Background
Migraine is a common and disabling neurological disorder characterized by episodic attacks of headache and associated symptoms. When the migraine progresses to a chronic form (chronic migraine), the frequency of migraine attack increases, and head pain can persist even between attacks. The strategy of treatment differs between episodic migraine (EM) and chronic migraine (CM).
To date, the diagnosis of migraine is based on the patients’ description of symptoms. Although the International Headache Society offers well-structured diagnostic criteria for migraine and its subtypes [
1], it is often challenged in the clinic by the language barrier (e.g. deafness or cognitively impaired patient), recall bias, and instability of patient-reported headache frequency. Therefore, researchers have been seeking a biomarker which can aid the diagnosis and follow-up of migraine. Calcitonin gene-related peptide (CGRP) is one of the most promising candidates, since interictal serum CGRP concentration was reported as a possible biomarker of CM [
2]. However, the diagnostic role of CGRP has not been validated yet. In this study, we aimed to reproduce the previous study results and validate with clinical data from normal subjects, patients with EM, and those with CM.
Methods
Subjects
We prospectively recruited 156 adult patients with migraine (106 episodic and 50 chronic migraine) in the Samsung Medical Center headache clinic from August 2015 to May 2016. The diagnosis of migraine was based on the International Classification of Headache Disorders 3rd edition beta version (ICHD-3 beta). The distinction between episodic migraine (EM) and chronic migraine (CM) was also based on the ICHD-3 beta. We included patients of > 1 year after migraine onset. Twenty-seven normal controls (NCs) were also recruited for this study. Subjects were considered as NC when they had no subjective headache, then investigators confirm that they did not have migraine, any headache of moderate or severe intensity, or any acute or chronic pain disorder and did not take any regular medications. This study was approved by the institutional review board of Samsung Medical Center. All the participants gave written consent.
Evaluation
All patients completed a structured questionnaire regarding headache characteristics, frequency, past medical history, and the use of acute and preventive medications. From the questionnaire, the presence of unilateral autonomic symptoms (UAS) and headache unilaterality were identified as clinical markers of trigeminal activation [
3,
4]. Two headache neurologists (M.J.L. and C.-S.C.) interviewed all patients. We used Allodynia Symptom Checklist-12 (ASC-12) to estimate allodynia during migraine attack. CGRP was followed up after 3 months in patients with CM who underwent Botulinum toxin treatment. Selected patients with EM also underwent the follow-up measurement.
Blood collection
Blood sampling was conducted in our central laboratory between 8 and 10 a.m. after overnight fasting. A serum separator tube was used for sampling. After clotting at room temperature for 30 min, samples were centrifuged for 15 min at approximately 2000×g. Aliquots were stored immediately at − 80 °C. At the day of sampling, all patients and NCs were asked about the presence of headache since the past two days. When present, we collected information about the presence and characteristics of headache at the day (day 0) and the day before (day − 1) the sampling. It was considered interictal if patients did not have moderate or severe headache at both day − 1 and day 0 for EM patients and at day 0 for CM patients. Patients who took acute abortive medication at day − 1 or day 0 were excluded from the study, regardless of the severity of headache.
Serum CGRP measurement
All serum CGRP concentration was measured by an experienced laboratory technician who was blinded to clinical information or group assignment, using a commercially available ELISA kit (Wuhan USCN Business Co., Ltd., Hubei, China) based on the manufacturer’s instructions. The principle of assay kit is the competitive inhibition enzyme immunoassay. The biotin labeled CGRP was added to reaction well with unlabeled CGRP from patient’s serum, incubated and measured by binding of avidin conjugated to horseradish peroxidase. Five-point standard curve was generated with serially diluted standards and the concentration of CGRP in the sample was derived from it, which was reverse proportional to the intensity of final reaction. In-house prepared quality control sample was included at every batch of test. Analyses of between-run precision of control 1 and control 2 showed coefficients of variation of 13.1% and 11.2%, respectively. The detection range of kit was 12.35–1000 pg/mL.
Statistical analysis
Data are presented as mean (SD) or number (%) unless otherwise specified. The student t-test or Mann-Whitney test was used depending on the distribution of continuous variables. The Chi-Square test or Fisher’s exact test was performed to compare categorical variables. The relationship between CGRP concentration and clinical characteristics including the presence of UAS and headache unilaterality was tested by using the linear regression analysis. All data analyses were performed using Stata (version 14). P values less than 0.05 were considered significant.
Discussion
In this study, we found no increase in serum CGRP concentration in patients with CM. It was not associated with headache days, allodynia severity, or the presence of headache on the day of measurement.
A biomarker is defined as an indicator of normal biological processes, pathogenic processes or pharmacological responses to a therapeutic intervention [
5]. Clinically, a good biomarker should aid the diagnosis of the disease, correlate with disease severity, or predict outcomes [
6]. It also should be easy to measure with acceptable inter-rater and intra-subject reliability. Finally, a good biomarker is linked to pathophysiological explanation [
5,
6].
