The primary objective of the MiBlock trial is to investigate the efficacy and safety of a single-session, bilateral navigation-guided onabotulinumtoxinA injection towards the SPG in participants with resistant chronic migraine.

The primary efficacy endpoint variable is moderate and severe headache days, defined according to the guidelines of the International Headache Society. ²⁰ Data capture is done prospectively as a patient-reported outcome in a study electronic headache Diary (eDiary). In the context of this study, a moderate or severe headache day is defined as a headache day with a recorded maximum pain intensity of ≥4 that lasts at least 4 hours or a day with a headache of any intensity which prompted the use of acute rescue medication. Rescue medications are defined as simple analgesics, triptans and opioids.

The primary efficacy endpoint is the change from the baseline in the number of moderate to severe headache days during weeks 5 to 8 following the intervention. The treatment effect will be assessed by comparing the average change in the treatment group to that in the placebo group. Post-injection weeks 5-8 are selected based on pilot study efficacy data suggesting that the frequency reduction the trial is designed to evaluate, reaches a plateau in this four-week period. ^{18, 19, 21}

The statistical analysis plan (SAP) highlights three key secondary endpoint variables; number of migraine days, headache intensity and number of 30% responders. Other secondary endpoint variables in this study are cumulative hours of moderate and severe pain intensity, days with rescue medication and quality of life measures. Several prespecified exploratory endpoints will also be analysed, including number of crystal-clear headache-free days, number of 50% responders, 75% responders and 100% responders, change in autonomic symptoms and patient-reported outcomes questionnaires.

Definition for safety reporting in this trial, including the definition of adverse event (AE) and serious adverse even (SAE), adheres to the European medicines agency standard. ²² Only treatment-emergent adverse events (TE-AE) and SAEs (TE-SAE) are collected. The primary safety endpoint is the occurrence of TE-AE and TE-SAE which will be compared between the placebo and the treatment arm.

This study is a multi-centre, randomised, quadruple-blinded, placebo-controlled trial. The study is investigator-initiated, exclusively publicly funded, and sponsored by a public hospital. It is designed according to the guidelines of the International Headache Society. ²⁰ The trial is conducted entirely within the framework of the public healthcare system in Norway, which also reimburses travel and accommodation expenses. In case a co-payment is required, the study will reimburse such expenses. No payment is made to the study subjects for their participation.

An overview of the study procedures is provided in Figure 3. Participants are recruited from four general neurology clinics located at university hospitals in Oslo, Trondheim, Bergen, and Bodø. Each site represents one of the four administrative healthcare regions in Norway, enabling nationwide participation. Participants may be recruited directly at the site, referred from other neurology departments across Norway, referred by their general practitioner, or may initiate contact themselves with the recruitment sites. Study information is made available at institution and governmental webpages and social media platforms during the recruitment period.

Study participation is commenced by an inclusion visit followed by a baseline period of at least 28 days. Participants are instructed to keep a daily electronic headache diary (eDiary) for the entire duration of the study. At visit 2, prior to randomisation, a formal re-evaluation of study eligibility including a review of eDiary adherence and headache frequency, is performed. Participants assessed ineligible up to or at this point, are considered screening failures and excluded from the study. Treatment allocation follows a 1:1 ratio, with a computer-generated randomisation scheme applying block randomisation with randomly selected block sizes. CT and MR scans of the region of interest are performed prior to the injection procedure to allow trajectory planning and a navigation-guided injection. The follow-up period of 12 weeks ends with a final visit, which concludes study participation. At the end of this visit, participants and investigator are asked to assess blinding.

The inclusion visit is conducted in accordance with a dedicated standard operating procedure (SOP), which standardizes both the sequence and the content of evaluations performed by the investigator. Patients receive a written copy of the study’s informed consent form (ICF) prior to the visit. The session begins with a review of eligibility criteria. For patients who meet these criteria, the investigator then delivers scripted, standardized study information orally, supported by 10 illustrated cards. This approach ensures that all participants receive the same essential information regarding the study design, procedures, and safety profile. Patients are encouraged to ask questions during and after this part of the session. Only at this stage, if the investigator deems the patient eligible and adequately informed, are patients invited to sign the ICF, formally consenting to participate in the study. The investigator also signs the ICF to confirm that the information was provided in accordance with the SOP governing this visit.

