Advertisement

Multiple cArdiac seNsors for mAnaGEment of Heart Failure (MANAGE-HF) – Phase I Evaluation of the Integration and Safety of the HeartLogic Multisensor Algorithm in Patients With Heart Failure

Open AccessPublished:April 20, 2022DOI:https://doi.org/10.1016/j.cardfail.2022.03.349

      Abstract

      Background

      Patients with heart failure (HF) and reduced ejection fraction suffer from a relapsing and remitting disease course, where early treatment changes may improve outcomes. We assessed the clinical integration and safety of the HeartLogic multisensor index and alerts in HF care.

      Methods

      The Multiple cArdiac seNsors for mAnaGEment of Heart Failure (MANAGE-HF) study enrolled 200 patients with HF and reduced ejection fraction (<35%), New York Heart Association functional class II–III symptoms, implanted with a cardiac resynchronization therapy-defibrillator or and implantable cardioverter defibrillator, who had either a hospitalization for HF within 12 months or unscheduled visit for HF exacerbation within 90 days or an elevated natriuretic peptide concentration (brain natriuretic peptide [BNP] of ≥150 pg/mL or N-terminal pro-BNP [NT-proBNP] of ≥600 pg/mL). This phase included the development of an alert management guide and evaluated changes in medical treatment, natriuretic peptide levels, and safety.

      Results

      The mean age of participants was 67 years, 68% were men, 81% were White, and 61% had a HF hospitalization in prior 12 months. During follow-up, there were 585 alert cases with an average of 1.76 alert cases per patient-year. HF medications were augmented during 74% of the alert cases. HF treatment augmentation within 2 weeks from an initial alert was associated with more rapid recovery of the HeartLogic Index. Five serious adverse events (0.015 per patient-year) occurred in relation to alert-prompted medication change. NTproBNP levels decreased from median of 1316 pg/mL at baseline to 743 pg/mL at 12 months (P < .001).

      Conclusions

      HeartLogic alert management was safely implemented in HF care and may optimize HF management. This phase supports further evaluation in larger studies.

      Trial Registration

      ClinicalTrials.gov (NCT03237858)

      Graphical abstract

      Key Words

      Heart failure (HF) is a major public health burden associated with high mortality and recurrent hospitalizations, with an economic burden in the United States alone estimated to be $70 billion by 2030.
      • Heidenreich P.A.
      • Albert NM
      • Allen LA
      • Bluemke DA
      • Butler J
      • Fonarow GC
      • et al.
      Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association.
      Approximately two-thirds of these costs are associated with worsening HF resulting in hospitalization. Although self-management by the monitoring of symptoms and weight is a mainstay of disease management, improving clinical outcomes is often difficult.
      • Takeda A.
      • Martin N
      • Taylor RS
      • Taylor SJ
      Disease management interventions for heart failure.
      Clinical decompensation can occur before the recognition of weight gain for a number of factors and may be overtly missed in the setting of cardiac cachexia in which weights are a poor indicator. Subclinical physiological changes often precede worsening HF and may be able to be used to predict and prevent hospitalizations.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      Patients may also have chronic symptoms from HF, making the detection of changes over time of worsening HF a difficult task.
      Remote monitoring and digital technologies are increasingly able to provide real-time, continuous data on patients.
      • Keesara S.
      • Jonas A.
      • Schulman K.
      Covid-19 and health care's digital revolution.
      Despite these advances, the majority of prior clinical trials have shown limited ability of such data to improve clinical outcomes.
      • Bourge R.C.
      • Abraham WT
      • Adamson PB
      • Aaron MF
      • Aranda Jr, JM
      • Magalski A
      • et al.
      Randomized controlled trial of an implantable continuous hemodynamic monitor in patients with advanced heart failure: the COMPASS-HF study.
      • van Veldhuisen D.J.
      • Braunschweig F
      • Conraads V
      • Ford I
      • Cowie MR
      • Jondeau G
      • et al.
      Intrathoracic impedance monitoring, audible patient alerts, and outcome in patients with heart failure.
      • Morgan J.M.
      • Kitt S
      • Gill J
      • McComb JM
      • Ng GA
      • Raftery J
      • et al.
      Remote management of heart failure using implantable electronic devices.
      More recently, the use of a wireless pulmonary artery hemodynamic monitoring failed to confirm improved outcomes in a broad range of patients with HF.
      • Lindenfeld J.
      • Zile MR
      • Desai AS
      • Bhatt K
      • Ducharme A
      • Horstmanshof D
      • Krim SR
      • et al.
      Haemodynamic-guided management of heart failure (GUIDE-HF): a randomised controlled trial.
      In the setting of HF management with the assistance of a remote HF diagnostic tool, 3 critical components are likely necessary: (1) diagnostic performance to detect HF worsening, (2) the ability to properly integrate into clinical workflow, and (3) an actionable treatment change in addition to standard clinical care.
      Cardiac implantable electronic devices with sensors have been used to develop HF diagnostics to detect early signals leading up to hospitalizations, with most using single physiological measures.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      ,
      • Abraham W.T.
      • Stevenson LW
      • Bourge RC
      • Lindenfeld JA
      • Bauman JG
      • Adamson PB
      CHAMPION Trial Study Group
      Sustained efficacy of pulmonary artery pressure to guide adjustment of chronic heart failure therapy: complete follow-up results from the CHAMPION randomised trial.
      • Conraads V.M.
      • Tavazzi L
      • Santini M
      • Oliva F
      • Gerritse B
      • Yu CM
      • et al.
      Sensitivity and positive predictive value of implantable intrathoracic impedance monitoring as a predictor of heart failure hospitalizations: the SENSE-HF trial.
      • Heist E.K.
      • Herre JM
      • Binkley PF
      • Van Bakel AB
      • Porterfield JG
      • Porterfield LM
      • et al.
      Analysis of different device-based intrathoracic impedance vectors for detection of heart failure events (from the Detect Fluid Early from Intrathoracic Impedance Monitoring study).
      • Auricchio A.
      • Gold MR
      • Brugada J
      • Nölker G
      • Arunasalam S
      • Leclercq C
      • et al.
      Long-term effectiveness of the combined minute ventilation and patient activity sensors as predictor of heart failure events in patients treated with cardiac resynchronization therapy: Results of the Clinical Evaluation of the Physiological Diagnosis Function in the PARADYM CRT device Trial (CLEPSYDRA) study.
      A more robust approach that combines information from a diverse set of sensors integrated into an algorithm and alert for detecting worsening HF may be an advantage with better operating characteristics.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      To address this need, the HeartLogic algorithm was developed that incorporated sensor data from accelerometer-based first and third heart sounds, thoracic impedance, respiration rate, relative tidal volume, heart rate, and patient activity. Previously, the Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients study showed the HeartLogic multisensor index and alert algorithm can provide a sensitive predictor of worsening HF.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      To further evaluate whether the HeartLogic algorithm can be integrated into HF care and improve outcomes, the Multiple Cardiac Sensors for Management of Heart Failure (MANAGE-HF) study was designed to evaluate HeartLogic integration, use, safety, and efficacy in HF in 2 phases. Phase I objectives were to develop an alert management guide (AMG), treatment process and evaluate associated outcomes including safety. Phase II will evaluate efficacy of HeartLogic alerts combined with an alert management protocol on patient outcomes. This report presents the main findings of phase I.

