Bioimpedance: Has Its Time Finally Come?

Bioimpedance, themeasurement of the frequency-dependent resistance to the flow of low energy electrical currents between two electrodes on or within the body, has been investigated for various medical applications for over 50 years. Applications in cardiology date back to at least 1966 and have included attempts to estimate stroke volume, cardiac output, cardiac contractility and lung fluid volume. The interest in lung fluid volume relates directly the goal of providing an objective, actionablemeasure of volume status in patients with heart failure to guide adjustments of medical therapies, predominantly diuretic doses. This is feasible since when salt and water content increase in the lungs, electrical currents passmore easily through the tissue and bioimpedance is therefore decreased. Despite prior advances in electrode technology, electric circuitry, signal processing algorithms and attempts to optimize the number and location of electrodes the accuracy, reproducibility and clinical utility of information retrievable from bioimpedance devices continue to be challenged. However, in this issue of the Journal, Shochat and colleagues report rather remarkable results of a study designed to test whether periodic assessment of lung volume status by a new bioimpedance device, can reduce hospitalizations and mortality in patients with heart failure. The study employed a device manufactured by RSMM Ltd (Tel Aviv) who supplied the devices for the study. In contrast to traditional bioimpedance devices that measure the total impedance between one or more pairs of electrodes placed on the body surface (generally on the same side of the body), the RSMM approach measures impedance across the chest (anterior to posterior, with the electrical currents transmitted through the lung) and with the thoracic wall impedance subtracted from the total impedance. This latter calculation allows for estimation of the impedance of the lung tissue itself, which normally accounts for only a relatively small portion of the total impedance. By isolating net lung impedance (LI) in this manner, theory suggests that this approach should yield greater sensitivity and specificity in detecting changes in lung fluid content. Because of this important differentiating technological feature, the results of this study should not be extrapolated to other bioimpeance systems. The present study randomized 256 patients from two hospitals to either standard of care (SOC, the Control group) or SOC plus monthly net LI measurement-guided medication adjustments (the Active group), predominantly changes in diuretic doses. All patients were seen in the clinic once per month, and more often if they had a change in medications. All patients underwent bioimpedance measurements but the results were made available to the treating physicians only for the active group. Adjustments in medications were protocolized to enhance uniformity of how the bioimpedance results were used in the Active group. The mean duration of follow up was 48 months. The time course of changes in net LI prior to, during and following heart failure admissions paralleled what has been reported for invasively measured pulmonary artery pressure, starting to decrease (signifying increased lung fluid) ~ 3 weeks prior to a hospitalization. Net LI measurements also showed that most patients were discharged from hospital prior to complete return of volume status to baseline values. The main findings of the study were a reduction in the rate of acute heart failure (AHF) hospitalizations at one year from 1.23 to 0.52 events/pt.yr (hazard ratio 0.51, P < .001) and a reduction in acute heart failure hospitalizations during the entire follow up period from 0.94 to 0.41 events/pt.yr (hazard ratio 0.63, P < .001). All-cause hospitalizations were also reduced, but this appeared entirely due to reductions inAHF and cardiac hospitalizations since the rate of non-cardiac hospitalizations, which accounted for ~1/3 of all hospitalizations in the Control group,werenot reduced.Remarkably, totalmortality, heart failurerelatedmortality and cardiacmortalitywere all decreased in the Active group compared to controls. A host of other secondary endpointswere also favorably impacted in theActivegroup.There was a NYHA functional class-dependence of the rates of hospitalizations (increasing with NYHA class), but the impact of monitoring was present in all classes. Shochat and colleagues discuss several of the limitations of the study. First and foremost, this was a single blind study From the Cardiovascular Research Foundation, New York, New York and Division of Cardiology, Columbia University, New York, New York. Manuscript received July 21, 2016; revised manuscript accepted July 21, 2016. Department of Medicine, Columbia University, 177 Fort WashingtonAve, NewYork, NewYork 10032. Tel.: +1 201 294 6081; Fax: +1 917 398 1690. E-mail: [email protected] See page 724 for disclosure information. 1071-9164/$ see front matter © 2016 Published by Elsevier Inc. http://dx.doi.org/10.1016/j.cardfail.2016.07.435