Demographic and headache characteristics at baseline

One hundred and forty-four patients were included in the study. One hundred and forty patients received at least one dose of erenumab and were concerned by effectiveness and safety assessment. One hundred and forty, 118, 107 and 102 patients were available for effectiveness and safety assessment at M3, M6, M9 and M12, respectively. Reasons for study discontinuation are reported in Fig. 1. Thirty-eight (27.1%) discontinued treatment during the 12-month follow-up, with 22 (15.7%), 11 (7.9%) and 5 (3.6%) patients before 3, 6 or 9 months of treatment respectively. The major cause of discontinuation was ineffectiveness (n = 31/38; 81.5%) followed by withdrawal because of patient’s choice (n = 4/38; 10.5%), adverse event (n = 2/38; 5.3%) and pregnancy (n = 1/38; 2.6%). The majority of patients who received at least one dose of erenumab and concerned by assessment were female (82.1%) with a mean age of 50.9 ± 11.4 (range: 19–75). Demographic and headache baseline characteristics of patients are reported in Table 1. Most of the patients (88.6%) suffered from CM. Most of the patients (90.7%) presented with a baseline MO which involved mainly nonsteroidal anti-inflammatory drugs (22.9% of the included population), weak opioids (25.7% of the included population) and triptans (71.4% of the included population). According to inclusion criteria of the compassionate erenumab use program, all patients had failed 11 oral preventive treatments and only 4 (2.9%) had continued oral prophylactic treatment.

Fig. 1

Flow-chart of patients from inclusion to study end with initial evaluation at baseline (M0) and evaluation at month 3 (M3), month 6 (M6), month 9 (M9) and month 12 (M12)

Table 1 Demographic and headache baseline characteristics of patients who received at least one dose of erenumab and concerned by effectiveness and safety assessment (n = 140)Unit dose throughout the compassionate erenumab use program

The repartition of erenumab unit dose (70 mg / 140 mg) was 114 / 26, 13 / 127, 8 / 110, 6 /101, 5 / 97 and 5 / 97 at M0, M1, M3, M6, M9 and M12 respectively (Fig. 1).

Primary efficacy endpoint

The proportion of ≥ 50% responders at M3, M6, M9 and M12 was 74/140 (52.9%), 69/118 (58.5%), 61/107 (57.0%) and 60/102 (58.8%) respectively (Fig. 2). Comparison of responder rates at M6, M9 and M12 with those at M3 showed no statistically significant difference (p = 0.505, p = 0.330 and p = 0.493 respectively). In univariate analysis, a positive association emerged between the 50% response rate at M3 and MMD (p < 0.001), MHD (p < 0.001) and EQ-5D (p = 0.025) at baseline (Table 2). At M6, a positive association appeared between the 50% response rate and age (p = 0.039), MHD (p = 0.011), MMD (p = 0.001) and psychiatric comorbidities (p = 0.012) (Table 2). Only psychiatric comorbidities (p = 0.002) were associated with the 50% response rate at M9 (Table 2). Medication overuse (p = 0.039), EQ-5D (p = 0.012) and psychiatric comorbidities (p = 0.002) were associated with the 50% response rate at M12 (Table 2). Multivariate logistic regression developed from the variables with p-value ≤ 0.05 in univariate analysis identified baseline variables predictive of therapeutic response only for M6. Three variables found to be associated to the 50% responder rate at M6 were a lower MHD (AdjOR = 0.93, 95%CI = [0.89; 0.99], p = 0.036), a lower age (AdjOR = 0.96, 95%CI = [0.93; 0.99], p = 0.045) and absence of psychiatric comorbidities (AdjOR = 2.61, 95%CI = [1.19; 5.87], p = 0.017) were associated to the 50% responder rate at M6 (Table 3).

Fig. 2

Reponder rates (≥ 30%, ≥ 50% as primary endopoint, ≥ 75%) et M3, M6, M9 and M12

Table 2 Univariate analysis of independent determinants of ≥ 50% response at M3, M6, M9 and M12Table 3 Logistic regression analysis of ≥ 50% reponse et M6Secondary efficacy endpoints

