Abstract 11692: First Evaluation of Corvue Algorithm Efficacy to Predict Heart Failure in Real Life: An Unnecessary Waste of Energy?

Introduction and hypothesis: Repeated hospitalizations for heart failure (HF) have a negative effect on quality of life and imply an important cost for the health care system. A drop of intrathoracic impedance due to pulmonary congestion detected by cardiac devices has been hypothesized to decrease hospitalization rate. Although Optivol algorithm (Medtronic) has been widely studied, we present the first study that evaluates the efficacy of CorVue algorithm (St. Jude Medical) to predict HF in a “real-life” cohort. Methods: CorVue algorithm (St Jude Medical) was activated in ICD or CRT-D patients to store information about intrathoracic impedance changes using nominal settings (with patient’s alarm emission off). Patients were followed every 3 months. Clinical events (new episodes of HF requiring ambulatory or emergency room (ER) treatment modification or hospitalization) and CorVue data were recorded. Appropriate CorVue detection for HF was considered if it occurred in the 4 prior weeks to the clinical event. Results: Sixty-five ICD/CRT-D patients (70 ± 1 year-old, 78% male) were included (36 ICD (55%) and 29 CRT-D patients). Device position was subcutaneous in 43 patients (69%) and submuscular in 19 cases. Time from implantation to algorithm activation was 21 ± 2 months. At inclusion, mean LVEF was 33.8 ± 1.7 % and 38 patients (59%) were in NYHA class I, 20 (31%) class II and 7 (11%) class III. After a mean follow-up of 14 ± 1 months, 51 patients (78%) remained in a clinically stable condition and 14 patients (21%) presented 20 episodes of HF: 9 hospitalizations and 11 ER/ambulatory treatment modification. On the other hand, a total of 83 CorVue algorithm alarm episodes were registered in 32 patients (49%), with a mean number of alarms per patient of 2.6 ± 0.2 (range 1 - 6). Only 5 alarm episodes (6%) resulted to be appropriate: 2 ambulatory diuretic therapy intensification and 3 hospitalizations. The sensitivity of CorVue algorithm was 25%. None of the variables analyzed to determine a better scenario for this algorithm improved significantly the diagnostic efficacy. Conclusion: CorVue algorithm (St Jude Medical) resulted to have a low sensitivity for the prediction of HF and its routinary activation could generate misleading information.