Elsevier

Journal of Cardiac Failure

Volume 22, Issue 11, November 2016, Pages 925-927
Journal of Cardiac Failure

Estimating Cardiac Output: Brief Report
Improved Accuracy of Automated Estimation of Cardiac Output Using Circulation Time in Patients with Heart Failure

https://doi.org/10.1016/j.cardfail.2016.05.007Get rights and content

Abstract

Background

Lung-to-finger circulation time of oxygenated blood during nocturnal periodic breathing in heart failure patients measured using polysomnography correlates negatively with cardiac function but possesses limited accuracy for cardiac output (CO) estimation.

Methods and Results

CO was recalculated from lung-to-finger circulation time using a multivariable linear model with information on age and average overnight heart rate in 25 patients who underwent evaluation of heart failure. The multivariable model decreased the percentage error to 22.3% relative to invasive CO measured during cardiac catheterization.

Conclusions

This improved automated noninvasive CO estimation using multiple variables meets a recently proposed performance criterion for clinical acceptability of noninvasive CO estimation, and compares very favorably with other available methods.

Section snippets

Material and Methods

Reference cardiac output was measured invasively in 31 patients who underwent evaluation of HF, principally using Fick's method.4 The patients also underwent polysomnography (PSG) (SAS-2100, Nihon Kohden Corp., Japan) within 2.0 ± 2.1 days of the catheterization. Patients were excluded if they had acute phase HF (eg, intubation, non-invasive positive pressure ventilation, high oxygen supplementation), and had to be stable enough to not necessitate a change of medication. The study protocol was

Results

Table 2 shows the coefficients, standard errors of the coefficients, and P values of the coefficients for the model that is based on the LFCT alone and the model that is based on the 3 predictor variables of LFCT, age, and AveHR. In both cases, the models were developed using the full set of 25 measurements. The R2 value, the fraction of CO variance accounted for by the predictor variables included in the model, was 0.443 (P < .001) for the single-variable model and 0.787 (P < .001) for the

Discussion and Conclusions

A limitation of this study is that the measurement using the proposed method and the reference invasive measurement would ideally be performed at the same time, but this is of course not possible because the patients are usually awake in the catheter laboratory. Therefore, to minimize errors that could result from the time separation between the 2 measurements, we included only patients whose cardiac condition was stable after recovering from decompensated HF. Furthermore, to quantify the

Disclosures

S.A. has received research funds from Teijin Home Healthcare (Tokyo, Japan) and Philips Respironics (Murrysville, Pennsylvania, USA).

Acknowledgments

H.D. was partially supported by a Discovery Grant from the Natural Sciences and Engineering Research Council of Canada. The authors also thank the anonymous reviewers for comments that helped improve the paper.

References (6)

There are more references available in the full text version of this article.
View full text
© 2016 Elsevier Inc. All rights reserved.