Inclusion and exclusion criteria are listed in Table 1 ³⁴ (Extended data). Eligible patients are adults 18-70 years of age, with resistant chronic migraine at time of inclusion; chronic migraine for 1 year prior to inclusion and debut of episodic migraine before the age of 50, and chronic migraine before the age of 65. Chronic migraine is defined according to ICHD-3-criteria as more than 3 months with 15 or more headache days each month, of which at least 8 days per month have pain and associate symptoms that comply with the migraine criteria. ²³ The definition of resistant adheres to that of the European headache federation consensus statement of 2020 ²⁴ and is in the study protocol operationalized as insufficient treatment effect, contraindication(s) or intolerable side effect(s) of at least three preventive medications from at least two prespecified drug classes.

Study treatment can be an add-on to a stable prophylactic regimen, provided it is kept unaltered during the entire trial participation and a minimum of 3 months before inclusion. The only exception is that use of botulinum toxin for any indication (including cosmetic) is not allowed for a period of four months prior to inclusion and during the study. Concomitant use of monoclonal antibody therapies targeting the CGRP system is allowed, but require a stable dose corresponding to 5 half-lives, i.e. a minimum of 20 weeks, prior to inclusion. Erenumab, fremanezumab, galcanezumab and eptinezumab are available in Norway. To date, atogepant is the only gepant reimbursed in Norway as a preventive treatment, but use of all gepants as prophylactic (in stable dose) or abortive drug is allowed. No restrictions on abortive medication use exist except for limiting opioids to 4 days per month. Participants cannot have secondary headaches, except for non-opioid medication overuse headache. Participants on antithrombotic medication could be considered for study participation if the treatment can be paused. In the case of monotherapy with acetylsalicylic acid, a pause is not required. A medical history with non-response to more than 6 preventive drugs from 7 predefined drug classes prior to participation is a cause for exclusion. As previous headache diary data may be unavailable, this assessment is at the investigators discretion. A patient-reported moderate response or intolerable side effect to a drug, is not classified as non-response to the drug.

The trial complies with the “Recommendations related to contraception and pregnancy testing in clinical trials” by the Clinical Trials Facilitation and Coordination group of Heads of Medicines Agencies (HMA). ²⁵ Woman of childbearing potential (WOCBP) are required to take pregnancy tests at visit 1 and 2 and use highly effective contraception (HEC) as defined in the in recommendation the first 4 weeks post-injection. The Competent authority authorized one exception to the recommendation document, in that sexual abstinence from heterosexual intercourse during the first 4-week period post-injection could be regarded as HEC, albeit not the preferred and usual lifestyle of the subject.

Injection procedure

The interventional procedure in MiBlock is a single session, bilateral, percutaneous injection towards the SPG performed using the MultiGuide® device aided by image-guided navigation. The MultiGuide® is a novel medical device developed by the research group from St. Olav’s Hospital and NTNU and holds a CE-label for the procedure. A commercially available surgical navigation system is used to plan an injection trajectory and provide real-time guidance of needle placement. The needle entry can either be supra- or infrazygomatical, depending on which trajectory provides best access to the sphenopalatine fossa. An example of an infrazygomatical trajectory is illustrated in Figure 2.

The intervention is performed at visit 2 in an outpatient setting at St. Olav’s University Hospital (Trondheim) and Oslo University Hospital by consultant neurosurgeons with prior experience in image-guided cranial procedures. Local surgeons are trained in the injection technique by the coordinating neurosurgeon (DB), who has extensive experience with the procedure. To ensure procedural consistency, the coordinating neurosurgeon is present at each site during the first procedures to directly supervise and oversee the injections. Under local anaesthesia a small incision, approximately 2 mm wide, is made in the skin prior to needle insertion. The participant is observed for at least one hour following the procedure.

Blinding and investigative medicinal product

This study employs rigorous allocation concealment and blinding strategies referred to as quadruple-blinding. This involves blinding of the (1) participants and (2) investigators (double blinding), in addition to blinding of the (3) statistician/outcome assessor, and (4) the pre-specified order in which the statistical analysis will be performed including preplanned tables for the analysis subject to blinded analysis.

