791 Optimal Timing of Implantable Cardioverter-Defibrillator Implantation After Myocardial Infarction: A Decision Analysis JONATHAN P. PICCINI, M.D., M.H.S., ∗ SANA M. AL-KHATIB, M.D., M.H.S., ∗ EVAN R. MYERS, M.D., M.P.H., ∗ KEVIN J. ANSTROM, Ph.D., ∗ ALFRED E. BUXTON, M.D.,† ERIC D. PETERSON, M.D., M.P.H., ∗ and GILLIAN D. SANDERS, Ph.D. ∗ From the ∗ Duke Clinical Research Institute and the Duke University School of Medicine, Durham, North Carolina, USA; and †Brown Medical School and Lifespan Academic Medical Center, Providence, Rhode Island, USA ICD Implant Timing. Background: The optimal timing of implantable cardioverter defibrillator (ICD) placement for the primary prevention of sudden cardiac death after myocardial infarction (MI) remains unknown. Methods and Results: We developed a Markov model to investigate the optimal timing of ICD implantation after MI (no ICD, ICD at 60 days, 6 months, and 1 year) in patients who meet current guidelines. Estimates of arrhythmic death (baseline risk 6%, range 1–20% per year), nonarrhythmic death, and ICD efficacy were based upon MADIT-II and other contemporary post-MI clinical trials. We used both deterministic and stochastic modeling processes in our analysis. After 10 years follow-up, the baseline probability of survival was higher in those treated with ICD implantation versus not (42% vs 30%, P < 0.001). Survival was highest with ICD implantation at 60 days versus 6 months versus 1 year: 42.4%, 42.3%, and 42.0% (P = 0.0028). ICD implantation at 60 days provided a mean incremental survival of 0.28 months and 0.84 months per patient (compared with implantation at 6 months and 1 year). In sensitivity analyses, patients’ competing risk for nonarrhythmic death was the primary determinant of benefit from ICD implantation. Overall, ICD implantation at 60 days resulted in the greatest life expectancy over a wide range of plausible nonarrhythmic and arrhythmic death rates. Conclusions: The benefits of early ICD implantation are modest when compared with delayed implanta- tion at 6 months/1 year. Our results suggest that making sure a patient receives an ICD, when appropriate, may be more important than the timing of the implantation procedure. (J Cardiovasc Electrophysiol, Vol. 21, pp. 791-798, July 2010) implantable cardioverter defibrillator, sudden cardiac death, myocardial infarction, decision analysis, congestive heart failure Introduction Despite improvements in the treatment of coronary artery disease, more than half of all cardiovascular mortality is due to arrhythmic death. 1 The implantable cardioverter defibrilla- tor (ICD) has been shown to improve survival in patients with significant left ventricular dysfunction following myocardial infarction (MI). 2,3 Accordingly, national guidelines advocate ICD implantation for the primary prevention of sudden car- diac death (SCD) in patients with a left ventricular ejection fraction (LVEF) less than 30% after MI (class I, Level of Ev- idence: A). 4 Consistent with these guidelines, between 2000 Dr. Piccini is supported by an American College of Cardiology Founda- tion/Merck Award, serves on an advisory board to Medtronic, and receives research funding from Boston Scientific. Dr. Al-Khatib receives research funding and speaking fees from Medtronic. Dr. Buxton reports honorarium for talk on appropriate use of ICD. Dr. Sanders participated in a research project funded by Medtronic. No other disclosures. Address for correspondence: Jonathan P. Piccini, M.D., M.H.S., Division of Cardiology, Duke Clinical Research Institute, Duke University Medical Center, DUMC #3115, Durham, NC 27705. Fax: 919-668-7058; E-mail: jonathan.piccini@duke.edu Manuscript received 3 August 2009; Revised manuscript received 23 November 2009; Accepted for publication 24 November 2009. doi: 10.1111/j.1540-8167.2009.01696.x and 2006, the number of ICDs implanted in the US doubled, from 32,000 to 64,000 per year. 5 The risk of SCD is greatest immediately after MI (as high as 14–24% in the first month after MI) before declining and reaching a plateau at 1 year. 6-10 Paradoxically, ICD implan- tation in the first 40 days after MI has not led to improved survival, despite reductions in arrhythmic death. 11,12 Addi- tionally, retrospective studies have suggested that ICD ther- apy may be of greatest benefit remotely after MI. 13 Given these conflicting data, the optimal timing of ICD implan- tation after MI remains unknown. 14 Since more than one million Americans suffer an MI each year, and 600 ICDs are implanted each day in the United States, 15 the question of optimal utilization of ICD therapy after MI is of vital importance to the public health. In order to ascertain the optimal timing of ICD implanta- tion after MI, we developed a Markov model to (1) assess the relative mortality benefits of early versus late ICD implanta- tion after MI and (2) to identify important factors that might influence these treatment strategies. Methods Decision Model The current American College of Cardiology/American Heart Association/Heart Rhythm Society guidelines advo- cate ICD implantation for primary prevention of SCD in