Clinical trial design in FDA submissions for neuromodulation devices, 1960-2023: a systematic review and meta-analysis.

Introduction: Integration of clinical trials into FDA approval processes is essential for ensuring the safety and efficacy of neuromodulation devices.

Methods: From 1960 to 2023, 125 FDA-approved neuromodulation records corresponding to 109 publications (64 randomized, 45 nonrandomized trials) were identified. Regulatory records and published trial characteristics were reviewed.

Results: Safety outcomes were reported in 95.4% of randomized and 82.2% of nonrandomized trials, with serious adverse events in ~ 75% of studies. Efficacy analyses showed small Cohen's d effect sizes in 65.85% of outcomes, medium in 24.39%, and large in 9.76%, with a median NNT of 4.65. Industry funded 87.5% of randomized trials, and conflicts of interest were disclosed in 92.2% of publications. Approximately 20% of devices were recalled - exclusively invasive. Randomized trials had larger sample sizes (median 152 subjects) than nonrandomized studies (median 53; p < 0.001), and study duration varied significantly by indication (p < 0.001).

Conclusions: Our findings reveal considerable heterogeneity in the evidence supporting neuromodulation devices, with trials showing systematic safety documentation but modest efficacy. Industry-funded trials dominate the evidence base, with recalls primarily limited to invasive technologies. Evidence gaps persist in minority representation and longitudinal data.