Study design, patients and randomisation

This double-blind randomised placebo-controlled trial investigating the clinical and neural effects of a structured exposure to odours in female patients with migraine with aura will be conducted at the Headache Clinic of the University Pain Center at TU Dresden. Using a between-subject design, 54 female patients (age between 18 and 55 years) with diagnosed migraine with aura (according to the international classification of headache disorders, third edition)35 will be recruited and randomly and blindly allocated to one of the two study arms (ie, 12-week odourless training—sham OT or 12-week training with odours—verum OT) according to an apriori randomisation list generated by an independent member of the clinic using block randomisation in Excel generating the allocation sequence. Patients will be enrolled by GG. Assigning the patients to interventions will follow the randomisation. In the randomisation list, numerical codes are assigned to patients. These number codes are noted on envelopes in which the verum or placebo smelling sticks are presorted, ensuring that the investigator is also blinded.

The first patient was included on 24 January 2022. Data acquisition is planned to be completed by the end of September 2023. Study design is illustrated in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Patient flow diagram and visit schedule. OT, olfactory training; TDI, threshold, discrimination, and identification test; TENS, transcutaneous electrical nerve stimulation; QST, quantitative sensory testing; d2-R, Test of attention; HIT-6, headache impact test; DASS, depression and anxiety stress scale; MIDAS, Migraine disability assessment questionnaire; STAI, state-trait anxiety inventory; MQS, migraine-specific quality of life questionnaire; BDI, Beck′s depression inventory; ASC, allodynia symptom checklist; CESD, center for epidemiological studies depression scale; fMRI, functional magnetic resonance imaging; Sham/Verum OT package, will contain four felt-tip pens (verum- training with odors, or sham- Odorless training) an instruction sheet, the headache calendar and the olfactory diary; Lateralization test, used to assess the trigeminal sensitivity.

Patient or public involvement

Participants’ previous experiences with OT27were used to improve the OT experience during this study and to improve the design of the study with for example the creation of a credible OT placebo arm. Moreover, in this study, patients will be asked how satisfied they were with the OT; and whether they would recommend the OT to a friend or relative who also suffers from migraine.

Patients eligibility

Patients’ selection is based on the following inclusion and exclusion criteria. Female participants meeting the international classification of Headache disorders (third edition)35 criteria for episodic migraine with aura as the primary diagnosis will be included (ie, ≥1 migraine episode a month, but <15 headache days a month). Inclusion criteria also entail participants’ age, which will range between 18 and 55 years. Additionally, all participants must be normosmic, that is, no olfactory loss (olfactory assessment by means of Sniffin’ Sticks battery (Burghart, Germany)36 the sum score of Threshold, Discrimination, Identification cut-off of 30.75),37 and free of any acute or chronic sinunasal ailments. When it comes to the use of prophylactic drugs, patients taking stable doses (for at least 3 months) of prophylactic drugs will be included. Regarding acute migraine medication, ≤9 days per month with acute analgesic medication will also be allowed.

Participants will be excluded if they report taking opioids, cannabis, serotonin and norepinephrine reuptake inhibitors, due to the known interference with emotional and sensory perception. Patients will also be excluded if they report taking more than >9 days per month of acute analgesic medication. Pregnancy, significant medical problems such as uncontrolled asthma or seizure disorder, acute cardiac disease, severe and instable psychiatric problems, other chronic pain syndromes, other (non-migraine) neurological diseases or any uncontrolled disease are also part of the exclusion criteria. Patients will also be excluded if there are any cognitive signs of dementia (evaluated through Montreal Cognitive Assessment).38 Patients taking part on the MRI session will be scanned during the interictal state. MRI compatibility and eligibility will be assessed by the local MRI technicians.

Recruitment

Participants will be recruited through local advertisements at the University Hospital of Dresden and through public billboards and online platforms such as websites and social networks. Potential participants will be informed that participation is voluntary and withdrawal is possible at any time without further explanations or any associated consequences. Study approvals were obtained from the Ethics Board of the TU Dresden (Protocol No. BO-EK-353082020) which complies with the standards established by the Declaration of Helsinki. Eventual protocol amendments will have to be approved by the ethics board of the TU Dresden. A written informed consent will be obtained from all participants after the nature of the study is explained and the possibility of clarifications is offered by a neurologist (GG). Participation will only be possible after written informed consent is provided (consent form and patient information are provided as online supplemental materials 1 and 2, respectively). Post-trial care will be conducted at the headache outpatient clinic at the University Hospital of Dresden if needed.