All personnel involved in the injection procedure were blinded to participant allocation. Designated unblinded study staff (“Unblinded Study Drug Coordinators”, USDCs) were automatically informed of each participant’s randomisation status via personal secure email notification. Based on this, the USDC prepared two sterile syringes (2 mL Omnifix Luer Lock, B. Braun, Germany) containing either: 1.

onabotulinumtoxinA 25 units (Botox®, Allergan/AbbVie, USA), diluted in 0.5 mL sterile 0.9% sodium chloride injection solution (B. Braun, Germany), or

All substances were sourced from the outpatient clinic’s medication stock; no catalogue numbers were available. Syringes were filled under sterile conditions following a standardized SOP, labeled with participant ID and initials, and sealed in sterile packaging. The USDC had no contact with participants or study staff involved in injections and placed the syringes in a predefined, access-controlled storage area for blinded staff to collect. Syringes were used within six hours of preparation, as per stability guidance from the manufacturer and SOP. The USDC was responsible for maintaining and storing the Drug Accountability Log, in which the participant ID number and initials, batch number, expiry date, and the date and time of preparation were recorded.

To assess the integrity of the double-blind design at the final visit, each participant and investigator is asked to indicate the treatment allocation of the subject, choosing from one of three options: Verum, Placebo, or “Don’t know”.

According to the SAP, independent statisticians will perform a blinded analysis without knowledge of the actual treatment allocation, using a dataset in which randomisation status is coded as Group A and Group B. The blinded analyses will follow a pre-specified order to minimize the risk of unintentional unblinding before completion. The result table templates for the blinded analysis are pre-specified for the statistical analysis report, and included as an appendix to the SAP, that will be finalized prior to the study data lock.

The Department of Neurology at the University Hospital in Trondheim provides a 24/7 emergency telephone service for all study participants. The on-call neurologist is responsible for responding to calls and may initiate unblinding if necessary. To enable continuous unblinding availability, a research nurse at the Department of Neurology in Trondheim, who is otherwise not involved in the study and has no contact with study participants, is automatically notified by email at the time of randomisation. This nurse prints the randomisation outcome for each participant and places it in a sealed envelope labelled only with the participant’s unique identification number. These envelopes are stored in a secure archive at the Department of Neurology, accessible to the on-call neurologist at all times.

Given that the effects of onabotulinumtoxinA cannot be reversed and each participant receives a single bilateral injection, the SOP states that any unblinding should, if feasible, be discussed in advance with the coordinating investigator.

A source data list specifies what constitutes source data. Data on visits, eligibility assessments, current medical history, and medication use, including headache diagnoses, and previous use of preventive headache medication, in addition to safety data has its source documentation in the patient electronic medical records (EPJ). Data handling is done according to the General Data Protection Regulation and institution guidelines. Data will be stored for 25 years after study completion for control purposes as required by Norwegian law.

eCRF

An electronic Case Report Forms (eCRF) applying the WebCRF3 software, was developed for this study. The webCRF3 software is provided and maintained by the research department at St Olav’s university hospital. Data collected at study visits, telephone consultations and safety information are entered. Data is stored in a pseudo-anonymous manner, by which each study participant is recognizable by his/her unique participant study identification (ID) numbers.

eDiary

Evaluation of the primary endpoint is solely based on patient-reported daily eDiary entries collected prospectively during the entire trial via a mobile phone application (app). The participant must make entries every day, including days when they do not experience headache. The tool is designed such that when participants report having headache, they will be prompted to answer additional questions about the maximum pain intensity, duration of headache, duration of headache of intensity ≥4, associated symptoms including aura, abortive medication use and self-reported identification of headache as migraine and/or other primary headaches. On days without headache, questions related to aura and potential prodromal or postdromal symptoms are asked, as well as on the presence of rhinorrhoea, nasal congestion and tearing. Regardless of headache status, data on impairment at work or in other activities, and potential adverse events are collected daily.

Investigators can enter an online dashboard that displays if daily entries to the diary have been done or not. The dashboard allows identification of participants not adhering to the diary. The system also prepares a report of patient-reported potential adverse events, to be followed up during the study. The eDiary app allows registration in a three-day interval (present day, and the two previous days). A back-up paper diary is provided to participants in case of technical problems.

