Insertable cardiac monitors after hf exacerbation: a target-trial emulation study
M Goldstein, R Kahwash, D Lanctin, S Sarkar, B Gerritse, J W DukesAbstract
Background
The ALLEVIATE-HF trial found high rates of arrhythmia detection and arrhythmia-directed interventions after insertable cardiac monitor (ICM) insertion in patients with heart failure (HF). However, real-world generalizability, the share of interventions attributable to ICM use, and impact on HF outcomes remain uncertain.
Methods
We emulated a target trial using United States Medicare claims comparing ICM insertion vs no ICM within 6 months after first HF exacerbation (index). A marginal structural model with inverse-probability of treatment weighting (IPTW) used demographic, comorbidity, and utilization data to address baseline and time-varying confounding during the 6-month initiation period. ICM status was time-varying during the initiation period and fixed thereafter; patients receiving ICM after the initiation period were censored at insertion. Outcomes included arrhythmia-directed intervention (pacemaker, implantable cardioverter-defibrillator, cardiac ablation, cardioversion, or left atrial appendage occlusion/exclusion), HF exacerbation or death, and HF exacerbation rates. Weighted Fine-Gray, Cox, and negative binomial regression models were used. Prespecified falsification endpoints (cataract extraction, urinary tract infection [UTI], community-acquired pneumonia [CAP]) assessed residual confounding. To assess trial generalizability, cumulative incidence of arrhythmia-directed interventions in real-world ICM recipients was compared with linked an ALLEVIATE-HF subset with Medicare enrollment.
Results
In the target trial, 1,040,258 patients met inclusion criteria; 2,818 initiated ICM within 6 months and 1,037,440 did not. While ICM recipients were younger and had more atrial fibrillation, IPTW achieved robust covariate balance (all standardized mean differences < 0.10). ICM was associated with higher arrhythmia-directed intervention (weighted HR: 2.7; 95% CI: 2.5-3.1; Figure 1a), with increases for all intervention types (Figure 1b-c). ICM was also associated with lower weighted risk of HF exacerbation or death (HR: 0.82; 95% CI; 0.77–0.88, Figure 2a) and lower HF exacerbation rates (IRR: 0.84, 95% CI: 0.75-0.94), driven by fewer inpatient events (IRR: 0.81; 95% CI: 0.72-0.91; Figure 2b). No association was observed with falsification endpoints: cataract extraction (HR: 1.00; 95% CI: 0.84-1.19); UTI (HR: 1.21; 95% CI: 0.87-1.67); CAP (HR: 0.99; 95% CI: 0.79-1.24). Among ICM recipients, cumulative incidence of arrhythmia-directed intervention was similar in real-world and trial cohorts at 3, 12, and 24 months (real-world: 15.9%, 30.1%, 37.3%; trial: 15.7%, 31.7%, 36.3%; HR=0.97; 95% CI: 0.78-1.37).
Conclusion
In real-world target-trial emulation, ICM insertion after HF exacerbation was associated with substantially higher arrhythmia-directed interventions and improved HF outcomes. Similar intervention rates between trial and real-world cohorts support generalizability of ALLEVIATE-HF to routine practice.