CM duration was 62.5 months.

Headache days and medication days both decreased from a median of 30 (IQR 25–30) at baseline to 15 (IQR 7–25; P < 0.001) (Table 2). The median headache intensity significantly decreased from 8 (IQR 7–9) to 5 (IQR 4–7; P < 0.001). The patients’ quality of life also improved, as suggested by the MIDAS score decrease from 87.5 (IQR 42.5–123.5) to 12 (IQR 3.5–51.5; P < 0.001) and by the HIT-6 score decrease from 65 (IQR 60–69) to 62 (IQR 56–65; P < 0.001).

The proportion of responders per cycle of treatment ranged from 18.3% to 41.7% over the study period (Fig. 2). At the end of follow-up, 66 patients (57.4%) patients were classified as anytime responders and the remaining 49 patients (42.6%) were classified as non-responders. Fifty-one (44.3%) patients achieved response to treatment within the third cycle; 33 (64.7%) of them were classified as sustained responders while 18 (35.3%) had fluctuations in treatment responses over the following cycles (Fig. 3).

Figure 4 shows the clinical data stratified according to response status. Not only responders, but also non-responders had a significant decrease in the number of headache days and in headache intensity as compared to baseline. The disability (MIDAS) and impact (HIT-6) of headache improved numerically but not significantly after treatment in anytime and sustained responders (Fig. 4).

Sustained responders had a lower CM duration (31 vs 49 months; P = 0.030) and a lower median number of monthly headache days at baseline (25 vs 30; P = 0.013) compared with non-responders (Table 3) while the baseline headache characteristics and changes in concurrent oral preventive treatments were similar in the two groups.

At the end of follow-up, 40 of the 115 included patients (34.8%) had mild adverse events; the most common were local tension (n = 14; 12.2%), local pain (n = 12; 10.4%), and muscle weakness or atrophy (n = 11; 9.6%).

Migraine frequency and intensity, as well as response to treatments, undergo fluctuations as a response to internal and external triggering factors; the fluctuation between satisfactory and unsatisfactory responses to treatment might be frustrating for patients with CM. A way to improve the management of patients with migraine is to evaluate not only the response to treatments at a given time point, but also sustained benefit over time. To our knowledge, our study is the first to address the concept of ‘sustained response’ to BT-A.

We found that more than half of patients had a ≥ 50% reduction in headache days from baseline after at least one cycle of BT-A treatment and that about two thirds of patients who achieved a ≥ 50% reduction in headache days within the third treatment cycle had a sustained response to BT-A. Our results should be read in the light of a difficult-to-treat study population of two tertiary level headache centers, who had a high mean age, a long history of migraine and of CM, and a high prevalence of medication overuse. The proportion of patients with sustained response to BT-A treatment would likely be higher in less difficult-to-treat populations.

In our real-life setting, patients who had higher likelihood of sustained response were those with a shorter CM duration and lower burden of headache days. No other studies have evaluated predictors of sustained response and we can compare our findings only with those studies who have evaluated anytime response [23‐33] (Table 4). Anyhow, our data suggest that some factors which predicted short-term response, including short CM duration, are predictors of sustained response [30, 32] (Table 4). The comparable studies assessing the response to BT-A had heterogeneous treatment protocols, follow-up durations, and patient populations (Table 4). Three studies found that shorter migraine duration predicted a favorable response to BT-A [23, 30, 32], in line with our finding regarding CM duration. Notably, a favorable response to BT-A was not predicted by the overall duration of migraine, but rather by the duration of its chronic stage. Hence, the detection of migraine chronification is crucial to provide proper access to specialist care and the most effective options for CM treatment as early as possible. Other studies found that ocular or imploding headache predicted response to BT-A [24, 26, 27, 29], while in our study the pain location did not influence BT-A response. Medication overuse predicted unfavorable response to BT-A in one study [33], which contrasts with our findings; however, a further study supports our finding of no difference in response between patients with and without medication overuse [40]. The optimal management of CM associated with medication overuse is in fact controversial. Medication overuse is not a contraindication to BT-A treatment for CM and does not preclude its effectiveness [39, 41]. However, detoxication prior to BT-A initiation may per se determine reversion of CM and may improve the efficacy [42]. In our study population, detoxication prior to BT-A initiation was not mandatory; however, we cannot exclude that an extensive use of detoxication might increase the proportion of anytime and sustained responders.

The present study has several strengths. The study sample included patients with CM from two tertiary headache centers with experienced personnel. Besides, patients received a homogeneous treatment protocol whose response was assessed using standardized criteria. However, the present study is limited by the lack of assessment of migraine comorbidities, especially the psychiatric ones, which negatively affect the course of migraine [43] and might have a negative impact also on BT-A treatment [33]. Data on psychological treatments, including behavioral therapy [44], which might have benefited some patients were not collected in this study. Besides, changes in oral preventive treatments during treatment with BT-A in some patients might have limited the homogeneity of the study population; however, this limitation is common to all real-life studies, in which the efficacy of a drug is assessed on top of the best practice. Changes in concurrent oral preventive treatments did not differ according to response status in our population (Table 3); however, we cannot exclude that non-significant results were due to small numbers. As a further limitation, we could not assess whether BT-A was more effective on migraine than on headache days, as we did not distinguish between them. Prior response to acute medications and especially triptans was not explored, such as the information regarding the imploding or exploding nature of headache, thus limiting the comparability of the present study data with those of other studies [24, 26]. We chose a 15-month follow-up to align with the 5 treatment cycles of the PREEMPT protocol; however, longer follow-up periods are needed to assess the possible long-term benefits of BT-A in sustained responders. Lastly, we chose a cutoff of ≥50% reduction in headache days to identify responders; a relevant proportion of patients might have experienced a clinical benefit from BT-A treatment even if not fulfilling the criteria to be defined as a ‘responder’ in our study (Fig. 2). Previous data suggest that even a ≥ 30% reduction in headache days might be considered relevant to patients [13], while such a reduction identified patients as ‘non-responders’ according to our study criteria. Further studies are needed to find a better definition of ‘response’ to BT-A treatment, which takes into account the patients’ quality of life.

According to our data, about two thirds of patients who achieve a clinically relevant response to BT-A within the third cycle of treatment maintain the response over time. More recent onset of CM and more headache-free days at baseline are associated with sustained response. Therefore, we suggest not to delay preventive treatment of CM with BT-A, in order to increase the likelihood of sustained clinical response.