Risks of ICD Implantation

Abstract & Commentary

By John P. DiMarco, MD, PhD, Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville

Source: Haines DE, et al. Implantable cardioverter-defibrillator registry risk score models for acute procedural complications or death after implantable cardioverter-defibrillator implantation. Circulation 2011;123:2069-2076.

In this paper, Haines and his coauthors propose a scoring system to predict risks associated with implantable cardioverterdefibrillator (ICD) implant procedures. The authors analyzed data from the national ICD Registry which tracks in-hospital complications after most ICD procedures performed in the United States. They analyzed data reported between January 1, 2006, and June 30, 2008. They identified 29 variables captured by the registry and tested them for inclusion in a risk model. This was accomplished using multivariate logistic regression analysis on the 29 selected variables with the endpoints of inhospital complications or in-hospital mortality. Once risk factors were identified, the risk score model was re-evaluated in a validation cohort.

The authors excluded 5012 implants that required thoracotomy for lead placement, leaving a final study cohort of 268,701 implants. In these patients, there were 8559 acute procedurerelated complications (3.2%). The most common complications were: hematoma (0.93%), lead dislodgement (0.93%), pneumothorax (0.42%), and cardiac arrest (0.03%). Other complications noted at lower frequencies included: cardiac perforation, myocardial infarction, pericardial tamponade, and infection. The in-hospital mortality rate was 0.38%. Patients who were admitted for reasons other than elective ICD implants had a significantly higher risk for both any complication or death (4.58% vs 2.51%, P < 0.0001) and of in-hospital death alone (0.91% vs 0.12%, P < 0.0001). Patients were more likely to suffer complications or death if they were older, female, had a lower left ventricular ejection fraction, or had a prolonged QRS duration on ECG. Complications were more common in those with nonischemic dilated cardiomyopathy, congestive heart failure, atrial fibrillation or flutter, prior valve surgery, cerebral vascular disease, chronic kidney disease, and chronic lung disease. The complexity of the procedure was also associated with an increased complication rate with the lowest rates seen with simple device changes for generator replacement and single-chamber ICDs and higher rates with dualchamber or biventricular devices. Ten variables were strongly associated with increased risk for complications and were assigned points in the scoring system. These included: age older than 70 years (1 point); female gender (2 points); New York Heart Association heart failure class III (1 point) or IV (3 points); atrial fibrillation or flutter (1 point); prior valve surgery (3 points); dual chamber (2 points) or biventricular (4 points) devices; chronic lung disease (2 points); BUN greater than 30 mg/dL (2 points); previous ICD reimplantation (6 points); and non-elective ICD implantation (3 points). The risk for any in-hospital complication increased from 0.6% with a total risk score < 5 (8.4% of the total population) to 8.4% among patients with a risk score of > 19. There was an almost linear increase in rate of complications as the risk score rose. A similar model using seven factors was constructed for in-hospital mortality.

The authors conclude that the risk for acute ICD complications can be predicted using a simple scoring system. For patients with the highest risk for complications, knowledge of this increased risk might change the risk-benefit equation leading to device implantation.


This is a useful paper for both implanting physicians and those who refer patients for ICD implants. For implanting physicians, this scoring system allows them to risk-adjust their own or their institution's performance against a national standard. For referring physicians, knowledge of the risk for complications will allow them to be more transparent in their discussions with patients about the risks and benefits of the procedure. We also must recognize that the data here refer to only in-hospital complications and are a lower estimate of the actual risks. Elective procedures are usually performed on an overnight stay basis. This, at least partly, explains why complications are lower with elective implants in patients admitted solely for the procedure. We also need to remember that some complications often typically occur outside the 1-2 day window often reported in the ICD registry. Pocket infections, lead dislodgements, and lead malfunctions, particularly with LV leads, fall into this early after discharge category.

ICDs are important tools for prevention of sudden death. However, we need to be frank with both ourselves and our patients that complications with these devices are common and the risk of complications should be part of any discussion with patients concerning an ICD procedure.