Abstract
Introduction
Administration of ATP-sensitive potassium channel opener levcromakalim triggers headache in healthy volunteers and migraine attacks in migraine patients. Here, we investigated the effect of ATP-sensitive potassium channel blocker glibenclamide on levcromakalim-induced headache in healthy volunteers.
Methods
In a randomized, double-blind, placebo-controlled, three-way cross-over study, 15 healthy volunteers aged 18–40 years were randomly allocated to receive glibenclamide and levcromakalim (day 1), glibenclamide and placebo (day 2), and placebo and placebo (day 3) on three different days separated by at least 1 week. The primary endpoints were the difference in incidence of headache and the difference in area under the curve for headache intensity scores (0–12 hours) between the days.
Results
Fifteen healthy volunteers completed the 3 days of the study. More participants (12/15, 80%) developed headache on the glibenclamide-levcromakalim day compared to the glibenclamide-placebo day (5/15, 33%) (p = 0.01; mean difference 47%; 95% confidence interval 18–75%) and compared to the placebo-placebo day (1/15, 7%) (p = 0.001; mean difference 73%; 95% confidence interval 48–99%). We found no difference in headache incidence between glibenclamide-placebo day and placebo-placebo day (p = 0.12; mean difference 27%; 95% confidence interval 1.3–52%). The area under the curve for headache intensity was significantly larger on the glibenclamide-levcromakalim day compared to the glibenclamide-placebo day (p = 0.003); and compared to the placebo-placebo day (p = 0.001). We found no difference in the area under the curve between the glibenclamide-placebo day compared to the placebo-placebo day (p = 0.07). The median time to onset for headache after levcromakalim infusion with glibenclamide pretreatment was delayed (180 min) compared to levcromakalim without pretreatment (30 min) from a previously published study.
Conclusion
Glibenclamide administration did not cause headache, and glibenclamide pretreatment did not prevent levcromakalim-induced headache. However, glibenclamide delayed the onset of levcromakalim-induced headache. More selective blockers are needed to further elucidate the role of the ATP-sensitive potassium channel in headache initiation.
Introduction
ATP-sensitive potassium (KATP) channels are expressed in the intra- and extracerebral arteries, trigeminal ganglion (TG) and trigeminal nucleus caudalis (TNC) (1,2), and several headache and migraine triggering molecules activate KATP channels (3). Intravenous infusion of KATP channel opener levcromakalim dilates cranial arteries and induces headache in healthy volunteers and migraine attacks in migraine patients (4–6). These data suggest an important role for KATP channels in signaling pathways underlying headache and migraine pathophysiology. Targeting these channels may be a possible therapeutic strategy for future treatment of migraine. KATP channels are composed of eight subunits belonging to two protein families and together the subunits form a stable and functional octameric complex (7,8). One class of subunits is a member of the inward rectifier potassium (Kir) channel family, while the other is a sulfonylurea receptor (SUR), a member of the ATP-binding cassette (ABC) transporter family (8).
Glibenclamide is a widely used antidiabetic drug that belongs to the second-generation sulfonylureas (9). It targets SUR and inhibits SUR-regulated channel activity in several ion channels (9). Preclinical experiments suggest that glibenclamide inhibits KATP channel activation (10), but whether glibenclamide prevents KATP channel-induced headache has not been tested in humans. Here, we investigated the effect of systemic glibenclamide on levcromakalim-induced headache in healthy volunteers. We hypothesized that glibenclamide would prevent levcromakalim-induced headache.
Materials and methods
We recruited 15 healthy volunteers through the Danish test subject website (www.forsøgsperson.dk). Written informed consent was obtained from all participants after detailed oral and written study information. The female participants were required to have sufficient contraception (contraceptive pill or intrauterine device/system (IUD/IUS)). Exclusion criteria were a) history of cardiovascular or cerebrovascular disease, diabetes mellitus, or hypercholesterolemia, b) abnormal electrocardiogram (ECG), c) hypertension at baseline on an experimental day (defined as a systolic blood pressure above 150 mmHg or a diastolic blood pressure above 100 mmHg), d) current pregnancy or breastfeeding, e) daily intake of medication (except oral contraceptives), f) daily smoking within the last 5 years, g) first-degree relatives with a history of diabetes mellitus, and h) history of any primary headache disorders (except episodic tension-type headache for <2 days per month during the last year) or first-degree family members with migraine as defined by the third International Classification of Headache Disorders (11). A full medical examination and ECG were performed on the day of recruitment. The study was approved by the Ethics Committee of the Capital Region of Denmark (H-18052188) and the Danish Data Protection Agency, and was conducted according to the Declaration of Helsinki of 1964, with later revisions. The study was also registered at ClinicalTrials.gov (NCT03886922).