      Methods

      MANAGE-HF included patients age 18 years or older with HF, New York Heart Association functional class II–III symptoms, and implanted with a cardiac resynchronization therapy-defibrillator or implantable cardioverter defibrillator with HeartLogic, who had either a HF hospitalization within 12 months or unscheduled visit with intravenous diuretic therapy for HF exacerbation within 90 days or an elevated natriuretic peptide concentration (brain natriuretic peptide [BNP] of ≥150 pg/mL or N-terminal pro-BNP [NT-proBNP] of ≥600 pg/mL) at any time during 90 days before enrollment. Patients also had to be remotely monitored via the LATITUDE (Boston Scientific, Marlborough, MA) system.
      Patients were excluded if they had New York Heart Association functional class IV symptoms, were implanted with unipolar right atrial or right ventricular leads, received or were scheduled to receive a heart transplant or ventricular assist device in the next 6 months, had a glomerular filtration rate of less than 25 mL/min while nonresponsive to diuretic therapy or on chronic renal dialysis, had been regularly scheduled for intravenous HF therapy (eg, inotropes or diuretics), had a life expectancy of less than 12 months, or were pregnant or planned a pregnancy.
      The Steering Committee designed the trial in collaboration with the study sponsor, Boston Scientific. An independent Data and Safety Monitoring Committee evaluated patient safety. A Clinical Event Committee provided an independent review of all serious adverse events (SAE) that resulted in a hospitalization or an outcome of death. The sponsor participated in the selection of participating centers, site monitoring, and data storage. Analyses were conducted by the sponsor, and the steering committee participated in the interpretation of the data. All authors had unrestricted access to the data and the first author drafted the initial version of the manuscript, which was reviewed and edited by all the authors. All the authors vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol.

