All patients that started treatment with erenumab in the Leiden Headache Center, a national referral centre, were invited to participate. They were all diagnosed with migraine, episodic or chronic, with or without aura, by a neurology resident in consultation with a neurologist with headache expertise or by a neurologist with a headache expertise, according to the ICHD-3 criteria . None of the patients had a second primary headache disorder. Only tension type headache was allowed, as this is common in patients with chronic migraine . Given the restricted availability of erenumab, all patients had at least 8 migraine days per month, and failed on at least 4 migraine prophylactics (meaning being ineffective, discontinued because of side effects or being contraindicated), including at least a betablocker, candesartan, valproate and topiramate. None of the patients had medication overuse headache.
Approval for this study was obtained from the LUMC Medical Ethical Committee and all participants gave written informed consent.
Patients were treated with erenumab 70 mg, administered subcutaneously once every four weeks. No additional prophylactic treatment was used.
The clinical response to erenumab was monitored using a validated daily headache e-diary [11, 15, 16]. This diary contains questions on the presence of headache, headache characteristics, accompanying symptoms and the use of acute migraine medication. In case of a headache, an automated algorithm based on the ICHD-3 criteria determined whether it was a migraine day. Additionally, days on which a triptan was taken, as well as aura without headache symptoms, were also counted as migraine days. Patients started this diary at least 4 weeks before starting treatment (the baseline period). In line with clinical trials , the clinical response was assessed by comparing MMD in week 9–12 (i.e. after three doses of erenumab) to that in the 4 week pre-treatment baseline observation period. A month is defined as 28 days (4 weeks).
Serum CGRP assays
Patients were invited to the hospital before starting treatment with erenumab (T0) and 2–4 weeks (after Tmax, but before the second dosing) after starting treatment with erenumab (T1). At both time points blood samples were collected from the antecubital vein, while subjects rested in a sitting position. The blood was then allowed to clot and was centrifuged at room temperature for 20 min at 622 g/2000 rpm to separate serum. Samples were then immediately stored at -80 °C in aliquots of 500 µL until analyzed.
For radioimmunoassay (RIA), a commercial kit (CGRP (Human) - RIA Kit (Phoenix pharmaceuticals, Burlingame, California, United States), detection range 0.53–660 pmol/l), was used following manufacturers’ instructions to measure CGRP-like immunoreactivity (CGRP-LI) levels. Biochemical assays were performed by an experienced lab technician who was blinded to the patient identity, study day and treatment effect of erenumab. All samples were analyzed in the same laboratory, under the same environmental conditions, and using the same batch for samples from different patients and different study days, to avoid a possible batch effect. Samples with values outside the detection range were set on the limits of the detection range.
Sample size was based on the available data. Baseline characteristics, including, sex, age, headache diagnosis and baseline headache measures were summarized using means and standard deviations or frequencies and proportions. For each patient the clinical response to erenumab was determined by calculating the absolute reduction in migraine days in the third month (week 9–12) after initiating treatment compared to the baseline month (4 weeks before starting treatment).
As serum CGRP-LI levels were highly skewed, a log transformation was applied, and these log-transformed values of CGRP-LI levels were used in all statistical analyses. However, for the sake of clarity, in the result section CGRP-LI levels are presented, without log transformation, as medians with interquartile ranges. To relate our CGRP measurements to measurements performed earlier by others, CGRP-LI levels at T0 were related to age and sex, with a Pearson correlation and an independent t-test, respectively. Comparisons between T0 and T1 were made using a repeated measurements model, with absolute reduction in monthly migraine days added as a covariate to assess the relation between change in serum CGRP-LI and change in migraine frequency.
To investigate the predictive value of serum CGRP-LI for the clinical response, two linear regression models were made with absolute migraine reduction as the outcome variable and with sex, age, migraine days at baseline as covariates. In our primary analysis, log serum CGRP-LI at T0, and in our secondary analysis log serum CGRP-LI at T1, was added as an independent variable.
In all analyses a two-sided p-value < 0.05 was considered to indicate significant differences. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 25 (IBM Corp., Armonk, N.Y., USA).