The main study, as described in the study protocol and in the ClinicalTrials.gov entry, comprises two experimental studies: i) A study concerning the cerebrovascular effects of glibenclamide and the effect of glibenclamide on levcromakalim-induced vascular changes in healthy volunteers (under review)); and ii) the present study.
Experimental design
In a double-blind, placebo-controlled, three-way crossover design, the participants were, in a balanced order, randomly allocated to receive a continuous intravenous infusion of 20 mL levcromakalim (0.05 mg/min (50 µg/mL) (Sigma-Aldrich, Darmstadt, Germany)) or 20 mL placebo (isotonic saline) over 20 min after oral administration of 10 mg glibenclamide (Hexaglucon, Sandoz) at the baseline (0 min) on three different days separated by at least 1 week (Figure 1 and 2). Oral glibenclamide followed by infusion of levcromakalim (day 1), oral glibenclamide followed by infusion of placebo (day 2), and oral placebo (multivitamin pill) followed by infusion of placebo (day 3). The levcromakalim or placebo infusion was administered 120 min after the pretreatment with glibenclamide or placebo. Randomization and preparation of the study drug was performed by the Capital Region Central Pharmacy. The randomization code remained in the hospital during the study and was not available to the investigators until the study was completed.
Overview of the study design. Fifteen healthy volunteers were randomly allocated to receive glibenclamide-levcromakalim on one day, glibenclamide-placebo on the second day, and placebo-placebo on the third day. The wash-out period between the days is at least 1 week.
Timeline of the study.
All participants were headache-free for at least 48 hours before each study day. All participants arrived at the clinic after 2-h fasting. The participants were placed in the supine position and a venous catheter was inserted into the left and right antecubital vein for drug (levcromakalim/placebo) and 20% glucose infusion. Then, the participants rested for at least 30 min before baseline measurements of vital signs were performed. The infusion started using a time and volume-controlled infusion pump. Vital signs were monitored during the experiments using an MR-compatible system (Veris Monitor, Medrad, Warrendale, PA). Mean arterial blood pressure (MABP), heart rate (HR), respiratory rate, blood oxygen saturation, nasal end-tidal CO2 tension (water trap and gas sample line, Medrad, Warrendale, PA), were continuously monitored and recorded every 10 min. Room temperature was continuously monitored and recorded every 5 min.
One of the authors (MMK) evaluated eligibility, obtained informed consent, and enrolled the participants. Experiments were carried out at the Danish Headache Center, Department of Neurology, Rigshospitalet Glostrup, from 5 April 2019 to 27 October 2019.
Headache and accompanying symptoms
Immediately before oral glibenclamide/placebo administration, and every 10 min from the administration, we asked participants specifically about the presence of headache, nausea, photophobia, and phonophobia. In case participants reported headache, the characteristics (intensity, throbbing/constant pain, aggravation by activity, location, associated photo/phonophobia, nausea or vomiting) were recorded using a standardized questionnaire. Headache intensity was recorded on a numerical rating scale (NRS 0–10) rating pain from none (NRS 0) to maximum imaginable (NRS 10). The participants completed the headache questionnaires hourly until 12 hours (waking hours) after discharge from the clinic. If the symptoms fulfilled ICHD-3 beta (11) criteria C and D for migraine without aura, they were characterized as migraine-like attacks.
Plasma glucose
Plasma glucose concentration was monitored during a 20-min baseline period before the administration of oral glibenclamide/placebo by ABL800 FLEX blood gas analyzer. After the start and when initial fasting glycaemia had declined by 10%, blood glucose concentrations were clamped around this level (4–6 mmol/L) by 20% glucose infusion. Infusion rates necessary to maintain blood glucose after drug intake were registered throughout the ensuing 220 min. The following standard formula was used to calculate glucose infusion rate (GIR):
Data analysis and statistics
Baseline was defined as T0 before the start of oral glibenclamide/placebo administration. For glucose measurement, the baseline was defined as T-20 ‒ 0 before the start. Calculation of sample size was based on detection of a difference between treatments in headache incidence of 80% (4) on the placebo day and 20% on the glibenclamide day at a 5% significance level with 80% power. Sample size was calculated at 12 participants (14), and 15 participants were included to ensure power.