Safety data

Safety data are collected prospectively through the eDiary app and through interviews as part of study visits and phone consultations. Study personnel register all AEs in a designated section of the eCRF.

Periprocedural adverse events is defined as AEs starting during or within one-hour post-injection, and are assessed for causal relationship to IMP, medical device and local anaesthesia. All other adverse event occurring in the 12 weeks postinjection are assessed for causal relationship to IMP and the medical device only. All adverse events will be coded using the MedDRA version 28.0 and severity is graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. Causality assessments and severity grading are performed by investigators at each site. Coding is performed according to the project management plan by one blinded medical monitor.

Previously reported adverse events to SPG-injections ( Table 2) and Summary of Product Characteristics (SmPC) for onabotulinumtoxinA are considered expected adverse events.

Acceptability of treatment

Short-lasting periprocedural pain and procedure duration are documented in the eCRF. Immediately after the needle is withdrawn following the second injection, the participant is asked to rate their pain using an 11-point Numerical Rating Scale (NRS) in the following way: (1) Rate the maximum pain experienced during the injection; (2) Rate the current level of pain. Additionally, at the final visit, the participant is asked whether they would be willing to undergo the injection again. Treatment acceptability will be evaluated based on this data.

Device related events

Any device deficiencies are collected after the injection procedure. Any adverse event that is deemed to have a possible, probable or causal relationship to the medical device is also registered as an adverse device event.

Patient-reported outcomes questionnaires

The following questionnaires are used in the study. All copyrighted instruments were used with proper permission or license. References to copyright and licensing are provided for each tool below: •

Headache Impact Test™ (HIT-6™; © 2001 QualityMetric Incorporated and the GlaxoSmithKline Group of Companies. All rights reserved). ²⁶ The HIT-6™ was licensed (Lic.no. QM49738) under a non-commercial license agreement for this study. All rights reserved.

Patients’ global impression of improvement (PGI-I) and of change (PGI-C) are registered at week 1, 2, 3, 4, 6, 8 and 12 postinjection. Depression and anxiety experienced by participant are evaluated with the 14-item Hospital Depression and Anxiety Scale© (HADS©) at visit 1 and post-intervention week 8 and week 12. At the same milestones, migraine-related quality of life measures and headache disability are mapped with the 6-question Headache impact test™ (HIT-6™) and the 14-question Migraine-Specific Quality-of-Life Questionnaire© Version 2.1 (MSQv2.1©). Work productivity and activity impairment: Migraine v2.1 (WPAI:M v2.1) questionnaire will be completed at visit 1 and weeks 12 post-intervention.

Blood samples and biobanking

Blood samples for the analysis of haemoglobin, leukocytes with differential count, thrombocytes, and C-reactive protein (CRP) are collected at visits 1, 2, and 3 ( Figure 3). At these same visits, participants may also provide additional consent to have blood drawn for biobanking. The latter follows a dedicated SOP and is stored at a research biobank at Oslo University Hospital.

Data management plan

Data management is guided by the data management plan. Data verification of questionnaire and haematology report data in the eCRF will be conducted on 10% of participants at each site.

Research hypothesis

This trial is designed to investigate the superiority of onabotulinumtoxinA compared to placebo when injected into the SPG using the MultiGuide® device with respect to the number of moderate to severe headache days: •

The null hypothesis is that the average change from baseline in the number of moderate to severe headache days during weeks 5-8 post-injection does not differ between participants receiving onabotulinumtoxinA or placebo.

Sample size

The MiBlock trial has a confirmatory statistical strategy that pre-specifies just one single hypothesis related to the primary endpoint. To that end, the sample size was calculated by a professional, independent biostatistician based on the data from the pilot study. ¹⁸ The standard deviation (SD) for the change in number of headache days per week between baseline and week 5 – 8 was 7. To detect a difference of 3.5 moderate to severe headache days per month between active and placebo groups in weeks 5 – 8, provided 80% power to reject the null hypothesis at 5% level nominal of significance and assuming a drop-out rate of 25%, a total of 170 patients will be randomised to treatment in this study. The chosen effect size was informed by results from the pilot study and by a clinical assessment of what would constitute a clinically meaningful benefit, particularly in light of the invasive nature of the treatment procedure.

