Small safety differences among ICD generator models may be unlikely to be identified through routine post-market surveillance, but large safety differences may be detected for most patients at anticipated average adverse event rates, according to a study published June 12 in Pharmacoepidemiology and Drug Safety.

Jonathan Bates, PhD, et al., used data from ACC's ICD Registry to detect safety differences among different ICD generator models, testing adverse event rates. The data included 100 percent of primary prevention ICD implantations and 80 percent of secondary prevention implantations conducted at U.S. hospitals between 2006 and 2010 and reimbursed by Medicare.

According to the results, 141 different ICD models were implanted at least 20 times in a total of 523,529 procedures. For a rate ratio of 1.05, only a small percentage of patients received an ICD model with sufficient utilization to detect safety differences. For a rate ratio of 1.15, 34 percent of ICD models had enough utilization to detect differences at an average event rate of 12.6 per 100 person years. At a rate ratio of 1.25, 30 percent of ICD models had sufficient utilization at 3.9 events per 100 person years, while 50 percent had sufficient use at 12.6 events per person years. For a rate ratio of 1.50, between 52 percent and 67 percent of ICD models had enough use at event rates of 3.9 and 12.6 per 100 person years, respectively, to detect safety differences.  

The researchers conclude that real-time post-market safety surveillance system would detect safety differences among ICD models for about 90 percent of ICD patients if the safety difference was 50 percent or higher or if the safety event rate was 6.1 per 100 person years or greater. They add that the results “should inform the expectations and limitations of real-time active post-market safety surveillance systems.”

Keywords: Pharmacoepidemiology, Secondary Prevention, Primary Prevention, Registries, National Cardiovascular Data Registries, ICD Registry

< Back to Listings