Annals of Cardiac Anaesthesia Annals of Cardiac Anaesthesia Annals of Cardiac Anaesthesia
Home | About us | Editorial Board | Search | Ahead of print | Current Issue | Archives | Submission | Subscribe | Advertise | Contact | Login 
Users online: 1129 Small font size Default font size Increase font size Print this article Email this article Bookmark this page


    Advanced search

    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Email Alert *
    Add to My List *
* Registration required (free)  

    Materials and me...
    Article Figures
    Article Tables

 Article Access Statistics
    PDF Downloaded392    
    Comments [Add]    
    Cited by others 5    

Recommend this journal


Table of Contents
Year : 2014  |  Volume : 17  |  Issue : 4  |  Page : 273-277
Comparison between continuous non-invasive estimated cardiac output by pulse wave transit time and thermodilution method

1 Department of Anaesthesia, Drexel University College of Medicine, Philadelphia, PA, USA
2 Department of Anesthesia, All India Institute of Medical Sciences, New Delhi, India
3 Department of Anesthesia and Perioperative Care, University of California, San Francisco, CA, USA

Click here for correspondence address and email

Date of Submission17-Jan-2014
Date of Acceptance17-Aug-2014
Date of Web Publication1-Oct-2014


Aims and Objectives: Cardiac output (CO) measurement is essential for many therapeutic decisions in anesthesia and critical care. Most available non-invasive CO measuring methods have an invasive component. We investigate "pulse wave transit time" (estimated continuous cardiac output [esCCO]) a method of CO measurement that has no invasive component to its use. Materials and Methods: After institutional ethical committee approval, 14 adult (21-85 years) patients undergoing surgery and requiring pulmonary artery catheter (PAC) for measuring CO, were included. Postoperatively CO readings were taken simultaneously with thermodilution (TD) via PAC and esCCO, whenever a change in CO was expected due to therapeutic interventions. Both monitoring methods were continued until patients' discharge from the Intensive Care Unit and observer recording values using TD method was blinded to values measured by esCCO system. Results: Three hundred and one readings were obtained simultaneously from both methods. Correlation and concordance between the two methods was derived using Bland-Altman analysis. Measured values showed significant correlation between esCCO and TD ( r = 0.6, P < 0.001, 95% confidence limits of 0.51-0.68). Mean and (standard deviation) for bias and precision were 0.13 (2.27) L/min and 6.56 (2.19) L/min, respectively. The 95% confidence interval for bias was - 4.32 to 4.58 L/min and for precision 2.27 to10.85 L/min. Conclusions: Although, esCCO is the only true non-invasive continuous CO monitor available and even though its values change proportionately to TD method (gold standard) with the present degree of error its utility for clinical/therapeutic decision-making is questionable.

Keywords: Continuous cardiac output monitor; Non-invasive cardiac output measurement; Pulse wave transit time

How to cite this article:
Sinha AC, Singh PM, Grewal N, Aman M, Dubowitz G. Comparison between continuous non-invasive estimated cardiac output by pulse wave transit time and thermodilution method. Ann Card Anaesth 2014;17:273-7

How to cite this URL:
Sinha AC, Singh PM, Grewal N, Aman M, Dubowitz G. Comparison between continuous non-invasive estimated cardiac output by pulse wave transit time and thermodilution method. Ann Card Anaesth [serial online] 2014 [cited 2022 Jan 23];17:273-7. Available from:

   Introduction Top

Cardiac output (CO) measurement is essential for the management of a number of surgical and Intensive Care Unit (ICU) patients. A continuous, non-invasive method would be ideal in assisting with many therapeutic decisions. Historically an accurate measurement has required a pulmonary artery catheter (PAC). The use of the PAC initially became extensive because of the ease of data collection and perceived accuracy of recorded values; however, over decades of PAC use, complications were eventually encountered, and some severe ones lead to its diminishing use. In addition inserting the PAC is inherently risky because it is invasive, time-consuming and requires a sterile field and a high level of skill and training in the operator. For this reason, among others, new devices have been developed to measure CO. The most commonly used among these is three-dimensional echocardiography. This is a device, with proven accuracy, however it requires significant skill to use and requires a transesophageal echocardiography probe to be utilized. Due to these limitations, long-term or continuous use of echocardiography based estimation of CO in non-cardiac ICU settings is limited.

It has been demonstrated that CO can be measured using the technology of pulse wave transit time (PWTT). PWTT is an established technology that uses the correlation of the heart activity to the electrocardiogram (ECG). It measures the time taken by the pulse (heartbeat) to propagate to another sensor on the arm, finger or leg. This sensor can be saturation of peripheral oxygen (pulse oximeter, SpO 2 ) or non-invasive blood pressure (NIBP) sensor. Currently, the technology is mostly utilized for atherosclerosis screening. Estimated continuous cardiac output (esCCO) is a technology that uses this timing to measure CO on a continuous yet completely non-invasive principle.

   Materials and methods Top

This prospective observer-blinded trial recruited postoperative patients (aged 21-85 years) requiring routine CO measurement using thermodilution (TD) in ICUs at a tertiary care hospital following ethical clearance from the hospital's Institutional Review Board. After obtaining informed/written consent from the patient or surrogate (as applicable in the case), 14 patients were recruited in the study over a period of 8 months. Patients with marked arrhythmias or requiring intra-aortic balloon pump or requiring circulatory support during the ICU course were excluded from the study.