      Study Design

      Patients who met the eligibility criteria were enrolled into a single arm, open-label study with active management of HeartLogic alerts with the assistance of an AMG developed by the steering committee.
      After enrollment, the baseline visit occurred after a minimum of 45 days of HF sensor data were collected and before 4 months from enrollment. At baseline, HeartLogic alerts were enabled and alert threshold was programmed to 16 (nominal alert threshold setting). After baseline, participants were monitored per routine clinical care and every 6 months per protocol to assess clinical status, events, and medication changes. After the baseline visit, adjustment of the alert threshold was allowed. All patients were followed until the last participant completed the 12-month follow-up visit.
      Providers received automated notice when an initial HeartLogic alert occurred and received weekly reminders (re-alerts) until the HeartLogic index recovered below the nominal alert recovery threshold (6). After all alerts, providers were required to attempt to contact the patient (Supplemental Appendix, e Fig. 1 ) and were encouraged to follow an AMG.
      Fig 1
      Fig. 1Illustration of HeartLogic alert definitions. On March 23, HeartLogic index (solid line) crosses a nominal threshold of 16 (dashed line), transitions from out-of-alert to in-alert state and issues an initial alert. It then issues re-alerts weekly on March 30, April 6, and April 16 until it crosses a nominal threshold of 6 on April 17th and transitions to out-of-alert state. The alert case lasts from March 23 (alert onset) to April 17 (alert recovery) and consists of 4 weekly alerts.
      The AMG was developed by the steering committee and integrated guideline directed medical therapy as well as best approaches for HF disease management. The AMG was submitted along with the protocol for approval by the US Food and Drug Administration (FDA).
      • Albert N.M.
      • Barnason S
      • Deswal A
      • Hernandez A
      • Kociol R
      • Lee E
      • et al.
      Transitions of care in heart failure: a scientific statement from the American Heart Association.
      ,
      • Yancy C.W.
      • Jessup M
      • Bozkurt B
      • Butler J
      • Casey Jr, DE
      • Colvin MM
      • et al.
      2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America.
      The AMG provided a treatment algorithm for clinical assessment, modification of diuretics and guideline directed medical therapy using an iterative process to integrate best practices observed in phase I (Supplemental Appendix, Final Version of the AMG). Treatments in response to a HeartLogic alert or in response to a worsening of HF were collected, including reinforcing adherence, change in current HF medications, addition of a new class of HF medications, or change within a HF class of medications. The first version (AMG Rev A, January 2018) contained a detailed treatment response flowchart and specific recommendation on treating fluid overload with mostly short duration or bursts of additional diuretics. The following year, the steering committee revised the AMG (Rev B, January 2019) to modify the diuretic response algorithm to encourage sustained uptitration rather than burst diuretics, link the changes to the behavior of HeartLogic index, and to change presentation from a table to an algorithmic flowchart. In March, 2020 the final version AMG (Rev C) was FDA approved with the major changes shifting the emphasis from diuretics in response to HeartLogic alerts to a larger set of HF treatments, and linking the behavior of HeartLogic index as an indication of whether the chosen treatment should be modified or maintained. More emphasis was also placed on addressing precipitating factors and optimizing GDMT than in the prior version of the AMG.
      The HeartLogic index has a range of 0–100 and a nominal value of 16 has been established to maximize operating characteristics and minimizing false positives for alerts. The evidence and data for using a nominal threshold of 16 for an alert primarily derives from the MultiSENSE study and was used by the FDA to support HeartLogic as diagnostic alert in HF.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      The HeartLogic alert state was determined per algorithm specification using the HeartLogic index data following the study baseline visit: the in-alert state begins when HeartLogic index exceeds the nominal alert threshold (16) and continues until the index falls below the nominal alert recovery threshold (6), an out-of-alert state occurs at all other times. A HeartLogic alert case is defined as a contiguous time period of in-alert state (Fig. 1). A HeartLogic alert case consists of weekly alerts: the initial alert and following re-alerts.

      Study End Points

      The primary evaluation of the phase I study was safe integration into clinical practice. Given this study was designed to evaluate HeartLogic integration into clinical practice, all analyses are considered exploratory. These analyses included an evaluation of HeartLogic performance, change in medical treatment, plasma natriuretic peptide concentrations, and HF hospitalization rate. An augmented HF treatment is defined as a higher equivalent dose of diuretics, vasodilators, angiotensin receptor blocker neprilysin inhibitors (ARNI), angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), beta-blockers, or mineralocorticoid receptor antagonists (MRA) in the 6 days after a weekly alert, as compared with the dose on the day before the initial alert. New additions of medications or increases in dosing of a HF medication were categorized as either bursts or sustained treatments, based on their duration: 1–3 days for bursts and more than 3 days for sustained treatments. A switch to a more bioavailable diuretic was also considered an augmented HF treatment, even if the equivalent dose stayed the same. Adverse events were collected through the study period and classified, by the sites, as related or not to the device or the study intervention.

      Statistical Analysis

      To evaluate differences in natriuretic peptide levels at the 12 month follow-up visits compared with baseline visits, a Wilcoxon signed-rank test for matched pairs was used. The rate of hospitalizations and death was determined as the sum of events divided by the sum of patient follow-up years for the HeartLogic follow up period (from baseline visit to end of study). A HF hospitalization was defined as a Clinical Event Committee–adjudicated hospitalization with a primary or secondary cause of HF.

