CorVue algorithm efficacy to predict heart failure in real life: Unnecessary and potentially misleading information?

Abstract

Heart failure (HF) hospitalizations have a negative impact on quality of life and imply important costs. Intrathoracic impedance (ITI) variations detected by cardiac devices have been hypothesized to predict HF hospitalizations. Although Optivol™ algorithm (Medtronic, Minneapolis, MN, USA) has been widely studied, CorVue™ algorithm's (St. Jude Medical, St. Paul, MN, USA) long-term efficacy has not been systematically evaluated in a "real-life" cohort. CorVue™ was activated in implantable cardioverter defibrillator (ICD)/cardiac resynchronization therapy defibrillator (CRT-D) patients to store information about ITI measures. Clinical events (new episodes of HF requiring treatment and hospitalizations) and CorVue™ data were recorded every 3 months. Appropriate CorVue™ detection for HF was considered if it occurred in the 4 prior weeks to the clinical event. Fifty-three ICD/CRT-D (26 ICD and 27 CRT-D) patients (67 ± 1 years old, 79% male) were included. Device position was subcutaneous in 28 patients. At inclusion, mean left ventricular ejection fraction was 25 ± 7% and 27 patients (51%) were in New York Heart Association class I, 18 (34%) in class II, and eight (15%) in class III. After a mean follow-up of 17 ± 9 months, 105 ITI drops alarms were detected in 32 patients (60%). Only six alarms were appropriate (true positive) and required hospitalization. Eighteen patients (34%) presented 25 clinical episodes (12 hospitalizations and 13 emergency room/ambulatory treatment modifications). Nineteen of these clinical episodes (76%) remained undetected by the CorVue™ (false negative). Sensitivity of CorVue™ resulted in 24%, specificity was 70%, positive predictive value of 6%, and negative predictive value of 93%. CorVue™ showed a low sensitivity to predict HF events. Therefore, routinely activation of this algorithm could generate misleading information.