Headache intensity scores are presented as median (range). The primary endpoints were the difference in incidence of headache and the difference in area under the curve (AUC) for headache intensity scores between glibenclamide-placebo day, glibenclamide-levcromakalim day, and placebo-placebo day. The incidence of headache was analysed as binary categorical data with Chi-square test. We calculated AUC according to the trapezium rule (15) to obtain a summary measure to analyse the differences in response between the 3 days. Differences in AUC for headache intensity scores were tested using two-way ANOVA test. To compare the median time to onset for headache after levcromakalim infusion, we used published data conducted in a double-blind, placebo-controlled design with 20 healthy volunteers randomly allocated in a 2:1 order to receive levcromakalim or placebo (4). In that study, 12 of 14 participants reported headache after levcromakalim infusion without pretreatment and the median time to onset for headache was 30 min (4). We used independent t-test to compare the onset of headache after levcromakalim infusion with and without glibenclamide pretreatment.
To compare the systemic hemodynamics from the glibenclamide-placebo day and placebo-placebo day, we used a mixed model with a covariance structure induced by random effects of person, person × day, person × part 1 measurements (part 1: Measurements at 60 and 120) or part 2 measurements (part 2: Measurements at 160 and 200). To compare the systemic hemodynamics from the glibenclamide-placebo day and glibenclamide-levcromakalim day, we used a mixed model with a covariance structure induced by random effects of person, person × day, person × treatment (treatments are placebo or levcromakalim). We tested for period and carry-over effects for all variables with Mann-Whitney test and independent t-test. All analyses were performed with SPSS Statistics version 23 for Windows and a two-sided probability level of less than 0.05 was considered statistically significant.
Data availability
The data supporting the findings in the present study are available from the corresponding author upon reasonable request.
Results
Fifteen healthy volunteers (nine women and six men) completed the study (Figures 1 and 2). The mean age was 24 years, (range 20–39) and mean weight 68.5 kg, range 65–90. There was no carry-over or period effect for value of headache. No differences were found for plasma glucose, pH, bicarbonate concentration (HCO3−), and lactate between the 3 days.
Headache
Twelve participants (80%) developed headache on the glibenclamide-levcromakalim day compared to five (33%) on the glibenclamide-placebo day (p = 0.01); the mean difference is 47% (95% confidence interval (CI) 18–75%), and compared to one (7%) on the placebo-placebo day (p = 0.001); the mean difference is 73% (95% CI 48–99%]. We found no difference in headache incidence between the glibenclamide-placebo day and placebo-placebo day (p = 0.12); the mean difference is 27% (CI 1.3–52%).
The AUC for headache intensity was significantly larger on the glibenclamide-levcromakalim day (13.06 ± 12.38) compared to the glibenclamide-placebo day (1.67 ± 3.12; p = 0.003); and compared to the placebo-placebo day (0.3 ± 1.2; p = 0.001) (Figure 3). We found no difference in the AUC between the glibenclamide-placebo day compared to the placebo-placebo day (p = 0.07).
Headache. (a) Headache intensity and localization on the glibenclamide-levcromakalim day. (b) Headache intensity on the glibenclamide-placebo day. (c) Headache intensity on the placebo-placebo day. The AUC for headache intensity was significantly larger on the glibenclamide-levcromakalim day (13.06 ± 3.2) compared to the glibenclamide-placebo day (1.67 ± 0.8; p = 0.0005); and compared to the placebo-placebo day (0.3 ± 0.3; p < 0.0001).