      Results

      Baseline Characteristics

      Between August 2017 and May 2019, 29 sites enrolled 200 patients with 191 completing baseline visit[s] and having HeartLogic activated (Fig. 2). Study follow-up was completed in July 2020 with an average patient follow up of 20.9 months from baseline. The mean patient age was 67 years, 68% were men, 81% were Caucasian, and 61% had a HF hospitalization in prior 12 months. Ischemic heart disease was present in 47% and the mean left ventricular ejection fraction was 26%. Comorbidities were common with 15% of patients with chronic obstructive pulmonary disease, 26% with renal dysfunction, and 36% with diabetes (Table 1). Baseline medications included a beta-blocker in 96%, ACEI or ARB in 54%, ARNI in 27%, MRA in 43%, and loop diuretic in 83% of patients.
      Table 1. Baseline Characteristics
      MANAGE-HFN = 191
      Demographics
       Age, mean ± SD67 ± 12
       Male gender, n (%)129 (68)
       Race (White, not of Hispanic origin), n (%)142 (81)
      Medical history, n (%)
       Ischemic heart disease90 (47)
       Dilated cardiomyopathy75 (39)
       Idiopathic cardiomyopathy20 (11)
       Valvular disease48 (25)
       Myocardial infarction73 (38)
       Coronary artery bypass grafting49 (26)
       Chronic obstructive lung disease28 (15)
       Pulmonary hypertension14 (7)
       Peripheral vascular disease25 (14)
       Cerebrovascular disease32 (17)
       Renal dysfunction50 (26)
       Hypertension144 (76)
       Diabetes69 (36)
       Hyperlipidemia134 (70)
       Sleep apnea45 (25)
       Depression34 (18)
       Hepatic disease11 (6)
       Anemia27 (14)
      Device type, n (%)
       Cardiac resynchronization therapy-defibrillator132 (69)
       Implantable cardioverter defibrillator59 (31)
      Laboratory measurements
       LVEF (%), mean ± SD25.6 ± 9.5
       NT-proBNP (pg/mL), median (Q1, Q3)1514 (685, 3480)
       Hemoglobin (g/dL), mean ± SD12.6 ± 2.1
       Creatinine (mg/dL), mean ± SD1.3 ± 0.6
      Concomitant medications, n (%)
       Loop diuretic158 (83)
       Thiazide diuretic17 (9)
       ACEI or ARB103 (54)
       ARNI51 (27)
       MRA82 (43)
       Beta-blocker184 (96)
       Vasodilators35 (18)
      ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blockers; ARNI, angiotensin receptor blocker neprilysin inhibitors; MRA, mineralocorticoid receptor antagonists; NT-proBNP, N-terminal pro-brain natriuretic peptide; SD, standard deviation.

