Abstract: The invention relates to an active implantable pacemaker. The device analyzes a ventricular electrogram signal (EGM) and is able to recognize, in a search window, an EA4 component of endocardial acceleration (EA) associated with atrial activity. In the presence of atrioventricular conduction, the search window is determined based on the temporal position of the EA1 and/or EA2 components of the EA signal. In the absence of atrioventricular conduction, a delay is counted from a paced ventricular event and applied to mask the EA1 and/or EA2 components in the EA signal, and the window for research of the EA4 component follows the masking delay. In the presence of a confirmed EA4 component, an atrioventricular delay is applied, counted from the EA4 component, and in the opposite case a predetermined escape interval is applied, counted from the last stimulated ventricular event.

Abstract: A cardiac implant includes a classifier configured analyze the data for comparison to thresholds prioritized according to a Boolean decision tree. The implant can generate an indicator of alert status (i.e., alert or no alert). A supervisory device operated, for example, by a doctor can associate each indicator a marker (AE/no AE) indicating the presence or absence of an observed adverse event. In the presence of a false positive, a command for update of the thresholds of the decision tree is transmitted to the implant. A database of reference patients can be used to recreate or further update the decision tree to avoid the occurrence of false negatives.

Abstract: Methods, devices, and processor-readable storage media are provided for the diagnosis of heart failure. One method includes collecting, using an implantable device, reference episodes; generating an in-suspicion model-cycle and an off-suspicion model-cycle based on the reference episodes; and determining whether to generate a heart failure alert, based on a difference between the in-suspicion model-cycle and the off-suspicion model-cycle.

Abstract: The invention relates to an active implantable pacemaker. The device analyzes a ventricular electrogram signal (EGM) and is able to recognize, in a search window, an EA4 component of endocardial acceleration (EA) associated with atrial activity. In the presence of atrioventricular conduction, the search window is determined based on the temporal position of the EA1 and/or EA2 components of the EA signal. In the absence of atrioventricular conduction, a delay is counted from a paced ventricular event and applied to mask the EA1 and/or EA2 components in the EA signal, and the window for research of the EA4 component follows the masking delay. In the presence of a confirmed EA4 component, an atrioventricular delay is applied, counted from the EA4 component, and in the opposite case a predetermined escape interval is applied, counted from the last stimulated ventricular event.

Abstract: Methods, devices, and processor-readable storage media are provided for the diagnosis of heart failure. One method includes collecting, using an implantable device, reference episodes; generating an in-suspicion model-cycle and an off-suspicion model-cycle based on the reference episodes; and determining whether to generate a heart failure alert, based on a difference between the in-suspicion model-cycle and the off-suspicion model-cycle.

Abstract: Methods, devices, and processor-readable storage media are provided for the diagnosis of heart failure. A method in this context includes collecting, using an implantable device, reference episodes, the reference episodes comprising, at least one of: electrical activity signals of a myocardium; myocardium hemodynamic activity signals, or indicators reflecting variation of physical parameters, variation of activity, and variation of hemodynamic phases between phases of effort and phases of recovery; generating an in-suspicion model-cycle and an off-suspicion model-cycle based on the reference episodes; and determining whether to generate an early heart failure alert, based on a difference between the in-suspicion model-cycle and the off-suspicion model-cycle.

Abstract: A cardiac implant includes a classifier configured analyze the data for comparison to thresholds prioritized according to a Boolean decision tree. The implant can generate an indicator of alert status (i.e., alert or no alert). A supervisory device operated, for example, by a doctor can associate each indicator a marker (AE/no AE) indicating the presence or absence of an observed adverse event. In the presence of a false positive, a command for update of the thresholds of the decision tree is transmitted to the implant. A database of reference patients can be used to recreate or further update the decision tree to avoid the occurrence of false negatives.