On the glibenclamide-levcromakalim day, eight participants reported headaches within 5 h and four reported headaches after 5 h from the start of levcromakalim infusion. The median time to onset for headache was 180 min (range 10 min to 10 h), the median duration of headache was 5.5 h (range 1–11 h), the median peak headache intensity was 2 (range 1–9 NRS), and the median time to peak headache occurred 6 h (3–10 h) after the start of levcromakalim infusion (Table 1). Three participants reported migraine-like attacks on the glibenclamide-levcromakalim day compared with one participant on the glibenclamide-placebo day and none on the placebo-placebo day. We found a significant difference between the median time to onset for headache on the glibenclamide-levcromakalim day in the present study compared to levcromakalim without pretreatment conducted in the previous study (p = 0.007) (4) (Table 2 and Figure 4).
Clinical characteristics of headache and associated symptoms in healthy volunteers after (0–12 h observation period).
h: hour; min: minutes; NR: not reported.
Localization/intensity/quality (throb = throbbing; press = pressing; diffuse)/aggravation (by cough during in-hospital phase and by movement during out-hospital phase).
Nausea/photophobia/phonophobia.
Migraine-like attacks are defined according to criteria described in methods.
Pain freedom or pain relief (≥50% decrease of intensity) within 2 hours.
Comparison of headache onset after levcromakalim infusion without glibenclamide pretreatment from a previously published study and headache onset after levcromakalim infusion with glibenclamide pretreatment in the present study.
Headache with and without glibenclamide pretreatment. To compare levcromakalim-induced headache with and without pretreatment with glibenclamide. The lines represent the median headache. The median time to onset for headache after levcromakalim infusion without glibenclamide pretreatment from a previously published study is 30 min, whereas the median time to onset for headache after levcromakalim infusion with glibenclamide pretreatment is 180 min.
Systemic hemodynamics
We found no difference in MABP between the glibenclamide-placebo day and the placebo-placebo day over time (p > 0.05). MABP decreased significantly after levcromakalim (78.76 ± 5.1 at 160 min and 79.42 ± 4.8 at 200 min) compared to placebo (86.16 ± 4.32 at 160 min (p < 0.001) and 86.69 ± 4.11 at 200 min (p = 0.001)) on the glibenclamide days and to the placebo-placebo day (84.0 ± 4.11 at 160 min (p = 0.03) and 84.35 ± 4.69 at 200 min (p = 0.02)) (Figure 5).
Systemic hemodynamic. Mean heart rate (HR) and mean blood pressure (MABP) after administration of oral glibenclamide/placebo and infusion of levcromakalim/placebo. Glibenclamide did not alter HR or MABP. Levcromakalim increased HR and decreased MABP, whereas glibenclamide did not inhibit levcromakalim-induced changes.
We found no difference in HR between the glibenclamide-placebo day and the placebo-placebo day over time (p > 0.05). HR increased significantly after levcromakalim (78.47 ± 10.68 at 160 min and 76.73 ± 11.0 at 200 min) compared to placebo (63.87 ± 7.43 at 160 min (p = 0.001) and 63.73 ± 4.94 at 200 min (p = 0.001)) on the glibenclamide days and to the placebo-placebo day (61.08 ± 5.58 at 160 min (p < 0.001) and 61.33 ± 5.45 at 200 min (p < 0.001)) (Figure 5).
No differences were found for respiratory rate, blood oxygen saturation, and nasal end-tidal CO2 tension.
Discussion
The major outcome of the present study was that glibenclamide administration itself did not cause headache and did not prevent levcromakalim-induced headache. Furthermore, glibenclamide did not attenuate levcromakalim-induced changes in mean arterial blood pressure (MABP) and HR.
It has been demonstrated that levcromakalim infusion induces immediate headache with the median time to onset for headache 30 min after the infusion start (Figure 4) (4). In the present study, participants were pre-treated with glibenclamide 2 h before the start of levcromakalim infusion. The peak plasma concentration of glibenclamide occurs 2 h after oral ingestion and its half-life is 3–4 h (16). A similar design has been used in multiple studies (17–20). Although glibenclamide pretreatment did not prevent levcromakalim-induced headache, the median time to onset for headache after levcromakalim infusion was 180 min. Hence, our data showed that glibenclamide delayed the onset of levcromakalim-induced headache. It is tempting to speculate that the increased headache incidence and intensity 180 min after levcromakalim infusion may be associated with the half-life of glibenclamide and thus, the decrease of plasma glibenclamide concentration. Consistent with previous studies investigating effects of 5–10 mg oral glibenclamide on systemic hemodynamic (17,18,20–23), we found no change in HR and MABP after glibenclamide compared to placebo. However, one study reported that oral glibenclamide (10 mg) increased systolic but not diastolic blood pressure compared with placebo (24).