To search for a biomarker of migraine, several candidates have been tested [
6,
7]. Based on an early finding that jugular venous CGRP level is elevated during acute migraine attack [
8], CGRP has been regarded as a key neuropeptide of migraine pathophysiology [
9,
10]. Indeed, increasing evidences on the role of CGRP in migraine headache exist. CGRP-containing neurons are most frequently found in the human trigeminal ganglion [
11]. CGRP antagonists and monoclonal antibodies to CGRP or its receptor have shown a good efficacy to prevent migraine attacks [
12‐
14]. Based on these results, serum CGRP concentration is one of the most attractive candidates of biomarkers of migraine. Recently, a possible role of interictal serum CGRP measurement in the diagnosis of CM and prediction of treatment outcome has been suggested by researchers [
2,
15,
16].
Our study results, however, do not support CGRP as a biomarker of CM. Serum CGRP concentration was not diagnostic for CM, did not correlate with disease severity (headache frequency or allodynia), and did not predict the treatment outcome. CGRP concentrations did not differ according to migraine subtype (migraine with vs. without aura) or comorbidities such as fibromyalgia and medication overuse. Clinical markers of trigeminovascular activation were not associated with increased CGRP concentration. CGRP changed significantly over time, but it did not correlate with disease course. Based on our study results, CGRP might be neither a static biomarker in determining disease condition, nor dynamic biomarker which can reflect the disease severity. In addition, serum CGRP concentration may be prone to inter-subject fluctuations, without regard to treatment response.
Biomarker studies using CGRP has been challenged because of its short half-life in venous blood. In earlier studies using plasma samples, investigators made a supreme effort to reduce the time from sampling to freezing [
8,
17]. However, Cernuda-Morollón et al. reported a promising result using serum samples which require a relatively long time to clot at room temperature [
2]. We followed their methods and the same manufacturer’s instruction with theirs. Our analysis of between-run precision showed acceptable variations to exclude batch effects. Role of CGRP as a biomarker of chronic migraine should be re-appraised after a critical review of detection methods.
Technical factors might have attributed to the discrepancy between our study results and the previous study results. In the study by Cernuda-Morollón et al. [
2], CGRP concentrations were low in the NC and EM groups, while subjects with a high (> 100 pg/mL) CGRP level were present only in the CM group. In contrast, such a high CGRP concentration was detected in all NC, EM, and CM groups in our study. Technically, a batch effect should be considered if different batches were used for different groups. Different lots of ELISA kits, environments of experiments, or personnel who performed the experiment can affect the result. For example, when a researcher collects and analyzes blood samples of patients prior to the recruitment of matched control subjects, the between-group difference may be affected by the order of experiment because the two experiments can be different at least theoretically in terms of lot numbers of kits, time delay from the sampling and experiment, timing of experiments, or even temperature or humidity in the laboratory. In our study, we analyzed blood samples from two or more groups in each batch and repeated experiment using some samples of previous batch when we started a new batch.
Differences in demographics and characteristics should be also considered. Our CM patients have less fibromyalgia and more medication overuse than patients included in the study of Cernuda-Morollón et al., but these two comorbidities were not associated with serum CGRP levels in both studies [
2]. According to Cernuda-Morollón et al. [
2], migraine with aura was associated with higher serum CGRP concentration in women with CM. While nearly half of CM patients had aura in their study, the prevalence of migraine with aura was less than one fifth in our CM patients. This is not surprising because Asians have less prevalence of migraine with aura, although the prevalence of migraine is overall similar across countries [
18‐
20]. This might explain the inconsistency in part between our and their study results. However, the association of migraine with aura and CGRP concentration was not reproduced in our study. Further validation studies are still warranted to further reproduce the association between migraine diagnosis, migraine subtype, and interictal CGRP concentrations before it can be implemented as a diagnostic procedure in clinics.
Taken together, our data raise questions regarding the validity of serum CGRP testing for the diagnosis of CM. In addition to clinical feasibility, fundamental questions also remain unanswered: whether the trigeminovascular system is persistently activated between attacks in CM, whether CGRP measured in peripheral blood can reflect the trigeminovascular activation, and what is the optimal method to detect and measure CGRP concentrations in human. In the future studies, it may be worthwhile to investigate if CM with a high level of interictal CGRP concentration is a clinically distinct subtype.
The strengths of our study are following. We conducted the CGRP measurement in our central laboratory which have been maintained under a strict quality control. Clinical information was completely blinded to the technician who performed the experiment. Also, several clinical features of migraine were tested. Our study also has limitations. Our NCs were not matched to patients in terms of age, sex, and comorbidities. However, we intended to determine the normal value in healthy young individuals. In addition, our study subjects were recruited from a single university hospital with a single ethnicity. Our study results might not be generalized to the whole migraine population until external validation is made.