      Remote Monitoring Alerts and Responses

      During the HeartLogic follow-up period between baseline and end of study (333 patient-years), there were a total of 585 alert cases computed at nominal threshold with a total of 2705 weekly re-alert reminders for a total of 3290 weekly alerts. The HeartLogic alert rate at nominal threshold was 1.76 alert cases per patient-year. The average alert duration was 36 days (median, 27 days) and 17% of the total follow-up time was associated with an in-alert state. Some alert communications were delayed or not generated because there were times when the patient's home communicator was not powered on or could not send data, or the patient was out of range, or the alert threshold was adjusted from nominal. Of the total 3290 weekly alerts, 2934 (89%) were communicated to the sites (median delivery time <1 day, Q3 <1 day, max 129 days), 2894 (88%) were documented as received by sites, 2677 (81%) were followed by an attempt to contact the patient, and 2402 (73%) were followed by a successful patient contact. Overall, there was successful patient contact during 93% of the alert cases (542/585).
      HF treatment was augmented during 74% of the 585 alert cases and during 54% of the 3290 weekly alerts. The following HF drug classes were augmented: 1590 (89%) diuretics, 185 (10%) beta-blockers, 132 (7%) MRA, 124 (7%) ARNI, 108 (6%) ACE/ARBs, and 69 (4%) vasodilators (hydralazine or nitrate). For subjects who completed the 12-month visit, the median NT-proBNP decreased from 1316 pg/mL (Q1–Q3, 664–2856 pg/mL) at baseline visit to 743 pg/mL (Q1–Q3, 336–1681 pg/mL) at the 12-month visit (P < .001).
      There was variability in alert response rates and treatments across sites, which ranged from augmenting diuretics only (loop or thiazide), diuretic + nondiuretic HF medication (beta-blockers, ACEI/ARB/ARNI, MRA, vasodilators), nondiuretic HF medication only, to not augmenting any HF treatment (Fig. 3A). The choice of response also varied by duration of the alert case (Fig. 3B). Diuretics were more frequently used initially, with movement toward more sustained diuretic changes and the addition of nondiuretic HF medications as the alert cases progressed.
      Fig 3
      Fig. 3Variability in response to alerts. (A) The use of heart failure (HF) treatments in response to weekly alerts by site, ordered by how often weekly alerts were treated with HF medications. The number of weekly alerts at each site shown above the bars. (B) Change in HF treatment choices throughout alert case duration. Burst = 1–3 days of increased diuretics. Sustained = >3 days of increased diuretic.
      Fig. 4 shows that the HF medication changes during first 12 months of the study were much more frequent while in-alert state than when out-of-alert state: 8.3-fold for all HF classes, 11.6-fold for loop diuretics, 24.7-fold for thiazide diuretics, 3-fold for ARNI/ACEI/ARBs, 2.6-fold for beta blockers, and 5.3-fold for vasodilators. While in alert, increases were more prevalent than decreases in all classes except for beta blockers (no difference): 1.44-fold for loop diuretics, 1.13-fold for thiazide diuretics, 1.25-fold for ARNI/ACEI/ARBs, and 3.75-fold for vasodilators. Overall, most changes were changes in loop diuretics.
      Fig 4
      Fig. 4Frequency of HF medication changes over first 12 months of the study, grouped by medication class, type of change, and alert state. ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blockers; ARNI, angiotensin receptor blocker neprilysin inhibitors; MRA, mineralocorticoid receptor antagonists.
      The HeartLogic index decreased quicker and alert cases resolved quicker when decongestive therapies were given (Fig. 5).
      Fig 5
      Fig. 5HeartLogic index temporal response to decongestive treatments. Average HeartLogic index value (± SEM) aligned on the day of alert case onset. The blue line and bars represent patients with an augmented loop diuretic, thiazide, or ARNI during first 2 weeks of the alert case. The red line and bars represent patients with no augmentation in loop diuretic, thiazide, or ARNI during first 4 weeks of the alert case. HeartLogic index data after a hospitalization or a corrective action other than oral medication change were removed. Horizontal grid lines represent the nominal alert onset threshold of 16 and recovery threshold of 6. Alert cases with an augmentation in loop diuretic, thiazide, or ARNI during the 14 days before alert onset and alert cases interrupted by death or withdrawal from the study were excluded from the analysis.

      HeartLogic Assessment

      At the initial alert, 64% of patients presented with at least 1 symptom of worsening of HF. There were 70 hospitalizations with HF as the primary reason for admission, 15 hospitalizations with HF as the secondary reason for admission, and 23 IV outpatient treatments during the HeartLogic follow-up period. The hospitalization rate for HF was 0.26 per patient-year. HeartLogic was in alert status within 30 days preceding 83% of hospitalizations with HF as the primary reason for admission and 77% of IV outpatient treatments. The rate of all-cause mortality was 0.05 per patient-year and the rate of all-cause hospitalization was 0.73 per patient-year for the HeartLogic follow-up period.

      Adverse Events

      A total of 691 AEs, 301 of which were SAEs, were reported. Fifty (7%) were deemed yes or possibly related to a qualified treatment resulting from a HeartLogic alert and 641 (93%) were deemed not related by the investigational sites. Five of the 50 AE, deemed yes or possibly related, were SAE. The 5 SAEs occurred in 4 of 147 patients with alerts (0.015/patient-year). The 5 SAE occurred in relation to 4 of the 585 alert cases (0.7%) and 1922 changes in HF medications (ie, increases, decreases, additions, and/or discontinuations) that occurred during alert cases (0.3%). The 5 SAEs were classified by the Boston Scientific internal safety team as abnormal laboratory values, renal insufficiency/failure–HF (2), dizziness–HF, and syncope–HF.