In preclinical experiments, glibenclamide attenuated potassium channel openers-induced vascular changes without altering the basal vascular tone (23). A few studies have reported that glibenclamide increased the basal tone in rat dural artery (25,26). Furthermore, preclinical studies showed that glibenclamide neither affected the spontaneous firing of nociceptive neurons nor nociceptive transmission upon cutaneous mechanical stimulation (27). In spontaneous trigeminal allodynic rats with increased nociceptive firing (28), glibenclamide increased the threshold to mechanical stimulation (27). Furthermore, glibenclamide had no effect on baseline CGRP concentration in the dura mater and TG but it attenuated capsaicin-induced CGRP release in the dura mater (27). These data indicate that glibenclamide only inhibits hyper-responsivity to sensory stimulus in hyperexcitable nociceptors without altering the baseline threshold. Although glibenclamide is an agonist of transient receptor potential ankyrin 1 (TRPA1), which leads to Ca2+ influx and thus depolarizing (29), the hyperexcitability reported by Christiansen et al. (27) was initiated before glibenclamide administration.
Interestingly, hyper-responsiveness to sensory stimuli and sensitization of peripheral trigeminovascular neurons are considered to be responsible for the characteristic throbbing pain of migraine (30–32). Whether glibenclamide may prevent levcromakalim-induced migraine would be the next question to address in this context. Glibenclamide is a non-selective blocker for SUR subunits (33). Three isoforms of SUR (SUR1, SUR2A or SUR2B) subunits are expressed in the KATP channels (34). SUR1-KATP channels are expressed in the brain and pancreas, SUR2A-KATP channels are expressed in cardiac and skeletal muscle, whereas SUR2B-KATP channels are mainly found in smooth muscle cells including VSMCs (3).
Glibenclamide has higher affinity to SUR1 than the other subunits with relatively limited access to the CNS (33,35). Therefore, hypoglycemia is a frequent side effect of the clinical use of glibenclamide. Although levcromakalim has high affinity to the SUR2B subunit, preclinical evidence demonstrates that glibenclamide attenuated levcromakalim-induced dural and pial artery dilation in an in vivo genuine closed cranial window model and in vitro organ bath (25). Investigations on the role of KATP channels in migraine pathophysiology must reveal whether a high affinity agonist to SUR1 can trigger headache and migraine. Future studies must elucidate the effect of glibenclamide on levcromakalim-induced migraine. Then, more selective blockers are needed to further examine the therapeutic strategies of targeting KATP channels.
Key findings
Glibenclamide administration did not cause headache and did not prevent levcromakalim-induced headache. Glibenclamide did not attenuate levcromakalim-induced changes in mean arterial blood pressure (MABP) and heart rate (HR). Glibenclamide delayed the onset of levcromakalim-induced headache.
Footnotes
Acknowledgements
The authors thank all participating subjects. Further thanks to Lundbeck Foundation and Aase & Ejnar Danielsen Foundation for their support.
Contributions
MMK and HG initiated the study and contributed to study design; protocol development; participant enrolment; data acquisition, data processing, analysis, statistics, and interpretation; drafting and revision of the paper. LK, JP and JMH contributed to data acquisition and critical review of the paper. MA initiated the study and contributed to study design, protocol development, statistics, data interpretation, and drafting and revision of the paper.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MMK has acted as an invited speaker for Novartis and received travel grant from ElectroCore, LLC. MA is a consultant, speaker or scientific advisor for Allergan, Amgen, Alder, ATI, Eli Lilly, Lundbeck, Novartis, and Teva, primary investigator for Alder, Amgen, Allergan, Eli Lilly, Novartis and Teva trials. MA has no ownership interest and does not own stocks of any pharmaceutical company. MA serves as associate editor of Cephalalgia, associate editor of Headache, and associate editor of the Journal of Headache and Pain. MA is President of the International Headache Society. HG, LK and JMH declare no conflicts of interest.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research leading to these results has received funding from Lundbeck Foundation (R155–2014–171) and the Aase and Ejnar Danielsen Foundation.
