Implantable Heart Defibrillators Too Widely Used in High-Risk Patients
More than one in five patients who receive implantable cardioverter-defibrillators (ICDs) do not meet evidence-based guidelines to receive them, significantly increasing their risk of complications and death, according to a new study published in the Journal of the American Medical Association and led by the Duke Clinical Research Institute.
Of the 111,707 patients who received ICD implants between January 2006 and June 2009, 25,145 (22.5 percent) did not meet evidenced-based criteria to receive the devices. The study is the first to examine in-hospital outcomes of patients receiving non-evidenced-based ICDs.
“During this time of increased emphasis on quality improvement by promoting evidence-based care, it is important to assess hospital performance and evaluate compliance with clinical guideline recommendations,” said Sana Al-Khatib, MD, the study’s lead investigator and associate professor of medicine at Duke.
“While clear guidelines are established, the rates of non-evidence-based ICD use are much higher than we expected, which had a direct impact on patient outcomes and overall health care use, including length of hospital stay and costs.”
ICDs are small generators designed to deliver an electrical jolt if potentially life-threatening heart rhythms are detected. Previous studies have shown ICDs are effective at preventing sudden cardiac death in patients with advanced heart failure.
Practice guidelines do not recommend use of a primary prevention ICD among patients who have recently had a heart attack or coronary artery bypass graft surgery and also advise against use for people with severe heart failure symptoms or a recent diagnosis of heart failure.
In 2005, the Centers for Medicare & Medicaid Services announced expanded coverage for ICD implantation for the primary prevention of sudden cardiac death. The Duke team sought to determine the rates at which physicians in clinical practice follow clinical practice recommendations.
They conducted an analysis of cases submitted to the National Cardiovascular Data Registry-ICD Registry, which was created by the Heart Rhythm Society and American College of Cardiology to track ICD implantations nationwide.
The researchers found 9,257 patients (37 percent) were within 40 days of a heart attack and 15,604 (62 percent) were in patients with newly diagnosed heart failure.
The risk of in-hospital death was significantly higher in patients who received a non-evidence-based device than in patients who received an evidence-based device (0.57 percent vs. 0.18 percent). The risk of any postprocedure complication also was significantly higher in the non-evidence-based group compared to the evidence-based group (3.23 percent vs. 2.41).
“While we observed modest differences when evaluating complication rates between the groups, the events that occurred in the non-evidenced based group may have been avoided since the procedures were clearly not indicated in that patient population,” Al-Khatib said. “While all procedures are associated with some risk, no degree of risk is acceptable if a procedure has no demonstrated benefit.”
Al-Khatib said one excess complication occurred for every 121 non-evidenced-based ICD nationwide.
The median length of hospital stay was significantly longer for patients who received a non-evidence-based ICD compared with patients who received an evidence-based ICD (three days vs. one day).
Researchers observed substantial hospital variation in non-evidence-based ICDs, comprising more than 40 percent of total implanted ICDs at some hospitals, and there was no clear decrease in the rate of use over time.
“These findings highlight the importance of continuing to enhance health care practitioner’s understanding of practice guidelines,” Al-Khatib said.
The study was funded by a grant from the National Heart Lung and Blood Institute.
Study co-authors include Anne Hellkamp, Jeptha Curtis, Daniel Mark, Eric Peterson, Gillian Sanders, Paul Heidenreich, Adrian Hernandez, Lesley Curtis, and Stephen Hammill.