Post-market safety signals of energy-based surgical devices: a cross-specialty disproportionality analysis of the FDA MAUDE database with robotic-platform validation and regulatory triangulation, 2015-2025.

Abstract

Post-market safety evidence for energy-based surgical devices is fragmented. We characterized adverse-event signals in the FDA MAUDE database, validated them on a third (robotic) platform, and triangulated across regulatory and international sources. Medical Device Reports (2015-2025) under electrosurgical and ultrasonic product codes were analyzed with reporting odds ratios (RORs), stratified by specialty and procedure, with robustness checks. All reports under the robotic-system code were retrieved as a third-platform validation cohort. Signals were triangulated with FDA recalls, 510(k) clearances, the Health Canada and Australian TGA databases, exposure descriptors, and a retrospectively applied Poisson CUSUM feasibility analysis. Among 110,183 deduplicated reports (65.8% electrosurgical, 34.2% ultrasonic), electrosurgical devices showed an electrical-thermal fingerprint (burns ROR 3.84; electric shock ROR 15.34) and ultrasonic devices a mechanical-integrity fingerprint (foreign-body removal ROR 92.74), replicated directionally in Canada and Australia. The fingerprint localized to generator/monopolar components and varied by specialty and procedure. Within the robotic validation cohort, component-stratified analysis partly reproduced these signatures-robot-mounted monopolar instruments retained a burn signal, and robot-mounted ultrasonic instruments carried the highest injury-report share-while cross-platform, same-modality contrasts are reported strictly as report-composition (share-level) comparisons. Reporting artifacts (19.5% placeholder narratives; a 2025 robotic mass-reporting episode comprising 55.8% of the robotic cohort) left fingerprint RORs materially unchanged. A retrospectively applied CUSUM flagged the burn theme more than five years before the 2023-2024 Class I recalls, illustrating-without establishing prospective lead time-the potential of automated monitoring. Energy-modality failure-injury fingerprints were consistent across specialties, delivery platforms, three national databases, and regulatory actions. Automated, artifact-aware surveillance of vigilance databases is a promising complement to pre-market review and warrants prospective evaluation.