      Discussion

      In this prospective observational study of patients with HF and reduced ejection fraction, the use of a HeartLogic multisensor algorithm with an AMG was integrated safely into clinical practice and associated with lower natriuretic peptide levels during the study period. Augmenting decongestive treatment early during an alert was associated with a shorter alert duration. Although there was significant variability in site responses to alerts, the majority of responses targeted decongestion.
      Remote management of HF will be needed owing to the growth in the population and the need for better, proactive population health management strategies to avoid progressive decompensation and hospitalization. To date, most strategies in remote monitoring and management have been limited to simple measures of HF, such as weight gain or worsening dyspnea or fatigue, with few incorporating physiological sensor monitoring. Physiological sensors can provide greater prediction of worsening HF with an opportunity to intervene before overt worsening of clinical symptoms develop, as well as provide objective data to guide intervention. Prior studies have generally relied on single sensors, which does not allow the comprehensive monitoring of the diverse physiological changes that occur in HF.
      • Conraads V.M.
      • Tavazzi L
      • Santini M
      • Oliva F
      • Gerritse B
      • Yu CM
      • et al.
      Sensitivity and positive predictive value of implantable intrathoracic impedance monitoring as a predictor of heart failure hospitalizations: the SENSE-HF trial.
      • Heist E.K.
      • Herre JM
      • Binkley PF
      • Van Bakel AB
      • Porterfield JG
      • Porterfield LM
      • et al.
      Analysis of different device-based intrathoracic impedance vectors for detection of heart failure events (from the Detect Fluid Early from Intrathoracic Impedance Monitoring study).
      • Auricchio A.
      • Gold MR
      • Brugada J
      • Nölker G
      • Arunasalam S
      • Leclercq C
      • et al.
      Long-term effectiveness of the combined minute ventilation and patient activity sensors as predictor of heart failure events in patients treated with cardiac resynchronization therapy: Results of the Clinical Evaluation of the Physiological Diagnosis Function in the PARADYM CRT device Trial (CLEPSYDRA) study.
      The HeartLogic diagnostic provides an alert algorithm integrating multiple sensors monitoring multiple facets of HF to allow an early prediction of worsening HF.
      • Boehmer J.P.
      • Hariharan R
      • Devecchi FG
      • Smith AL
      • Molon G
      • Capucci A
      • et al.
      A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
      In the current study, HeartLogic was integrated into clinical practice with alerts transmitted to clinical teams and responses to alerts performed locally. Importantly, some initial alerts were presymptomatic, allowing treatment to begin before overt decompensation and before hospitalization was required. In response to alerts, the augmentation of diuretics was the principal treatment pattern. The early use of decongestive treatment in response to alerts was associated with a shorter alert duration. Conversely, the response to alerts was not typically associated with an increase in guideline-directed medical therapies, such as ACEI/ARB/ARNI. These results indicate a gap in optimization of HF management that may be addressed more fully in future efforts to enhance guideline-directed medical therapy with physiologic monitoring. In contrast, the results potentially highlight the importance of monitoring and/or managing volume control more actively in patients with HF.
      It is important to understand why HeartLogic may have had an impact on management of patients with HF patients compared with prior studies. In other studies, remote monitoring was challenging because alerts were not successfully transmitted to clinical teams.
      • Crossley G.H.
      • Boyle A
      • Vitense H
      • Chang Y
      • Mead RH
      CONNECT Investigators
      The CONNECT (Clinical Evaluation of Remote Notification to Reduce Time to Clinical Decision) trial: the value of wireless remote monitoring with automatic clinician alerts.
      ,
      • Bohm M.
      • Drexler H
      • Oswald H
      • Rybak K
      • Bosch R
      • Butter C
      • et al.
      Fluid status telemedicine alerts for heart failure: a randomized controlled trial.
      In addition, there was a general lack of standardized guidance for how to act on alerts, unlike the current study, which provided an AMG to health care providers and clinicians that encouraged taking action for presymptomatic alerts. MANAGE-HF did not only test HeartLogic alerts, but also critically included an AMG that provides stepwise actions for diuretics escalation, improvements in guideline-directed medical therapy, and addressing underlying causes.
      As new models of care go forward for HF, complex diagnostics and algorithms will proliferate. Although many of these tools will offer important predictive characteristics, it will be critical to evaluate whether such algorithms can be incorporated effectively into HF management. In addition, it will also be important to establish the optimal care model to support the human resources, either locally or centrally, to manage alerts when they occur. In an ideal situation, a larger volume of patients can be managed remotely, alleviating the use of on-site clinical activities for those who are the sickest or need more direct management. Implementation safety becomes even more critical with the increased interest in more remote or virtual care that the coronavirus disease 2019 pandemic has caused throughout care delivery systems. The ability to provide better monitoring and management may also provide more efficient pathways for effectively triaging clinical visits and automating guideline-directed medical treatment.
      Study limitations include its nonrandomized observational design and lack of a uniform adherence to the AMG throughout the study. The study protocol was intentionally flexible to allow learning and the development of guidance to safely integrate responses to alerts into routine clinical practice. However, iterative adjustments of the guide, a delay in its initial release, and absence of protocol-required actions led to its variable implementation by sites. Finally, given the observational nature, residual confounding may be present and influence the findings. This study’s sample size is modest and an assessment of the efficacy of clinical outcomes should be interpreted with caution because this phase was the first of 2 phases of MANAGE-HF.

      Conclusions

      The HeartLogic multisensor index and alert algorithm management was safely implemented in HF care. The implementation was associated with HF treatment augmentation, in response to alerts, resulting in more rapid reduction of HeartLogic index values. Based on the results of this study, initiating phase II of MANAGE-HF as a large, randomized clinical trial is fully supported.

      Brief Lay Summary

      MANAGE-HF was a multicenter, observational study that enrolled 200 patients with HeartLogic enabled devices. The HeartLogic Index uses multiple sensors to track patient health indicators and predict worsening heart failure. In this study, alerts were sent to health care providers when the HeartLogic Index increased over an alert threshold. The health care providers evaluated the patients with alerts and documented treatment response. A timely response to the alerts coupled with changes in heart failure medications was safe and associated with more rapid improvement in the patients HeartLogic Index, perhaps signaling improved heart failure management.

