Is Repetitive Transcranial Magnetic Stimulation Effective in Reducing Endometriosis-associated Pain

Is Repetitive Transcranial Magnetic Stimulation Effective in Reducing Endometriosis-associated Pain: A Randomized Controlled Trial

The goal of this research is to improve pain outcomes for the over 500K Canadian women, girls and gender-diverse individuals who are newly diagnosed with endometriosis each year. Chronic pain that persists after interventions for endometriosis is a huge problem. There is some evidence that endometriosis-associated pain (EAP) is, at least to some extent, associated with changes in pain physiology, particularly central sensitization of pain. There is currently no effective evidence-informed intervention that addresses EAP. Yet a recent feasibility trial on a repetitive transcranial magnetic stimulation (rTMS) intervention demonstrated promising results compared to a sham intervention for reducing pain in a sample with EAP.

The objectives of this trial are:

1. to evaluate the effectiveness of an rTMS intervention for pain reduction among those with recalcitrant post-operative EAP,
2. to inform on the utility of a long (10 session) vs short (5 session) protocol for pain reduction among those with recalcitrant post-operative EAP
3. to determine if any improvements in pain observed 30 days after an rTMS intervention are retained 6 months later
4. to identify physical and psychosocial mediators that impact the successful reduction of pain among patients with EAP treated using rTMS.
5. to describe patients' perceptions of and satisfaction with rTMS as an intervention for EAP.

This will be a double-blind, randomized-controlled trial (RCT) using a 2X2 factorial design. The protocol was developed following the Consolidated Standards of Reporting Trials (CONSORT) statement and also follows the Canadian Institutes for Health Research (CIHR) Strategy for Patient-Oriented Research (SPOR) to adapt research activities and dissemination strategies to patients equitably and effectively.

The interventions include repetitive transcranial magnetic stimulation (rTMS) delivered using either a real or a sham coil and one of two intervention durations (short: 5-session, long: 10 session), thus four study groups with 1:1:1:1 allocation.

Intervention: The rTMS intervention will be delivered to the primary motor cortex (M1) on the left side, over the hand representation using high frequency (HF) rTMS. Since there is no known somatotopy for pelvic pain, handedness is not thought to be important, however, handedness will be recorded to verify this. The rTMS intervention will be applied by way of a Magstim Rapid² system (Whitland, UK) coupled with a 70mm cooled figure-of-8 coil. The coil position and orientation will be standardized between treatment visits through using a Brainsight neuronavigation system (Rogue Research, Canada). A standard Montreal Neurological Institute (MNI) atlas will be used. The stimulation target will be individually defined at baseline as the site eliciting the highest averaged motor evoked potential (MEP) peak-to-peak amplitude in the first dorsal interosseous (FDI) muscle. The resting motor threshold (RMT) will be defined as the lowest intensity stimulus that elicits a MEP in the FDI of the right hand, with a peak-to-peak amplitude of at least 50 uV in 50% of trials. MEPs will be recorded using the on-board EMG system. Each rTMS session (real or sham) will consist of a total of 1500 pulses: 15 sets of pulses delivered at 10Hz for 10s at 80% RMT, separated by 50s intervals. The short protocol will involve interventions on 5 consecutive days, while the long protocol will include a 2-day break and another 5- day intervention. Acceptability, adverse events and side effects will be recorded after each session. Missed visits will be rescheduled wherever possible (but tracked). and adherence will be recorded as a proportion of sessions attended.

Participant allocation: Concealed, computer-generated allocation will be performed using permuted block randomization (block size= 4) only after a participant has consented and completed all questionnaires at baseline. The allocation sequence will be managed by an independent investigator not associated with the project. It will be concealed from the study team through storage in a password protected file on an encrypted computer, backed up in a password-protected directory. Groups will be labelled as A-short, A-long, B-short, B-long, to keep investigators and participants blinded as to which coil (A/B) provides real/sham rTMS (only one investigator who will have no direct involvement with participants will know which coil provides real/sham rTMS; they look, sound and feel identical). Treatment duration will be revealed only to the recruitment officer, to the individual delivering the intervention, and the participant, since they will need to know the length of the protocol to schedule intervention sessions and follow-up. Gender will be recorded and if possible, gender-based sub-analyses will be performed. the sample will not be stratified by race/ethnicity, however efforts will be made to ensure diversity in the sample, as the prevalence of endometriosis also appears to be influenced by race/ethnicity. Recruitment bias will be minimized through using multiple recruitment sources with broad reach. Consistent with the CONSORT extension recommendations for trials involving non-pharmacological interventions, the researchers who perform baseline and follow-up assessments will not be involved in the treatment, and will be blinded to group allocation.