Subject adherence, protocol deviations and violations

The definition of protocol deviation and violation is based on the following statement ³¹ and should in this study cover all incidences of participant non-adherence to study procedures after enrolment (ie. post-randomisation). Protocol violations in this study are headache diary nonadherence defined as <90% adherence and/or alterations in headache preventive treatment during trial participation. Electronic headache diary adherence will be calculated at 4-week (28 days) intervals. Protocol violators are excluded from the per protocol analysis.

The main analysis is planned when all participants have concluded the 12-week follow-up after the injection and the database has been locked. All statistical analyses are pre-planned and will be described in a SAP finalised prior to data lock. The SAP will also include a preplanned order for data analysis, as well as the results table templates for the blinded analysis. The order of data analysis is defined a priori in order to ensure that the main efficacy analyses are completed without the risk of unblinding which may occur when analysing the safety data.

The full analysis set (FAS) will include all randomised participants who received the injection. To adhere to the intention-to-treat ideal for the FAS population, participants that undergo unsuccessful injections (i.e. Investigative medicinal product (IMP) not administered bilaterally or unilaterally) are included in the FAS.

The FAS will be used for all efficacy analyses. Briefly, for the primary outcome, a mixed logistic regression model will be used to estimate the difference in the change from baseline in mean monthly headache days between the treatment and placebo group. The model will include the presence or absence of headache on each day with recorded information from baseline until 12 weeks after the injection as a dependent binary variable, treatment allocation, time-period (baseline, week 1-4, week 5-8 and week 9-12) and an interaction term between treatment allocation and post-injections time-periods are included as categorical fixed effects, and with participant ID as a random effect. Site of inclusion will also be assessed as a potential random effect prior to unblinding of the statistician to treatment allocation. All available data will be included in this analysis, regardless of proportion of eDiary days completed. The model works under the assumption that data are missing at random. No formal missing-not-at-random analysis are preplanned.

The analysis of secondary outcomes will use the same strategy, although using mixed linear regression models for continuous outcomes. Additionally, sensitivity analyses for the primary and secondary outcomes will be performed, which will include restriction analysis to the subgroup of participants with technically successful bilateral injections and the per-protocol set. Analysis across all three postintervention 28 day periods will be performed, to allow evaluation of temporal patterns of treatment response.

For responder endpoints, responder status is defined based on the percentage change in the number of moderate to severe headache days, comparing each 28-day post-intervention period with the 28-day baseline period. To reduce the risk of uncertainty or misclassification of responder status due to incomplete data, responder endpoints will be calculated only for participants who have completed at least 90% of daily headache diary entries in both 28-day periods used for the respective comparisons.

Prespecified subgroup analyses will be conducted for each of the following baseline factors: sex, baseline medication overuse, patient-reported accompanying cranial autonomic symptoms during migraine attacks, and concomitant use of preventive treatments targeting the CGRP system.

A prespecified assessment of the success of the double blinding is included in the SAP and will be included in the statistical analysis report. The assessment will use the blinding index proposed by Bang et al. ³² In addition, the blinding index proposed by James et al. ³³ will be calculated, and the results will be compared as recommended by Bang et al. ³²

The study is monitored by independent monitors according to a monitoring plan developed based on their own project risk analysis. In addition, two independent, unblinded monitors are assigned to the two injection sites to monitor the drug accountability log.

An independent data safety monitoring committee consisting of one consultant anaesthesiologist with previous service on ethical boards, one neurologist with headache sub-specialisation and one biostatistician will assess safety at 10%, 25%, 50% and 75% of treated participants and advise the sponsor on the safety profile.

Upon data lock following study completion, only the blinded study statistician will have access to the complete dataset, in accordance with procedures prespecified in the SAP. Once the analyses outlined in the SAP have been completed, the full dataset will be made available to all study investigators to facilitate preparation of the results for publication and to allow their independent review of the data. No contractual or other restrictions apply to this process, beyond compliance with applicable data protection regulations.

The results of this trial will be published in international open access peer-reviewed medical journals and presented at relevant scientific meetings. The SAP and statistical analysis report will be made available upon publication. Publication will adhere to the Vancouver guidelines.