High-Intensity, Low-Frequency Periodic rTMS Over the Right Dorsolateral Prefrontal Cortex on Cardiac Autonomic Regulation in Women With Recurrent Pregnancy Loss and Anxiety

Heart rate (HR) will be continuously recorded via ECG (1,000 Hz) during a single rTMS session comprising 20 stimulation cycles (each 60 s: 40 s stimulation at 1 Hz, then 20 s rest). For each cycle, a 50-s analysis window will be defined (-10 s to 0 s pre-stimulation baseline; 0 s to +40 s post-stimulation). Baseline-corrected mean HR for each cycle will be computed as the mean HR in the post-stimulation segment (0-40 s) minus the mean HR in the 10-s pre-stimulation baseline (-10-0 s). Per-participant values will then be averaged across the 20 stimulation time windows, and the primary endpoint will compare these averaged baseline-corrected means between the active and sham rTMS groups.

HRV spectral power at the target frequency will be derived from R-R interval (RRI) series obtained from continuous ECG (1,000 Hz). RRI will be resampled at 1 Hz via cubic-spline interpolation, linearly detrended, and band-pass filtered. Time-frequency analysis will use a derivative-of-Gaussian (DoG) wavelet. For each stimulation cycle (60 s; 40 s stimulation + 20 s rest), a 50-s stimulation time window will be defined (-10 s to +40 s relative to stimulation onset). Frequency-specific entrainment power will be computed as the mean wavelet power within 0.015-0.018 Hz inside this window. Per-participant estimates will then be averaged across the 20 stimulation time windows, and active versus sham groups will be compared.

For each participant in the active rTMS group, the linear correlation (Pearson's r) between these two window-wise measures will be calculated across the 20 stimulation time windows.

For each participant in the sham rTMS group, the linear correlation (Pearson's r) between these two window-wise measures will be calculated across the 20 stimulation time windows.

This outcome will capture the incidence (proportion of participants with ≥1 event) of prespecified mild adverse events during the rTMS session and the immediate 1-hour post-session observation period. Mild events include scalp discomfort at the stimulation site, facial muscle twitching, mild headache, and dizziness. Events will be solicited and recorded in real time by trained operators using a standardized checklist, with participant self-report. Vital signs (blood pressure and resting heart rate) will be reassessed at the end of the session. Events will be classified as mild if transient, self-limited, not requiring medical treatment, and not meeting serious adverse event criteria (e.g., syncope or seizure). Incidence will be summarized by group (active vs sham) for comparison.

This outcome will quantify the incidence (proportion of participants with ≥1 event) of serious adverse events (SAEs) occurring during the rTMS session and the subsequent 1-hour observation period. SAEs will be defined per ICH-GCP criteria (events that are fatal, life-threatening, require inpatient hospitalization or prolong an existing hospitalization, result in persistent or significant disability/incapacity, or are otherwise medically important). Events of special interest include syncope and seizures. Trained operators will actively monitor and document SAEs in real time using standardized case-report forms, including onset time, duration, actions taken, outcome, and relatedness to the intervention. Any suspected SAE will trigger protocol-specified safety procedures (immediate termination of stimulation, vital-sign reassessment, medical evaluation, and referral as needed). Incidence will be summarized by group (active vs sham).

In the prespecified exploratory Stage 2 subgroup, change in HAMA total score from baseline will be assessed at treatment Week 3, at the end of treatment (Week 4), and at 1 month, 3 months, and 6 months after treatment. The HAMA total score ranges from 0 to 56, with higher scores indicating greater anxiety severity.

In the prespecified exploratory Stage 2 subgroup, CGI-S, CGI-I, and HAMD-17 will be assessed at treatment Week 3, at the end of treatment (Week 4), and at 1 month, 3 months, and 6 months after treatment. The CGI-S score ranges from 1 to 7, with higher scores indicating greater illness severity. The CGI-I score ranges from 1 to 7, with lower scores indicating greater improvement. The HAMD-17 total score ranges from 0 to 52, with higher scores indicating greater depressive symptom severity.

In the prespecified exploratory Stage 2 subgroup, response and remission rates will be calculated separately by group at treatment Week 3, at the end of treatment (Week 4), and at 1 month, 3 months, and 6 months after treatment. Response will be defined as a reduction of at least 50% in HAMA score from baseline. Remission will be defined as HAMA <8 together with a CGI-I score of 1 or 2. Between-group differences between the active stimulation group and the sham stimulation group will be evaluated at each time point.

Exploratory mediation analyses will use group as the independent variable, the magnitude of baseline-corrected heart rate deceleration during the first acute session as the mediator, and the change in HAMA score at the end of treatment, 1 month after treatment, and 3 months after treatment as separate dependent variables. The HAMA total score ranges from 0 to 56, with higher scores indicating greater anxiety severity. These analyses are intended to evaluate whether the acute physiological response during the first session may mediate subsequent improvement in anxiety symptoms. Results will be used for mechanistic hypothesis generation and will not be interpreted as confirmatory.

Blinding assessment will be conducted at the 3-month follow-up visit after treatment in the prespecified exploratory Stage 2 subgroup. Participants will be asked for their subjective judgment of treatment assignment using a 5-point scale: 0, definitely not in the active group; 1, inclined to think not in the active group; 2, unable to judge; 3, inclined to think in the active group; and 4, very certain of being in the active group. These results will be described by group, and group differences in distribution will also be compared.

Safety outcomes will include adverse events and serious adverse events occurring during the 4-week treatment period and within 30 days after the final treatment session, together with unresolved adverse events, hospitalization, unintended pregnancy, significant worsening of psychiatric symptoms, and major changes in concomitant treatment at the 3-month and 6-month follow-up visits after treatment. Safety outcomes will be summarized by group and compared between the active stimulation group and the sham stimulation group. No numerical symptom scale applies to this outcome.