Some small studies suggested that telemonitoring may improve HF outcomes and reduce readmissions. The advantage of this approach is its scalability, because this protocolized approach utilized all the standard instruments (telephone and weight scale) and approach (patient outreach and proactive interventions) commonly deployed in everyday clinical practice. A prospective multicenter, randomized, controlled trial involving HF patients compared usual care with a telephone-based interactive voice-response system to collect weight and symptom data and then discuss this information with the clinician. They found no meaningful reduction of death or readmission when the telemonitoring arm was compared to usual care. One potential issue is the sustainability and adherence of the strategy. Of note, 14% of the patients randomized to the telemonitoring arm never used the system, and by the end of the study, only 55% of the patients were still using the system at least 3 times/wk.⁷ Additionally, this approach may have created significant delays as it required all the data to be discussed with the physician, which may delay the ability to intervene and inadvertently increased clinicians’ work burden.⁶ Clearly, such findings do not necessarily negate the use of these widely available tools for patient communications,
but its routine implementation with such a design may not be effective in reducing hospitalizations.

There have been additional randomized, controlled, telemedicine trials, notably Telemedical Interventional Monitoring in Heart Failure (TIM-HF) and Efficacy of Telemedical Interventional Management in Patients with Heart Failure (TIM-HF2) based out of Germany.⁸,⁹ TIM-HF was published back in 2010 and it highlighted the study design behind this prospective, randomized, multicenter trial. The study enrolled 1571 patients and evaluated the impact of telemedicine on mortality in HF patients over the course of 21 months. Patients were randomized in a 1:1 manner and the intervention arm (remote monitoring) consisted of daily electrocardiograms (ECGs), blood pressures, and body weights in tandem with telephone support. The data obtained were analyzed by a physician-led team to help guide patient care in real time. The allcause mortality rate in the control arm was 11.34% and in the remote patient management arm was 7.86% (hazard ratio [HR] 0.70; 95% confidence interval [CI] 0.50-0.96; P = 0.0280), though cardiovascular mortality was not statistically significantly different (HR 0.671; 95% CI 0.45-1.01; P = 0.0560).⁸,⁹ Additional data also showed a reduction in lost days because of hospitalization with the intervention arm. Although the results are promising, important limitations to the generalizability of this approach include that the remote system was designed for German health care systems and the full-time staff necessary to deliver real-time care may limit the scalability of this approach.

Although telemonitoring with telephone-based systems did not prove to be beneficial in large randomized trials, some investigators have explored the use of measuring thoracic impedance as the basis of home surveillance tools. Impedance, defined as the resistance to the flow of current between two points (measured by special equipment, denoted with the symbol Z, and expressed in ohms), is not a common parameter measured in the clinical setting. When this is used to measure flow across the chest, it has been shown to correlate with thoracic fluid content because fluid conducts electricity better than air or tissue (hence lower impedance with more fluid accumulation as the thoracic tissue has higher impedance than blood or fluid). This difference allows us to use this property to determine changes in filling pressures, but there are also many factors that can confound the results of measurement including body posture, underlying valve disease, arrhythmias, sensor positioning, and also the computer algorithms used. Additionally, there is currently no standardized definition of fluid index, and more guidance and clarity are required on the proper timing of interventions. Impedance can be measured using two approaches, including an external band electrode method and implanted device-based method.¹⁰

The concept and feasibility of remote monitoring using these techniques has been demonstrated by implantable hemodynamic systems such as CardioMEMS. CardioMEMS is implanted in the distal branch of the descending pulmonary artery and provides pressure readings using a home console. The data are immediately available for review by the health care team and provide trends of information. The CardioMEMS Heart Sensor Allows Monitoring of Pressure to Improve Outcomes in NYHA Class III Heart Failure Patients trial was a prospective, multicenter, clinical trial consisting of 550 patients that showed reduction in hospitalizations for patients with New York Heart Association Class III HF, but it is not feasible for a large cohort of HF patients because of cost and invasive nature.¹¹ Other implantable monitoring studies all demonstrated that changes in filling pressures precede significant changes in weight.¹⁰,¹¹,¹²,¹³ Specifically, HF rehospitalizations are preceded by rising filling pressures
even 2 weeks prior to detectable weight gain or symptoms.³ Studies have shown that half of the patients admitted for acute decompensated HF had a weight gain of 2 lb or more.¹⁴ Continuous monitoring of hemodynamic and impedance parameters would theoretically allow to detect signs of acute decompensated HF earlier when compared to monitoring systems relying entirely on weight trends and patient symptoms.

Currently, there are various wearable devices available including skin patches, necklaces, and bracelets that aim to accomplish this goal. Inan and colleagues at the Georgia Institute of Technology studied a wearable patch sensor that measured seismocardiogram signals that are waves that correspond to the opening and closing of the aortic valve and ejection of blood.² The battery life for the sensor on one charge is about 50 hours, which would allow for 24-hour recordings. They hypothesized that with the 6-minute walk test (6MWT), seismocardiogram signals would change less in American College of Cardiology Stage C HF patients when compared to compensated HF patients, because they would have less cardiac reserve. The study enrolled 45 patients (13 decompensated HF patients, 32 compensated HF patients) and did not have a control arm. They compared the seismocardiogram signals at rest and after 6MWT for similarity using graph mining: a higher graph similarity score (GSS) suggested reduced cardiovascular reserve. A significant difference in GSS was identified between the two groups. Additionally, GSS decreased in patients who were admitted with decompensated HF and became clinically euvolemic by discharge. These findings suggest that seismocardiogram signals along with machine learning algorithms can analyze important variables to assess the volume status of HF patient, but additional research is needed to replicate their findings.