Data collection: A series of questionnaires will be completed using an on-line platform before and after the intervention and laboratory-based assessments will occur on the day of the first and final (5th or 10th) intervention session, at the later of which participants will report their patient global impression of change (PGIC) in pain, and their patient satisfaction with treatment (PST). All participants will be followed daily to record their daily pain (NRS 0-10) in the 30 days before the first and 30 days after the last rTMS intervention session using automated text messaging or e-mail notifications, or phone calls if that is their preference. 30 days after the intervention (primary end point), and 6 months later, participants will provide their PGIC for pain, PST, and complete the Brief Pain inventory (BPI), the Endometriosis Health Profile (EHP-30) and the Beck Depression Index (BDI), through an on-line platform (reminders will be sent and non-completion will be monitored, with follow-up by the protocol officer to ensure completion wherever possible). Individuals randomized to the real intervention (both short and long protocols) will be invited to participate in a semi-structured interview through which we aim to capture a nuanced impression of the effectiveness of the intervention, its acceptability, and other observations.

Target Sample size: The target sample size is 152 participants (38 per group). The sample size was estimated based on findings of a prospective cohort study by Pinot-Monange et al. which suggested that 75% of participants (9 of 12) will be improved on the PGIC, and that there will be a reduction in reported pain sensitivity (d=0.62) associated with the short rTMS protocol. The rTMS literature suggests that a moderate effect size (Cohen's d=0.05), will be achieved, suggesting that a sample of 30 per group will provide sufficient power to detect an intervention effect. A simulation in R based on 100 replications using a small (d=0.2) effect suggested that such an interaction would be detected (power = 0.80) with a sample size of n=100 (25 per group). The sample size includes estimated conservative drop-out rates, 15%, at the primary end point, and 25% after 6 months.

Recruitment sources: Recruitment will be from a broad range of sources within the Ottawa, Ontario and Gatineau Quebec regions of Canada, including local gynecology clinics, local physiotherapy clinics, social media platforms (Instagram, X, Facebook), printed posters at sites within the community related to social interests, cultural practices, and religious beliefs, and sexual health resource centers.

Planned analyses: Baseline participant characteristics will be aggregated by group and summarized using descriptive statistics. Intent-to-treat analyses will be performed to address Objectives 1-3. To answer Objectives 1 and 2, and to address relevant secondary objectives, 2-way repeated-measures ANOVAs will be performed for continuous variables, including intervention (2 levels: real, sham) and duration (2 levels: short, long) as main effects, and the interaction between intervention and duration, using α=0.05 and adjusting for unequal variances if necessary. Where there are significant interaction effects, Cohen's d and marginal means will be used to estimate effect sizes. Where no significant interactions are found, within- and between-group main effects will be estimated using Cohen's d. Chi-Square analyses will be used to evaluate group differences in the proportion of participants reporting meaningful clinical changes in pain based on the average daily pain (>30% improvement) and the PGIC (somewhat to very much improved), generating Odds Ratios and estimating the number needed to treat with 95% confidence. Analyses after the 6- month follow-up will include time as a main effect. Wherever possible, the data will be disaggregated by gender and race/ethnicity, and sub-analyses will be performed. Where this is not possible, descriptive analyses by gender and race/ethnicity will be reported, highlighting observations that may inform future studies.

To address Objective 4, univariate analyses using only data from those who received real rTMS will be used to identify trends (p<0.15) in baseline and demographic data such as, age, race, gender, daily pain, EHP-30, BPI, Central Sensitization Inventory (CSI), Pain Catastrophizing Scale (PCS), BDI, State Trait Anxiety Inventory (STAI), pressure pain threshold (PPThresh), pressure pain tolerance (PPTol), temporal summation of pain (TS), conditioned pain modulation (CPM) and Tampon Test, suggestive of differences between successes (>30% improvement) and failures. Separate binary logistic regression models will examine the relationship between independent variables tending to differ by group and clinically relevant improvements in dependent variables [PGIC (somewhat to very much improved), daily pain (>30% improved), BPI (> 30% improved) and EHP-30(> 30% improved). Only three predictors will be included in the final model- the one's with the largest effect sizes in univariate testing. Bootstrapping (X1000) will be used to improve the robustness of the models.

Recruitment rate, adherence, adverse events and use of rescue medications (reported on the BPI) will be analyzed descriptively, and any protocol deviations will be described. Post-intervention interview data will be coded, and a thematic analysis will be completed in duplicate using N-Vivo software. Interviews will continue until 15 have been completed, thematic saturation is reached, and no new themes emerge after three consecutive interviews.

Planned Interim analyses: One interim analysis will be performed once n=16 per group have reached the primary endpoint and based on Question 1. The trial will be deemed futile if the conditional power to detect a treatment effect for all intervention groups is <30%. The trial will not be deemed futile based on the power of the interaction effect.

Inclusion Criteria:

-self-reported endometriosis-associated pain (> 3 in the numeric rating scale, 0-10) that has persisted following medical or surgical intervention,

Exclusion Criteria:

• Contra-indications to rTMS (e.g., metal/implants around the head/neck, pacemaker), history of epilepsy (history of seizures) in the family
• Pain symptoms initiated by other known causes (e.g., infections, thyroid disease, autoimmune diseases, gastrointestinal disease)
• Experience a more severe, extra-pelvic pain than that associated with endometriosis
• Pregnancy