      Funding Source

      Funding for this research was provided by Boston Scientific.

      Disclosures

      Adrian Hernandez reports research grants from American Regent, AstraZeneca, Boehringer Ingelheim, Bristol Myers Squibb, Janssen, Merck, Novartis, Pfizer, and Verily; consulting fees from AstraZeneca, Amgen, Bayer, Boston Scientific, Cytokinetics, Merck, Myokardia, Novartis, and Relypsa. Larry Allen reports research Grants from the American Heart Association, National Institutes of Health, and Patient Centered Outcomes Research Institute; and consulting fees from ACI Clinical, Amgen, Boston Scientific, Cytokinetics, and Novartis.
      Christopher Chien reports consulting fees from Abbott Laboratories, Astra Zeneca, and Boston Scientific. Martin Cowie reports consulting fees from Abbott, Boston Scientific, Medtronic, AstraZeenca, Novarits, Pfizer, Bayer, Boehringer Ingelheim, Roche Diagnostics, and Servier. Nancy Albert reports consulting fees from Amgen, AstraZeneca, Boston Scientific, Merck, and Novartis. Liviu Klein reports research grants from Ancora Heart, CVRx, Optima Integrated Health, Respicardia, and V-Wave Medical; and consulting fees from Abbott, Boston Scientific, and Medtronic.
      Carolyn Lam reports research support from Boston Scientific, Bayer, Roche Diagnostics, AstraZeneca, Medtronic, and Vifor Pharma; consultancy fees from Abbott Diagnostics, Amgen, Applied Therapeutics, AstraZeneca, Bayer, Biofourmis, Boehringer Ingelheim, Boston Scientific, Corvia Medical, Cytokinetics, Darma Inc., Us2.ai, JanaCare, Janssen Research & Development LLC, Medtronic, Menarini Group, Merck, MyoKardia, Novartis, Novo Nordisk, Radcliffe Group Ltd., Roche Diagnostics, Sanofi, Stealth BioTherapeutics, The Corpus, Vifor Pharma, and WebMD Global LLC; and is the co-founder and a nonexecutive director of Us2.ai.
      Rezwan Ahmed, Viktoria Averina, Brian Kwan, Stephen Ruble, Kenneth Stein, and Craig Stolen are employees and shareholders of Boston Scientific. Marie Galvao reports consulting fees from Boston Scientific.

      Enrolling Sites

      Heart Center Research, LLC, Huntsville, AL; Rex Hospital, Raleigh, NC; Cardiology Associates of Northeast Arkansas, P.A., Jonesboro, AR; Catholic Medical Center, Manchester, NH; Hopital Prive du Confluent SAS, Nantes, France; Community Heart and Vascular Hospital, Indianapolis, IN; Sacred Heart Medical Center at Riverbend, Springfield, OR; Advanced Cardiovascular Specialists, Shreveport, LA; Sentara Norfolk General Hospital, Norfolk, VA; Emory University Hospital, Atlanta, GA; Penn State Milton S Hershey Medical Center, Hershey, PA; Centracare Heart and Vascular Center, St. Cloud, MN; Sharp Memorial Hospital, San Diego, CA; Universitaetsklinikum Wuerzburg, Wuerzburg, Germany; Parkview Hospital, Inc., Fort Wayne, IN; SouthEast Texas Clinical Research Center, Beaumont, TX; Stern Cardiovascular Foundation, Inc., Memphis, TN; United Heart and Vascular Clinic, St. Paul, MN; University of Southern California Hospital, Los Angeles, CA; Bethesda North Hospital, Cincinnati, OH; Cardiology Consultants of Philadelphia, Yardley, PA; CHRU Hopital Pontchaillou, Rennes, France; Duke University Medical Center, Durham, NC; Memorial Hospital, Colorado Springs, CO; Montefiore Medical Center, Bronx, NY; University of California, San Francisco, San Francisco, CA; Cardiovascular Consultants, Concord, CA; Lindner Center for Research and Education at Christ Hosp, Cincinnati, OH; University of Colorado Hospital, Aurora, CO

      Acknowledgments

      DSMB Members: DSMB Chair, Inder Anand, University of Minnesota, Minneapolis, MN; Akshay Desai, Brigham and Women's Hospital, and Harvard Medical School, Boston, Massachusetts; David DeMets, University of Wisconsin-Madison, Madison
      Clinical Event Committee Members: Clinical Event Committee Chair, Brian Jaski, Sharp Memorial Hospital, San Diego, CA; Teresa De Marco, University of California, San Francisco, CA; J. Thomas Heywood, Scripps Clinic, San Diego, CA.

