Non-Evidence-Based Defibrillator Implantations? Based on What Evidence?

Bradley P. Knight, MD, FACC, FHRS, Editor-in-Chief
Bradley P. Knight, MD, FACC, FHRS, Editor-in-Chief
Dear Readers, Dr. Al-Khatib and colleagues recently published a study in the Journal of the American Medical Association that received a great deal of attention.1 The title alone is an attention getter: “Non-evidence-based ICD implantations in the United States.” What exactly, if anything, did this study teach us? The authors used the National Cardiovascular Data Registry-(NCDR) ICD Registry to identify patients who underwent implantable cardioverter defibrillator (ICD) implantation for primary prevention, but did not have an “evidence-based indication.” They excluded patients who received a cardiac resynchronization pacing device. The following criteria were used to identify patients who did not have an evidence-based indication:
  1. Myocardial infarction (MI) within 40 days before implantation.
  2. Newly diagnosed heart failure (HF) at the time of ICD implantation.
  3. New York Heart Association (NYHA) functional class IV symptoms.
  4. Coronary artery bypass surgery within 3 months.
They found that 23% of patients underwent ICD implantation for a non-evidence-based indication, mostly because they were coded as having newly diagnosed HF or having had a recent MI. It is important that the heart rhythm community police their own, and identify physicians and hospitals where patients are undergoing ICD implantation unnecessarily. However, there are two major problems with this paper. The most important limitation is that the data entered into the NCDR database is not perfect. At many institutions, the information is gathered by coders and chart abstractors who likely do not have a complete understanding of the patient’s history. There are several examples of mistakes that could occur, ranging from a simple coding error to a situation where a patient is admitted with NYHA class IV symptoms, and codes as such, but improves clinically during the hospitalization and has class II symptoms at the time of ICD implantation. One could argue that the physician should complete this form to improve its accuracy, and this is done at some hospitals. However, the length of the data collection form and the time required to complete it leads to opposition by most physicians. At the minimum, however, an implanting physician at each institution should review key data for quality and accuracy before quarterly submission to the NCDR. It would be reasonable to review any cases where a complication has been noted, or where a patient has been identified as having one of the non-evidence-based indications. A second, and very different, limitation is that many if not all patients who undergo ICD implantation have a legitimate indication that is not captured in the guidelines. The ‘guidelines’ are merely ‘guidelines,’ and there are times when ICD implantation is clinically appropriate outside of the guidelines. The following could be considered as examples of legitimate indications for an ICD:
  1. A patient with heart failure and an LVEF of 20% who has complete heart block after CABG and valve replacement and needs pacing.
  2. A patient with severe class IV heart failure and a recently diagnosed dilated cardiomyopathy who is listed from transplantation and the ICD is implanted as a ‘bridge to transplant.’
  3. A patient admitted with a nonischemic cardiomyopathy and chronic heart failure who is admitted with an acute episode of heart failure and a mild troponin elevation, resulting in a diagnosis of myocardial infarction in the chart.
Continued efforts are needed to make sure that physicians who are implanting ICDs are doing what is in the best interest of the patient. However, using the NCDR ICD database to conclude that a quarter of patients are receiving an ICD without an evidence-based indication without being certain that the data is accurate, and without determining what the actual indication was for each implant, is counterproductive to legitimate efforts to reach the thousands of patients who have an indication for ICD therapy, but who are not being treated.


  1. Al-Khatib SM, Hellkamp A, Curtis J, et al. Non-evidence-based ICD implantations in the United States. JAMA 2011;305:43-49.