Verum and sham OT procedure

During the OT programme, all patients (N=54) will be offered to choose four out of eleven, commercially available, pleasant odours (citronellal: lemon odour (#ABX321602), orange odour (#FJJ232836D) and eugenol: cloves odour (#S0100148), coffee odour (#P0604646), vanilla odour (#DGI178677B), dark chocolate odour (#JVA332845B), strawberry odour (#ABX321354A), apple odour (#P0602153), lavender odour (#DGFLO794), bergamont odour (#A1285) and sage odour (#A1700), presented in felt-tip pens (‘sniffin’ sticks’), based on their individual preferences. Lemon, orange, vanilla, dark chocolate, strawberry and lavender will be provided by Takasago International Corporation, Japan. Cloves, coffee and apple will be provided by Frey+Lau, Henstedt-Ulzburg, Germany and bergamont and sage by Primavera Life GmbH Oy-Mittelberg, Germany. The chosen odours will be used to perform the daily olfactory exposure. All pens will be labelled with the odour name and patients will be instructed to sniff each pen for 20 s twice a day (during the morning and in the evening). The treatment groups (sham and verum OT) will differ solely with regard to the ingredient contained in the training felt-tip pens given to the participants. Whereas in the verum group, the training pens will contain the active solution previously selected, in the placebo group, the training pens will be filled with a placebo odourless solvent (propylene glycole). To increase the overall compliance and to ascertain that the placebo (odourless) group remains blinded to the treatment allocation, all patients are provided with a cover story. Patients are informed that the training concentrations of the therapy odours are different from the ones that participants have previously smelled in the lab, when choosing the odours. Patients are told that whereas the odours used during the therapy tend to be less strong, and may be insufficiently intense to produce an odour sensation (ie, subliminal odours) they are intense enough to be perceived by the brain and may influence mental processes and behaviour. The training will last for 12 weeks. All participants will receive a phone call by an experimenter every 4 weeks to assess participant’s olfactory function and to reinforce compliance with the training programme (reminder of the cover story). After study completion, all participants will be individually contacted and their group allocations will be revealed. All patients will be offered the possibility of performing or continuing a training with odours.

Spontaneously reported adverse events will be collected by VF and JD, assessed and managed by GG, and reported in the results section.

Primary, secondary and exploratory outcomes

All outcome measures will be assessed by a blinded researcher at the pain centre, TU Dresden or self-administered online using REDCap electronic data capture tools hosted at the University Hospital Dresden.39 40

Pain thresholds have been previously reported to significantly increase in patients with chronic low back pain after OT.27 Hence, the primary outcomes that will be used to assess the effects of OT in pain in migraine patients will comprise the observed change (from pre-OT to post-OT) in mechanical pain threshold (MPT) and in EPT, assessed using quantitative sensory testing41 and transcutaneous electrical nerve stimulation (TENS),42 respectively. MPT will be carried out on the right forehead with the right volar arm used as reference and will be performed with seven pinpricks, with standardised stimulus intensities (8, 16, 32, 64, 128, 256 and 512 mN) and a flat contact surface (0.25 mm diameter). The threshold will be determined in five series of ascending and descending stimulus intensities. EPT will be carried out on the right volar arm using TENS (frequency Hz, pulse duration µs, 2 circular electrodes diameter 3,2 cm). TENS is a well-tolerated system to measure electric thresholds, commonly used in therapeutic settings.43 Electrical detection threshold will be measured by single stimuli of increasing electric current until participants detect the stimulus. After that, a stepwise increase in mA will lead to the level of perception of pain.

Secondary outcomes

Secondary outcomes will comprise changes in olfactory threshold and the number of headache days from pre to post OT. With regard to olfactory threshold, participants will undergo a standardised, validated psychophysical olfactory test, the ‘Sniffin’ sticks’ test battery36 that includes the assessment of odour sensitivity or threshold. Odour threshold is evaluated using phenyl ethyl alcohol (PEA, rose like odour), using a three-alternative forced choice presented by means of a single staircase, using stepwise dilutions in a row of 16 felt-tip pens. A change of 2.5 in the threshold after treatment is considered a significant change. An olfactory diary, where patients report their daily training, the perceived odour intensity during training (range 0–10), and the occurrence of events that can potentially affect patients’ olfactory sensitivity during OT, such as allergies, or viral infections like COVID-19, will be collected and further considered during analysis. To assess OT-related changes on headache days, a standardised headache pain diary44 and the migraine disability assessment questionnaire45 commonly used to evaluate headache related disability as lost time due to migraine, will be used.

