Clinical outcomes of automatic algorithms in cardiac resynchronization therapy: Systematic review and meta-analysis

Leonardo Knijnik; Bo Wang; Rhanderson Cardoso; Colby Shanafelt; Michael S Lloyd

doi:10.1016/j.hroo.2023.09.001

Clinical outcomes of automatic algorithms in cardiac resynchronization therapy: Systematic review and meta-analysis

Heart Rhythm O2. 2023 Sep 13;4(10):618-624. doi: 10.1016/j.hroo.2023.09.001. eCollection 2023 Oct.

Authors

Leonardo Knijnik¹, Bo Wang¹, Rhanderson Cardoso², Colby Shanafelt¹, Michael S Lloyd¹

Affiliations

¹ Emory University Adult Congenital Heart Center, Atlanta, Georgia.
² Heart and Vascular Center, Brigham and Women's Hospital, Boston Massachusetts.

Abstract

Background: Algorithms to automatically adjust atrioventricular (AV) and interventricular (VV) intervals in cardiac resynchronization therapy (CRT) devices are common, but their clinical efficacy is unknown.

Objective: The purpose of this study was to evaluate automatic CRT algorithms in patients with heart failure for the reduction of mortality, heart failure hospitalizations, and clinical improvement.

Methods: We performed a systematic review and meta-analysis of randomized controlled trials (RCTs) in patients with CRT using automatic algorithms that change AV and VV intervals dynamically without manual input, on a beat-to-beat basis. We performed a subgroup analysis including intracardiac electrogram-based (EGM) algorithms and contractility-based algorithms.

Results: Nine RCTs with 8531 participants were included, of whom 4275 (50.1%) were randomized to automatic algorithm. Seven of the 9 trials used EGM-based algorithms, and 2 used contractility sensors. There was no difference in all-cause mortality (10.3% vs 11.3%; odds ratio [OR] 0.90; 95% confidence interval [CI] 0.71-1.03; P = .13; I² = 0%) or heart failure hospitalizations (15.0% vs 16.1%; OR 0.924; 95% CI 0.81-1.04; P = .194; I² = 0%) between the automatic algorithm group and the control group. Study-defined clinical improvement was also not significantly different between groups (66.6% vs 63.3%; risk ratio 1.01; 95% CI 0.95-1.06; P = .82; I² = 50%). In the contractility-based subgroup, there was a trend toward greater clinical improvement with the use of the automatic algorithm (75% vs 68.3%; OR 1.45; 95% CI 0.97-2.18; P = .07; I² = 40%), which did not reach statistical significance. The overall risk of bias was low.

Conclusion: Automatic algorithms that change AV or VV intervals did not improve mortality, heart failure hospitalizations, or cardiovascular symptoms in patients with heart failure and CRT.

Keywords: Algorithm; Cardiac resynchronization therapy; Heart failure; Optimization; Pacing.