European cardiovascular registries as reliable data sources to assess implants safety and performances across the implant lifecycle

Abstract Introduction Cardiovascular device registries could be important data sources for manufacturers to provide data required by the recently implemented Medical Device Regulation. However, quality and reliability of device registries for regulatory purposes has not been reported. Purpose To assess European cardiovascular registries on their design and methodology used, to gain insight into the quality of the evidence provided. Methods Seven medical literature libraries were searched to identify regional/national/multi-country European cardiovascular registries capturing data on coronary stents and/or valve repair and/or replacement, following the International Medical Device Regulators Forum definition of a medical device registry. Registries’ annual reports, peer-reviewed publications, and websites were reviewed to extract information on 32 variables in 6 different domains: identification (e.g., geographical representativeness); maturity (e.g., starting year); governance (e.g., funding); coverage, design & organisation (e.g., hospital coverage); data quality (e.g., completeness of procedures); safety & performance (e.g., outlier procedures). Additionally, reported outcomes with their definition and follow-up duration were collected. An active registry was defined as supplying recent data (2018 or beyond). Results 20 registries were identified, of which 11 registries capturing data solely on valve repair and/or replacement, 2 solely on coronary stents, and 7 capturing data on both valves and coronary stents (Table 1). The majority (55%) of included registries were active. Overall, registries reported on a median of 33% of variables (IQR 14-67%), varying from 0% (IQR 0-4%) for the domain safety & performance to 83% (IQR 67-100%) for the domain identification (Figure 1). 6 (30%) registries published annual reports, but most registries publish papers for which selected patient groups are included, so that the total number of patients receiving an implant were reported by only 4 (20%) registries. No registry reported procedure completeness or outlier procedures. The most frequently reported outcome was mortality, reported by 18 (90%) registries, but using 70 different time-points ranging from in-hospital mortality to mortality at 21-years. Most registries (80%) reported 30-day mortality. Major cardiovascular events (MACE) were reported as combined end-points by 8 registries (40%), but with 17 different combinations of complications included and 7 different time intervals; most (33%) registries reported 1-year MACE. This large heterogeneity in outcomes, definitions, and follow-up durations makes comparing outcomes, pooling data, and/or benchmarking implants across registries impossible. Conclusion European cardiovascular registries need better reporting on variables related to the design and methods used as well as agreement on definitions and outcomes used, to achieve their potential as a reliable data source for regulatory purposes.Recent activity of included registriesReported items by registries