Other exploratory measurements such as headache associated pain intensity, acute analgesic intake (both assessed by a standardised headache diary),44 the impact of headaches on life (assessed by the Migraine-Specific Quality of Life Questionnaire,46 Headache Impact Test),47 allodynia (evaluated by Allodynia Symptom Checklist),48 anxiety & depression (evaluated by Depression and Anxiety Stress Scale,49 State-Trait Anxiety Inventory,50 Beck’s Depression Inventory,51 and the Centre for Epidemiological Studies Depression Scale)52 and cognitive function53 will also be assessed before and after OT.

Additionally, this protocol evaluates the neurofunctional changes and the OB volume changes associated with the OT programme. Structural and functional imaging data will be acquired with a 3 Tesla MRI scanner (Siemens Prisma, Erlangen, Germany) using a 32-channel head coil. Structural scans will consist of a T1-weighted 3D MPRAGE (voxel size: 1×1×1 mm, FOV: 256 mm, slices: 160, TR: 2300 ms, TE: 3.43 ms, TI: 900 ms, flip angle: 9°, TA: 5:11 min, acceleration mode: GRAPPA, acceleration factor PE: 2) with full-brain coverage and a T2-weighted turbo-spin echo (voxel size: 0.2×0.2×1 mm, FOV: 87 mm, slices: 28, TR: 6200 ms, TE: 117 ms, flip angle: 160°, TA: 8:30 min, averages: 3) with coverage of the olfactory bulb in coronal direction. The functional scans will use slightly different protocols for olfactory and visual stimuli: Olfactory EPI (voxel size: 3×3×3 mm, FOV: 192 mm, slices: 45, TR: 1000 ms, TE: 40 ms, flip angle: 60°, TA: 3:33 min, acceleration mode: slice acceleration, acceleration factor PE: 2, acceleration factor slice: 3, volumes: 203, multislice mode: interleaved); visual EPI (voxel size: 3×3×3 mm, FOV: 192 mm, slices: 45, TR: 2500 ms, TE: 30 ms, flip angle: 80°, TA: 8:53 min, acceleration mode: slice acceleration, acceleration factor PE: 2, acceleration factor slice: 1, volumes: 210, multislice mode: interleaved). In addition a fieldmap will be acquired (voxel size: 3×3×3 mm, FOV: 192 mm, slices: 48, TR: 508 ms, TE1: 4.92 ms, TE2: 7.38 ms, flip angle: 60°, TA: 3:33 min, multislice mode: interleaved).

Neurofunctional tasks

Interictal (migraine free at least 48 hours before and 24 hours after the imaging session) female patients with migraine with aura will undergo two functional MRI sessions (before and after the structured OT programme). During each imaging session, participants will undergo four tasks: (A) one olfactory, (B) two trigeminal and one (C) emotional task:

1. The olfactory task consists of passive intranasal birhinal administration of peach (#LA1300245, Frey+Lau, Henstedt-Ulzburg, Germany), a pleasant odorant stimulus, commonly used in olfactory imaging studies.54 Peach will be presented in undiluted concentrations and is expected to be clearly perceivable, without causing any trigeminal sensation.

2. Intranasal administration of peppermint (#ABX321352, Takasago, Japan) will be used as a pleasant trigeminal stimulant and will be administered in undiluted concentrations. For trigeminal nociceptive stimulation, we will use intranasal administration of propionic acid (#W292400, Sigma Aldrich, Germany) known to induce nasal irritation.55 The intensity of the propionic acid will be calibrated/adjusted individually using airflow (L/min). For the individual calibration of propionic acid, before the MRI scan, patients will be instructed to lie on their backs, place the intranasal tubes in their noses and assess the discomfort associated with the Propionic administration (aimed to be a 7 on NRS range 0–10). The propionic acid stimulation starts with an airflow 2.5 L/min and the patients will be asked how they perceived the stimulation until an airflow is found at which the patients give a stable (three consecutive) rating of 7. For both olfactory and trigeminal stimulations, odourless/unscented air will serve as a control/baseline stimulus.