      Appendix. Supplementary materials

      References

        • Heidenreich P.A.
        • Albert NM
        • Allen LA
        • Bluemke DA
        • Butler J
        • Fonarow GC
        • et al.
        Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association.
        Circ Heart Fail. 2013; 6: 606-619
        • Takeda A.
        • Martin N
        • Taylor RS
        • Taylor SJ
        Disease management interventions for heart failure.
        Cochrane Database Syst Rev. 2019; 1CD002752
        • Boehmer J.P.
        • Hariharan R
        • Devecchi FG
        • Smith AL
        • Molon G
        • Capucci A
        • et al.
        A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE study.
        JACC Heart Fail. 2017; 5: 216-225
        • Keesara S.
        • Jonas A.
        • Schulman K.
        Covid-19 and health care's digital revolution.
        N Engl J Med. 2020; 382: e82
        • Bourge R.C.
        • Abraham WT
        • Adamson PB
        • Aaron MF
        • Aranda Jr, JM
        • Magalski A
        • et al.
        Randomized controlled trial of an implantable continuous hemodynamic monitor in patients with advanced heart failure: the COMPASS-HF study.
        J Am Coll Cardiol. 2008; 51: 1073-1079
        • van Veldhuisen D.J.
        • Braunschweig F
        • Conraads V
        • Ford I
        • Cowie MR
        • Jondeau G
        • et al.
        Intrathoracic impedance monitoring, audible patient alerts, and outcome in patients with heart failure.
        Circulation. 2011; 124: 1719-1726
        • Morgan J.M.
        • Kitt S
        • Gill J
        • McComb JM
        • Ng GA
        • Raftery J
        • et al.
        Remote management of heart failure using implantable electronic devices.
        Eur Heart J. 2017; 38: 2352-2360
        • Lindenfeld J.
        • Zile MR
        • Desai AS
        • Bhatt K
        • Ducharme A
        • Horstmanshof D
        • Krim SR
        • et al.
        Haemodynamic-guided management of heart failure (GUIDE-HF): a randomised controlled trial.
        Lancet. 2021; 398: 991-1001
        • Abraham W.T.
        • Stevenson LW
        • Bourge RC
        • Lindenfeld JA
        • Bauman JG
        • Adamson PB
        • CHAMPION Trial Study Group
        Sustained efficacy of pulmonary artery pressure to guide adjustment of chronic heart failure therapy: complete follow-up results from the CHAMPION randomised trial.
        Lancet. 2016; 387: 453-461
        • Conraads V.M.
        • Tavazzi L
        • Santini M
        • Oliva F
        • Gerritse B
        • Yu CM
        • et al.
        Sensitivity and positive predictive value of implantable intrathoracic impedance monitoring as a predictor of heart failure hospitalizations: the SENSE-HF trial.
        Eur Heart J. 2011; 32: 2266-2273
        • Heist E.K.
        • Herre JM
        • Binkley PF
        • Van Bakel AB
        • Porterfield JG
        • Porterfield LM
        • et al.
        Analysis of different device-based intrathoracic impedance vectors for detection of heart failure events (from the Detect Fluid Early from Intrathoracic Impedance Monitoring study).
        Am J Cardiol. 2014; 114: 1249-1256
        • Auricchio A.
        • Gold MR
        • Brugada J
        • Nölker G
        • Arunasalam S
        • Leclercq C
        • et al.
        Long-term effectiveness of the combined minute ventilation and patient activity sensors as predictor of heart failure events in patients treated with cardiac resynchronization therapy: Results of the Clinical Evaluation of the Physiological Diagnosis Function in the PARADYM CRT device Trial (CLEPSYDRA) study.
        Eur J Heart Fail. 2014; 16: 663-670
        • Albert N.M.
        • Barnason S
        • Deswal A
        • Hernandez A
        • Kociol R
        • Lee E
        • et al.
        Transitions of care in heart failure: a scientific statement from the American Heart Association.
        Circ Heart Fail. 2015; 8: 384-409
        • Yancy C.W.
        • Jessup M
        • Bozkurt B
        • Butler J
        • Casey Jr, DE
        • Colvin MM
        • et al.
        2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America.
        Circulation. 2017; 136: e137-e161
        • Crossley G.H.
        • Boyle A
        • Vitense H
        • Chang Y
        • Mead RH
        • CONNECT Investigators
        The CONNECT (Clinical Evaluation of Remote Notification to Reduce Time to Clinical Decision) trial: the value of wireless remote monitoring with automatic clinician alerts.
        J Am Coll Cardiol. 2011; 57: 1181-1189
        • Bohm M.
        • Drexler H
        • Oswald H
        • Rybak K
        • Bosch R
        • Butter C
        • et al.
        Fluid status telemedicine alerts for heart failure: a randomized controlled trial.
        Eur Heart J. 2016; 37: 3154-3163