For MMDs, the median (IQR) was 18.0 (13.0–26.0), 9.0 (5.0–17.0), 7.5 (5.0–14.0), 8.0 (5.0–12.5) and 8.0 (5.0–12.0) at M0, M3, M6, M9 and M12, respectively (Fig. 3). For MHDs, the median (IQR) was 23.0 (16.0–30.0), 11.0 (6.0–22.3), 11.0 (6.0–19.0), 9.0 (6.0–16.0) and 9.0 (6.0–15.8) at M0, M3, M6, M9 and M12, respectively (Fig. 3). For HIT-6 score, the median (IQR) was 68.0 (63.8–73.3), 60.0 (54.0–65.0), 60.0 (50.3–53.0), 59.0 (50.0–63.0) and 58.0 (50.0–62.9) at M0, M3, M6, M9 and M12, respectively (Fig. 3). For MMDs, MHDs and HIT-6 score, paired t-test of M3, M6, M9 and M12 with M0 showed statistically significant differences (p < 0.001). For MMDs, MHDs and HIT-6 score, changes from baseline (M0) at M3, M6, M9 and M12 are presented in Table 4.

Fig. 3

Median with IQR of monthly migriane days (MMDs), monthly headache days (MHDs) and HIT6-score (HIT) at M0, M3, M6, M9 and M12

Table 4 Monthly migriane days (MMDs), monthly headache days (MHDs) and HIT6-score (HIT) at M0, M3, M6, M9 and M12 with change from baseline (BL or M0) at M3, M6, M9 and M12

The proportion of ≥ 75% responders at M3, M6, M9 and M12 was 22/140 (15.7%), 21/118 (17.8%), 25/107 (23.4%) and 24/102 (23.5%), respectively (Fig. 2). Comparison of 75% responder rates at M6, M9 and M12 with those at M3 showed no statistically significant difference (p = 1.000, p = 0.297 and p = 0.275 respectively). The proportion of ≥ 30% responders at M3, M6, M9 and M12 was 114/140 (81.4%), 95/118 (80.5%), 85/107 (79.4%) and 85/102 (83.3%) respectively (Fig. 2). Comparison of 30% responder rates at M6, M9 and M12 with those at M3 showed statistically significant differences (p = 0.001, p < 0.001 and p = 0.002 respectively).

The proportion of patients with CM was reduced at M3 (53/140, 37.9%), M6 (41/118, 34.7%), M9 (32/107, 29.9%) and M12 (25/102, 24.5%) compared with M0 (124/140, 88.6%) (p < 0.001) (Fig. 4). At M3, the rate of reversion from chronic migraine to episodic migraine was 57.3%. The proportion of patients with MO was reduced at M3 (68/140, 48.6%), M6 (51/118, 43.2%), M9 (43/107, 40.2%) and M12 (39/102, 38.2%) compared with M0 (127/140, 90.7%) (p < 0.001) (Fig. 4). At M3, the rate of transition from M0 to non-overuse was 46.5%

Fig. 4

Proportion of medication overuse (MO) (yes/no) and migraine form (episodic/chronic) at M0, M3, M6, M9 and M12

PGIC at M3, M6, M9 and M12 are presented in Fig. 5. Patients who reported considerable or major improvement were 78/140 (55.7%), 68/118 (57.6%), 62/107 (57.9%) and 67/102 (65.7%) at M3, M6, M9 and M12, respectively.

Fig. 5

Patient Global Iimppression of Change (PGIC) at M3, M6, M9 and M12

LOCF analysis

LOCF analysis showed similar results in terms of responder rates, MMDs, MHDs, HIT-6 score. Results are presented in Table 5.

Safety and tolerability assessment

Fifty-three (37.9%) patients reported at least one adverse event including: cutaneous erythema and/or pain at the injection site for 42 (30%) patients, constipation for 22 (15.7%) patients, muscle spasm for 2 (1.4%) patients, alopecia for one (0.7%) patient and blood pressure increase in one (0.7%) patient. There was no serious adverse event. In two (1.4%) patients, constipation led to discontinuation of treatment at M3 in one patient and at M6 in the other. The patient who experienced an increase in blood pressure did so after 3 months of erenumab use. This patient had a history of well-controlled HBP at the time of inclusion in the compassionate erenumab use program. This worsening of HBP was judged by the patient's referring cardiologist to be of moderate intensity and was rapidly brought under control by adapting the antihypertensive treatment. In view of these elements and the fact that the concerned patient had a 75% response rate, it was decided that the benefit/risk ratio was in favor of erenumab and treatment was continued. No worsening of HBP was observed during the additional 9 months of the compassionate erenumab use program for this patient.

One female patient became pregnant after 5 months of exposure to erenumab. Treatment was stopped as soon as the pregnancy was known. The pregnancy and delivery proceeded normally with no neonatal problems.