Circ: Automated safety surveillance tools may help spot faulty devices sooner
Automated safety surveillance tools, such as one tested by researchers at the Minneapolis Heart Institute Foundation, may be able to identify poor performing cardiac devices earlier than the existing postmarket surveillance methods used in clinical practice, according to a study published March 6 in Circulation: Cardiovascular Quality and Outcomes. This usage could help expose fewer patients to potentially hazardous devices.

“The goal of postmarket surveillance is to enhance the public health by reducing the incidence of medical device adverse experiences," Robert G. Hauser, MD, of the Minneapolis Heart Institute Foundation, and colleagues wrote. “However, medical device surveillance in the United States depends largely on voluntary reporting of adverse events, and, consequently, early safety signals may be missed.”

Because postmarket medical device surveillance may miss early device safety signals, Hauser and colleagues conducted a study to evaluate the feasibility of using of an automated safety surveillance tool—DELTA—to detect the early signals that an implantable cardiac device was underperforming.

Using a commercially available automated surveillance tool (DELTA, Coping Systems), the researchers performed a simulated prospective monthly cohort and propensity-matched comparative analyses on a three-center database of Sprint Fidelis and Quattro Secure (both Medtronic) implantable cardioverter-defibrillator (ICD) leads. Patients who received leads at Minneapolis Heart Institute, Mayo Clinic in Rochester, Minn., and Beth Israel Deaconess Medical Center in Boston, between November 2001 and December 2008 were included. The surveillance tool was set to trigger device alerts if the log rank probability dipped below 0.05.

Metronic suspended the worldwide marketing of its Sprint Fidelis platform of defibrillation leads due to the potential for lead fractures in October 2007.

Hauser and colleagues found that 84 of the 1,035 Fidelis leads and 23 of the 1,675 Quattro leads failed during the study period. “The simulated full-cohort analysis triggered a sustained alert for Fidelis leads beginning 13 months after the first implant and two years before Fidelis leads were removed from the market,” the authors wrote.

Additionally, the authors noted that of the 1,035 patients implanted with Fidelis leads, 93.6 percent were successfully matched to Quattro patients. During the propensity-matched analysis, the alert triggered 22 months post-Fidelis lead implant and more than one-year prior to the lead recall.

By January 2005, the authors noted that 95 Fidelis implants had occurred without failure, and 848 Quattro leads were implanted with a single failure at 18 months.

The first DELTA alert within the full-cohort unmatched simulation occurred in October 2005, 13 months after the Fidelis lead was released on the market. At this time, the authors noted that there was one Fidelis lead failure among 419 device implants. All subsequent monthly unmatched survival analyses triggered a DELTA alert, leaving the authors to speculate that by early 2007, Fidelis leads would have either been removed from the market or infrequently used.

"This proactive approach to postmarket surveillance has the key attributes of timeliness and scientific rigor while minimizing the chance that clinical or operator variables could account for the observed failures," the authors wrote.

During the simulated model, the authors found that the DELTA alert triggered in July 2006 and persisted until January 2007. Between the 4th quarter of 2005 and the first five months of 2006, DELTA alerts occurred in 12 percent to 26 percent of simulations. In June 2006, DELTA alerts triggered in more than 60 percent of simulations except for December 2006 and January 2007, when the alert occurred in 43 percent of patients and 53 percent of patients, respectively.

“The results of this multicenter study suggest that an active automated surveillance system could have identified this cardiovascular device problem substantially sooner than was achieved through existing postmarket surveillance methods,” the authors wrote.

Initial reports of Fidelis lead failure surfaced in July 2007; however, it wasn't until Oct. 15, 2007, that the manufacturers decided to pull the device from the market. By that time, 268,000 Fidelis leads were implanted worldwide. The current study authors added that an alert triggered in 2006 or early 2007 could have "spared thousands of patients the consequences of receiving a device prone to failure, including the risks and costs of lead replacement.

"If our simulation is supported by prospective studies using active databases, a significant gap in the postmarket surveillance of medical devices can be closed by the application of tools such as DELTA to national clinical registries and data warehouses," the authors noted.

"This observation suggests that the application of such tools to clinical registries and databases may limit the exposure of large patient populations to potentially hazardous devices," Hauser et al summed.