Modeling defibrillation benefit for survival among cardiac resynchronization therapy defibrillator recipients

Abstract

Background: Patients with heart failure having a low expected probability of arrhythmic death may not benefit from implantable cardioverter defibrillators (ICDs).

Objective: The objective was to validate models to identify cardiac resynchronization therapy (CRT) candidates who may not require CRT devices with ICD functionality.

Methods: Heart failure (HF) patients with CRT-Ds and non-CRT ICDs from the National Cardiovascular Data Registry and others with no device from 3 separate registries and 3 heart failure trials were analyzed using multivariable Cox proportional hazards regression for survival with the Seattle Heart Failure Model (SHFM; estimates overall mortality) and the Seattle Proportional Risk Model (SPRM; estimates proportional risk of arrhythmic death).

Results: Among 60,185 patients (age 68.6 ± 11.3 years, 31.9% female) meeting CRT-D criteria, 38,348 had CRT-Ds, 11,389 had non-CRT ICDs, and 10,448 had no device. CRT-D patients had a prominent adjusted survival benefit (HR 0.52, 95% CI 0.50-0.55, P < .0001 versus no device). CRT-D patients with SHFM-predicted 4-year survival ≥81% (median) and a low SPRM-predicted probability of an arrhythmic mode of death ≤42% (median) had an absolute adjusted risk reduction attributable to ICD functionality of just 0.95%/year with the majority of survival benefit (70%) attributable to CRT pacing. In contrast, CRT-D patients with SHFM-predicted survival <median or SPRM >median had substantially more ICD-attributable benefit (absolute risk reduction of 2.6%/year combined; P < .0001).

Conclusions: The SPRM and SHFM identified a quarter of real-world, primary prevention CRT-D patients with minimal benefit from ICD functionality. Further studies to evaluate CRT pacemakers in these low-risk CRT candidates are indicated.

Figure 1 --. Method Used to Calculate Component Relative Risk Reductions.

As described in the text, hazard ratios for survival for CRT-D versus No Device (HR-CRTD) and CRT-D versus a Standard ICD (HR-CRTP) were generated from CRT-D, Standard ICD, and No-Device Cohorts. The hazard ratio for the ICD component of CRT versus No Device (HR-ICD) was estimated based on dividing the two measure hazard ratios. Risk reductions were calculated based on subtracting each hazard ratio from 1. Please also refer to the text. CRT=Cardiac Resynchronization Therapy; ICD=Implantable Cardioverter Defibrillator.

Figure 2 –. CONSORT Diagram.

A flow diagram showing the derivation of the CRT-D, ICD, and No Device groups is shown. CRT=Cardiac Resynchronization Therapy. ICD=Implantable Cardioverter Defibrillator.

Figure 3 –. Survival with CRT-D, Standard ICD, and No Device by SHFM/SPRM Subgroups.

Kaplan-Meier curves adjusted by the SHFM demonstrate survival in groups 1–4 (corresponding to panels A-D) in CRT-D patients. CRT-D=Cardiac Resynchronization Therapy Defibrillator; SHFM=Seattle Heart Failure Model; SPRM=Seattle Proportional Risk Model.

Figure 4 –. Relative Risk Reductions from the CRT Pacing and Defibrillation Components of CRT-D by SHFM and SPRM Quintiles.

The adjusted relative risk reductions attributable to CRT pacing (RESYNCH) and the defibrillation (DEFIB) components of CRT-D by SHFM and SPRM quintiles are shown. CRT=Cardiac Resynchronization Therapy; DEFIB=Defibrillation; ICD=Implantable Cardioverter Defibrillator; RESYNCH=Resynchronization.

Figure 5 –. Absolute and Relative Adjusted Risk Reductions from the ICD Component of CRT-D.

The survival benefit associated with the ICD component of CRT-D increases from risk group 1 to risk group 4 (A), both in terms of adjusted relative risk reduction (B), and adjusted annual absolute risk reduction (C). CRT=D=Cardiac Resynchronization Therapy Defibrillator; ICD=Implantable Cardioverter Defibrillator; RESYNCH=Resynchronization; SHFM=Seattle Heart Failure Mode; SPRM=Seattle Proportional Risk Model.

Figure 6 –. Demonstration of the Effect of SHFM Adjustment in Control Group Patients.

The panels demonstrate the effectiveness of SHFM adjustment for accounting for survival differences in patients on and off ACE inhibitor/ARBs (panels A and B), beta blockers (panels C and D), and registry versus clinical trial subsets of the control group (panels E and F). These were chosen as representative variables with similar results obtained for other covariates.

Central Illustration --. Seattle Proportional Risk Model and Seattle Heart Failure Model for CRT-D Risk Stratification.

The concept of differential benefit from the resynchronization and defibrillation components of a CRT-D is explained. Patients in group 1 are expected to have the lowest benefit from the defibrillator component of CRT and thus derive most of the benefit from CRT-pacing, while group 4 patients derive greater survival benefit from the defibrillator component relative to CRT pacing.