Pre-ictal and ictal ventricular repolarization changes in epilepsy: a systematic review

This registration deposits the protocol for a systematic review of ventricular repolarization changes in the pre-ictal and ictal periods in people with epilepsy. The review asks what repolarization changes have been reported for these periods, and how the seizure-related time windows and comparator periods used to identify them are defined across studies. PubMed and Scopus were searched on 10 August 2026, and the PubMed search was re-run without a date limit on 30 October 2026 after a publication-date filter was found to have been unintentionally active in the original run. Eligible studies report a QT/QTc-based, T-wave-based or ST-segment repolarization measure together with at least one seizure-linked temporal comparison (pre-ictal versus interictal, pre-ictal versus ictal, or ictal versus interictal). Synthesis is narrative; no meta-analysis is planned. This registration is RETROSPECTIVE. Registration with PROSPERO was attempted and was not possible, because PROSPERO does not accept an initial registration for a review that has progressed beyond the data extraction stage. At the date of this deposit, searching, screening and data extraction were complete, risk-of-bias assessment had been carried out by both reviewers with consensus in progress, and data synthesis had not begun. The retrospective nature of registration is stated in the attached protocol and will be declared in the Methods section and in the limitations of the resulting manuscript. Methodological decisions taken after the review began, including the correction to the PubMed search, are listed in Section 13 of the attached protocol rather than presented as pre-specified. Two files are attached. The protocol sets out the review question, eligibility criteria, data extraction procedure, risk-of-bias method and synthesis plan. The search strategy file gives the full search strings for both databases, the deduplication method, the correction to the PubMed search, and the automation declaration required by PRISMA 2020 item 8, including which parts of the automated concept filter could be recovered from the preserved record sets and which could not.