Comparative Effectiveness of Implantable Defibrillators for Asymptomatic Brugada Syndrome: A Decision-Analytic Model

Abstract

Background Optimal management of asymptomatic Brugada syndrome (BrS) with spontaneous type I electrocardiographic pattern is uncertain. Methods and Results We developed an individual-level simulation comprising 2 000 000 average-risk individuals with asymptomatic BrS and spontaneous type I electrocardiographic pattern. We compared (1) observation, (2) electrophysiologic study (EPS)-guided implantable cardioverter-defibrillator (ICD), and (3) upfront ICD, each using either subcutaneous or transvenous ICD, resulting in 6 strategies tested. The primary outcome was quality-adjusted life years (QALYs), with cardiac deaths (arrest or procedural-related) as a secondary outcome. We varied BrS diagnosis age and underlying arrest rate. We assessed cost-effectiveness at $100 000/QALY. Compared with observation, EPS-guided subcutaneous ICD resulted in 0.35 QALY gain/individual and 4130 cardiac deaths avoided/100 000 individuals, and EPS-guided transvenous ICD resulted in 0.26 QALY gain and 3390 cardiac deaths avoided. Compared with observation, upfront ICD reduced cardiac deaths by a greater margin (subcutaneous ICD, 8950; transvenous ICD, 6050), but only subcutaneous ICD improved QALYs (subcutaneous ICD, 0.25 QALY gain; transvenous ICD, 0.01 QALY loss), and complications were higher. ICD-based strategies were more effective at younger ages and higher arrest rates (eg, using subcutaneous devices, upfront ICD was the most effective strategy at ages 20-39.4 years and arrest rates >1.37%/year; EPS-guided ICD was the most effective strategy at ages 39.5-51.3 years and arrest rates 0.47%-1.37%/year, and observation was the most effective strategy at ages >51.3 years and arrest rates <0.47%/year). EPS-guided subcutaneous ICD was cost-effective ($80 508/QALY). Conclusions Device-based approaches (with or without EPS risk stratification) can be more effective than observation among selected patients with asymptomatic BrS. BrS management should be tailored to patient characteristics.

Keywords: Brugada syndrome; cost‐effectiveness; implantable defibrillator.

Figure 1. Overview of model structure.

Depicted is an overview of model structure. Health states comprising each of the 3 strategies modeled (observation [with ICD for survived arrest], electrophysiologic study [with ICD for positive findings], and upfront ICD) are depicted from top to bottom. Initiation states (states at which simulation starts) are depicted in green, standard states in orange, and terminal states (states where individuals must remain) in gray. Events mediating transitions between states are depicted with colored arrows. Additional events occurring within health states are depicted in the box. EPS indicates electrophysiologic study; and ICD, implantable cardioverter‐defibrillator.

Figure 2. Effects of diagnosis age and initial rate of cardiac arrest on clinical effectiveness.

Depicted is the effect of varying Brugada syndrome diagnosis age (top panels) and initial yearly arrest rate (bottom panels) on the clinical effectiveness of: observation (gray), electrophysiologic study (orange), and upfront implantable cardioverter‐defibrillator (green) strategies, using either subcutaneous (left panels) or transvenous (right panels) devices. The y‐axis depicts total quality‐adjusted life years lived. For the initial rate of arrest, the x‐axis depicts the baseline rate at the start of simulation before applying exponential decay (see text). BrS indicates Brugada syndrome; EPS, electrophysiologic study; and ICD, implantable cardioverter‐defibrillator.

Figure 3. Optimal strategy as a function of electrophysiologic study risk stratification efficacy and implantable cardioverter‐defibrillator utility.

Depicted are the results of 2‐way sensitivity analyses across varying efficacy of electrophysiologic study for risk stratification (rows) vs utility of implantable cardioverter‐defibrillator (columns), for models including subcutaneous devices (upper panels) and transvenous devices (lower panels). In each plot, the optimal strategy for each set of conditions (green: observe, yellow: EPS, red: upfront implantable cardioverter‐defibrillator) is depicted in each box, where optimal is defined as the strategy maximizing quality‐adjusted life years (effectiveness tables), or the most effective strategy having an incremental cost‐effectiveness ratio under the willingness‐to‐pay threshold of $100 000 per quality‐adjusted life years (cost‐effectiveness tables). Parameters representing the base case scenario (implantable cardioverter‐defibrillator utility 0.95 and relative risk of arrest after positive electrophysiologic study 1.7) are starred. EPS indicates electrophysiologic study; and ICD, implantable cardioverter‐defibrillator.

Figure 4. Clinical and cost‐effectiveness in probabilistic sensitivity analyses.

Depicted are the results of 1000 runs of probabilistic sensitivity analysis, which estimates the effects of parameter uncertainty on clinical and cost‐effectiveness estimates. A, The proportion of times each strategy resulted in the greatest overall clinical effectiveness (ie, highest quality‐adjusted life years). B, A cost‐effectiveness acceptability curve, which depicts the probability that each strategy is the most cost‐effective option across increasing willingness‐to‐pay (x‐axis). C, A cost‐effectiveness acceptability frontier, which depicts the preferred strategy (by color) and its probability of cost‐effectiveness across increasing willingness‐to‐pay (x‐axis). For (B and C), the willingness‐to‐pay threshold of $100 000/quality‐adjusted life year used to define cost‐effectiveness in this study is depicted by the vertical hashed line. EPS indicates electrophysiologic study; and ICD, implantable cardioverter‐defibrillator.