|| Checking for direct PDF access through Ovid
Importance: There is increasing emphasis on getting medical devices to market more quickly by use of priority review pathways, promoted in recent US Food & Drug Administration (FDA) guidance documents and the 21st Century Cures Act of 2016.Objective: To evaluate the impact of priority review pathways for high-risk cardiovascular devices by analyzing the quality and characteristics of premarket clinical studies in support of their approval and the frequency of postmarket recalls.Design and Setting: Descriptive study of premarket clinical studies supporting all high-risk cardiovascular devices approved via the priority review premarket approval (PMA) pathway between January 1, 2005 and June 15, 2017.Exposure: Priority Review PMA Main Outcome(s) and Measure(s): For all clinical studies which supported priority review PMAs we analyzed: number of patients enrolled, number of patients analyzed for each primary endpoint, demographic information, randomization, blinding, use and type of controls, use of training patients, number and location of study sites, and time of FDA review to approval. We also determined if a post hoc analysis was conducted. We counted all Class 1 and Class 2 recalls and examined FDA expert device panel recommendations.Results: There were 16 priority review cardiovascular devices approved, with mean 1.4 (SD 0.6, range 1-3) pivotal studies per device. Eleven of 16 (69%) were approved based on a single pivotal study. Of 23 pivotal studies, ten (43%) were randomized, 3 (13%) were single-blinded, and none were double-blinded. Twelve (52%) used a noninferiority primary endpoint hypothesis for at least one endpoint and 18 (78%) used a surrogate endpoint. Of the 11 votes on device effectiveness by FDA expert panels, 6 (55%) voted devices were effective. There were a total of 18 Class 1 and 2 recalls for 7 of these devices.Conclusions and Relevance: Most priority review pathway high-risk cardiovascular devices are approved without a randomized or blinded trial. Noninferiority design and use of surrogate endpoints are common. These devices are frequently subject to recalls.