During the functional runs for the olfactory and trigeminal stimulations, a block design will be adopted. During each functional run, each block will last for 20 s. In each block, peach or peppermint will be delivered to the patients’ (8 s) alternatively with the baseline stimulus, that is, unscented air (12 s) (figure 2). In the case of propionic acid, the ON stimulation will be divided into 8 blocks of 1 s each that consist of 500 ms of propionic acid and 500 ms of unscented air. In both, olfactory and trigeminal stimulations there will be a total of 10 blocks per run, and each run will last for 200 s. Both odour and trigeminal stimulations will be presented orthonasally using a portable and scanner compatible custom-design, computer controlled olfactometer. Pulses of stimulus embedded in clean air will be delivered birhinally via Teflon tubing (flow rate 2.5 L/min for both Peach and Peppermint). Immediately after each olfactory and trigeminal stimulation, participants will be requested to assess, the intensity of the olfactory and trigeminal stimulation on a scale from 0 to 10 (0 corresponding to ‘not perceivable’ and 10 ‘extremely intense’), and the valence on a scale from −5 (extremely unpleasant) to +5 (extremely pleasant). An additional question related to the level of pain/discomfort felt during the propionic acid stimulation, and regarding the cooling effect felt during the peppermint stimulation will be assessed on a similar scale from 0 to 10.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Schematic diagram of the experimental paradigm regarding olfactory and trigeminal stimulations during the fMRI scan. a) Representation of the functional runs. In total, there will be three functional runs – one for each stimulus ie. peach, peppermint, and propionic acid. Each run will be composed of 10 blocks, lasting 200 seconds each. b) Representation of each block. Each block will comprise an ON (olfactory or trigeminal stimulation lasting for 8 seconds) and OFF (unscented air that lasts for 10 seconds) stimulation. Note that for propionic acid the ON stimulation will be divided into 8 blocks of 1 second each that consist of 500ms of propionic acid and 500ms of unscented air. Image source Free Stock photos by Vecteezy.com.

1. The emotional visual task will consist in the presentation and evaluation of emotional loaded pleasant, unpleasant and neutral photographs selected from the International Affective Picture System (IAPS).56 The IAPS is a standardised, emotionally evocative visual stimulus that has been widely used in imaging studies of emotion and pain. A total of 90 pictures, that is, 30 pleasant, 30 unpleasant and 30 neutral, were selected and will be used. The picture selection will be based on the standardised reported mean valence and arousal.56

The experimental paradigm will consist of 10 blocks of emotionally evocative visual stimuli that will be alternated with a crosshair presented in the middle of the screen. During each block, 10 emotionally evocative pictures will be presented for 2.5 s each followed by a baseline cross displayed for 30 s. Each block (ON and OFF) will last for 55 s and there will be a total of nine blocks lasting in total 495 s (figure 3). The order of the ON blocks as well as the order of the images in each block will be randomised for every participant and session. The presentation of the visual stimuli will be programmed using stimulus presentation software Presentation V.20.1 (Neurobehavioral Systems). While laying in the scanner, patients will view the images displayed on a large screen (1920×1080 pixel) reflected through a mirror.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Schematic diagram of the visual emotional experimental paradigm during the fMRI scan. a) Representation of the functional run which will be composed of nine blocks (3 pleasant, 3 neutral, and 3 unpleasant) presented in a randomized order during 495 seconds. b) Representation of the blocks. Each block will consist of an ON phase (comprising 10 pleasant, 10 unpleasant, or 10 neutral pictures - each picture presented for 2.5 seconds (25 seconds)) and an OFF phase (comprising the baseline cross with a duration of 30 seconds). Image source Free Stock photos by Vecteezy.

When it comes to the assessments of the affective picture system, in line with previous procedures,34 once outside the scanner, patients will be presented with the same blocks of images and will be requested to rate each block of pictures for emotional valence and arousal, using the nine point Self-Assessment Manikin Scale (SAM). SAM scale is an affective rating system that uses graphical figures to depict emotional reactions for valence and arousal.57 The procedure in the scanner takes about 45 min.

Credibility ratings

To assess the credibility of the cover story, three questions are added to the last day of training on the patient’s diary. Patients will be asked how much they think their migraine-related pain has improved since they started training; how satisfied they were with the OT; and whether they would recommend the OT to a friend or relative who also suffers from migraine. Responses to these questions will be given on a scale from 0 to 10, 0 indicating ‘not at all’ and 10 indicating ‘very much’. Additionally and right before unblinding of the study, participants will be asked to which group they think they were allocated (placebo or OT group) and how sure they were that the treatment they had been assigned to was effective for migraines on a scale from 0 (not at all) to 10 (very much).

