Prospective evaluation of the multisensor HeartLogic algorithm for heart failure monitoring

Abstract

Background: The HeartLogic algorithm measures data from multiple implantable cardioverter-defibrillator-based sensors and combines them into a single index. The associated alert has proved to be a sensitive and timely predictor of impending heart failure (HF) decompensation.

Hypothesis: We describe a multicenter experience of remote HF management by means of HeartLogic and appraise the value of an alert-based follow-up strategy.

Methods: The alert was activated in 104 patients. All patients were followed up according to a standardized protocol that included remote data reviews and patient phone contacts every month and at the time of alerts. In-office examinations were performed every 6 months or when deemed necessary.

Results: During a median follow-up of 13 (10-16) months, the overall number of HF hospitalizations was 16 (rate 0.15 hospitalizations/patient-year) and 100 alerts were reported in 53 patients. Sixty alerts were judged clinically meaningful, and were associated with multiple HF-related conditions. In 48 of the 60 alerts, the clinician was not previously aware of the condition. Of these 48 alerts, 43 triggered clinical actions. The rate of alerts judged nonclinically meaningful was 0.37/patient-year, and the rate of hospitalizations not associated with an alert was 0.05/patient-year. Centers performed remote follow-up assessments of 1113 scheduled monthly transmissions (10.3/patient-year) and 100 alerts (0.93/patient-year). Monthly remote data review allowed to detect 11 (1%) HF events requiring clinical actions (vs 43% actionable alerts, P < .001).

Conclusions: HeartLogic allowed relevant HF-related clinical conditions to be identified remotely and enabled effective clinical actions to be taken; the rates of unexplained alerts and undetected HF events were low. An alert-based management strategy seemed more efficient than a scheduled monthly remote follow-up scheme.

Keywords: CRT; ICD; decompensation; heart failure; telemedicine.

FIGURE 1

Report of automatic diagnostics available for review through the LATITUDE remote monitoring platform. It includes the HeartLogic index and the contributing sensors: accelerometer‐based first and third heart sounds, intrathoracic impedance, respiration rate, night heart rate, and patient activity. In this example, a 67‐years‐old man with dilated cardiomyopathy, left bundle branch block, left ventricular ejection fraction 31%, NYHA Class II, underwent implantation of Resonate X4 CRT‐D. After implantation, the patient was fine and the index remained under the threshold until January 8, 2019, when a HeartLogic index was notified to the center (blue bar). The nurse responsible for remote monitoring contacted the patient who did not report worsening HF symptoms. After 1 week, the alert state was persisting. The nurse contacted again the patient who continued to report no symptoms, but he referred that he had discontinued diuretic therapy at the end of December (red bar). After consulting with the doctor, the nurse suggested the patient to restore diuretic therapy (green bar). The HeartLogic index decreased to below the recovery threshold value of 6 on January 29th

FIGURE 2

Upper panel. Flowchart depicting the adjudication of HeartLogic alerts. Clinically meaningful alerts: associated with HF events or alerts that resulted in active clinical actions; New information: no reported previously sign or symptom, no action already taken to treat the clinical condition; Actionable alerts: alerts that resulted in active clinical actions to manage the HF condition. Middle panel. HF‐related conditions associated with clinically meaningful alerts (multiple conditions were reported per alert). Lower panel. Actions taken to manage the HF condition detected by the alert (multiple actions were reported per alert)

FIGURE 3

HF signs detected during in‐office visits according to the HeartLogic alert state

FIGURE 4

Symptoms of HF reported during in‐office or remote examinations according to the HeartLogic alert state