Sample size calculations

Based on the available literature and previous clinical results,27 we expect an increase in either mechanical or EPT in 55% of patients in the verum group and in 25% of patients in placebo group. The null hypothesis of no difference between increase rates of the two groups will be tested using a binomial test for two independent proportions. If a significance level of 5% (two sided) is assumed, at least 42 patients (21 per group) have to be analysed in order to achieve an 80% test power. To account for large reported drop-out rates of 30%,58 we plan to enrol N=54 participants (27 per group).

Statistical analysis

The primary endpoint of the study is the improvement in either mechanical or EPT after the treatment and will be quantified using the percentage of patients whose pain thresholds increased by at least 30% in comparison to their baseline value. A binomial test for two independent proportions will be performed in order to test the difference in improvement rates between the verum and placebo groups (with a significance level of 5%, two sided). Additionally, independent samples t-test will be carried out in order to compare the average change in pain thresholds between the study groups.

Data collected will be first presented using relevant descriptive statistics, namely, mean and SD or median and IQR (continuous variables), absolute or relative frequencies (count data). The association between study outcomes will be summarised using either a correlation coefficient or OR, as appropriate, with a 95% CI. Questionnaire data measured over time will be presented graphically, highlighting possible differences in average between two arms. Additionally, two-sample tests, for example, Mann-Whitney-Wilcoxon U test, t-test, χ² or Fisher’s exact test, will be performed to compare the study outcomes between arms and within each arm (over time). Finally, a model-based analysis will be carried out, aiming to explore a potential influence of patient characteristics on the endpoints of interest. Mixed-effects regression models will be employed to investigate the time trends in pain level and quality of life. For secondary outcomes and relevant effects, standardised mean differences and CIs will be provided, in addition to p values. Statistical analysis will be implemented using R (R Foundation for Statistical Computing), the latest version available by the time of analysis. Missing data will be handled by methods recommended in ICH E9 (R1).59 If data imputation is performed, sensitivity analyses will be carried out, in order to assess the effect of imputation on the study results.

Neuroimaging data will be preprocessed using the latest fmriprep pipeline.60 Individual difference images, representing changes in blood oxygenation level dependence reactivity (ie, olfactory/trigeminal stimulation vs unscented air; emotional vs neutral visual stimuli) will be calculated by subtracting pre from post-OT contrast maps and further used in second level group analysis. During the second level analysis, between group t-tests will be used to assess the neural reactivity changes associated with OT. Based on previous migraine imaging findings, a region of interest approach will be used. Additionally, a more liberal and general whole brain approach will be also performed.

Based on previous studies evaluating the effect of OT, we expect to see an increase in pain thresholds27 (mechanical and EPT—study primary outcomes), in patients with migraine that will undergo verum as compared sham OT. Additionally, we expect to see an increase in olfactory threshold and a reduction in the number of headache days in patients that were enrolled in verum OT as compared with sham OT. With regard to the functional brain imaging data, we hypothesise that a 3-month OT results in increased olfactory related activity, whereas decreased activations are expected in pain-related areas.

Data management

Behavioural data will be partially collected via EDC-System REDCap (Vanderbilt University, Tennessee). Data evaluation will be pseudonymised after patients have given their consent. The data will be stored pseudonymously for 10 years on a secure database server of the REDCap infrastructure. Pseudo anonymisation is carried out according to a fixed one way algorithm. Lists will be stored separately on secure media. When it comes to neuroimaging data acquisition, personal data of the participants will also be pseudonymised. During conversion, imaging data will be anonymised to omit personal identifiable data from the field headers. Imaging data will be stored (anonymised) on a cloud storage provided by the University Hospital of Dresden https://caruscloud.uniklinikum-dresden.de/. Only authorised study personal will have access to the media and servers and only the principal investigator will be able to export data. All data will be made available on request.

Ethics and dissemination

This study was approved by the Ethics Board of the TU Dresden (protocol no. BO-EK-353082020) and will be carried out according to the principles of Declaration of Helsinki. Patients will be informed that they will have a chance to receive either verum or sham OT. All steps will be taken to minimise any discomfort associated with the study. Immediately on study termination, or voluntary ending of the study, a debriefing session will occur and to minimise any potential disappointment related to randomisation, patients will be offered the possibility of enrolling a verum OT programme. Findings of the study will be disseminated through peer-reviewed journals and scientific conferences.