For the surgical procedure a balanced anesthesia technique using propofol (1-2 mg/kg), fentanyl (1-2 mg/kg) and intermediate acting muscle relaxant was used for all the patients. After induction, a 7.5F PAC (Swan-Ganz Standard TD PAC; Edwards Life Sciences, USA) was placed in the internal jugular vein via an 8.5F introducer sheath. After confirmation of the position of PAC, CO monitoring was initiated using the TD technique (Vigilance II Monitor, Edwards Life sciences, USA). The esCCO LifeScope monitor was then connected to the patient. This esCCO LifeScience monitoring system (Main Unit Model MU-910RK, Bedside Model BSM-9101K) included a SpO 2 sensor (pulse oximeter), NIBP cuff, and set of EKG electrodes. The esCCO system estimates the PWTT from the time gap between ECG-R wave and the peripheral pulse detected by the pulse oximeter. The principle of esCCO is an inverse correlation between stroke volume (SV) and PWTT. Based on this principle, esCCO is calculated by the following equation: EsCCO = K × (a × PWTT + b) × heart rate (Where a is a fixed value which is decided experimentally by the past esCCO clinical studies). The constants K and b are individualized for each patient. These values are dependent on physical profile (age, weight, height) of the patient . The present protocol (software) tends to negate the effect of the change of vascular properties (vasoconstriction or arteriosclerosis) on the effect of travel time of pulse wave thus reflecting changes in SV alone to measure the CO.

Blood pressure for the esCCO data was collected from an Invasive arterial blood pressure line if present or else NIBP was used for the same (Arterial catheter was not placed for the purposes of the study). SpO 2 was measured using a finger sensor. The NIBP cuff was placed on the opposite arm from the SpO 2 sensor and the readings were recorded at 1-5 min depending on the patient's hemodynamic status. The patients' demographic data (age, gender, weight, height) were entered into the LifeScope system, thereafter, NIBP cycle was allowed to run for the initial calibration of the esCCO system. The measuring program was then initiated on a computer connected to the esCCO monitor. This program continued to record the esCCO readings until the monitor was removed on discharge from ICU or completion of study period. This monitor and the data it generated were kept covered in the ICU, away from the treating team, so as not to affect any clinical decisions .

As per the hospital protocol TD based reading were only taken if any clinical intervention for patient management warranted the use of present CO. However, if no repeat measurements were needed for any intervention/management routine repeat readings were noted at least every 6 hourly. In addition, readings were taken after any clinical intervention that would be expected to change CO, like a fluid bolus or blood transfusion. The TD readings were recorded using an electronic system after ensuring that hemodynamic variables were stable during the preceding 5 min interval, and no intervention was done immediately prior to readings (within these same last 5 min). Estimated CO shown by esCCO were noted at time points that coincided with the time when CO values were estimated using the TD method by an independent blinded observer. Continuous CO with esCCO was monitored till the patient was discharged from the ICU or until completion of study duration.

Additionally, the following data were also collected - type of surgery, duration of surgery, American Society of Anesthesiologists physical status and vital signs (hemodynamic parameters and pulse oximeter hemoglobin saturation) at time of CO estimation using both the methods. The data obtained was analyzed using the Statistical Package for Social Sciences (SPSS) 21 (IBM Inc, Chicago) for Macintosh. Data was tested for normality using Kolmogorov-Smirnov test. Parametric data was expressed as mean ± standard deviation (SD). Descriptive statistics was used for demographic data. Correlation analysis and Bland Altman method were performed to compare esCCO and CO. [1] Further as suggested by original paper by Bland and Altman, for analyzing repeated values from a single subject (as in our case CO values measured from the same subject at different chronological points), principle of "comparing varying true values" was used. As different time points where measurements were made on the same patient were independent of each other assumption as "all repeated differences for a single subject are independent" was assumed. [2] As described in the above statistical analysis methodology, we interpreted the results with the possibility of actual confidence intervals (CIs) of resulting variables to be slightly wider than those calculated. Correlation coefficient (r) and P value were obtained using above modifications in the Bland and Altman analysis. The statistical significance level, a, was set at 0.05. Sample size calculation was done by using the two means equation in the SPSS 21 (IBM Inc.). Mean and standard deviation of CO obtained from the study by Yamada et al. were used to calculate the sample size. [3] The mean difference between study and control group was 0.13, and the detectable difference was set at 80% of SD. A two-tailed hypothesis testing concluded that 787 data points were needed to achieve a power of 80%, with an alpha error rate of 5%. With an estimation of 10 data points collected from each patient, 40 patients were to be enrolled. An interim analysis after 14 patients showed a statistically significant difference between the two methods. The percentage error between esCCO and CO (2 SD of bias divided by mean CO) was 69%. Such a high estimated error disapproved its clinical utility and thus further enrolment was stopped.

   Results Top

The patient characteristics are shown in [Table 1]. The indications of surgery are shown in [Table 2]. A total of 301 data points were obtained from the 14 patients. The data points showed a significant correlation between esCCO and CO (r = 0.6, P < 0.001, 95% confidence limits of 0.51-0.68) [Figure 1]. Mean and (SD) for bias and precision were 0.13 (2.27) L/min and 6.56 (2.19) L/min, respectively. The 95% CI for bias was - 4.32-4.58 and for precision it was 2.27-10.85 [Figure 2]. Percentage error between esCCO and CO (2 SD of bias divided by mean CO) was 69%.
Table 1: Demographic parameters of included population

Click here to view
Figure 1: Correlation between estimated continuous cardiac output (pulse wave transit time) and cardiac output by thermodilution

Click here to view
Figure 2: Bland-Altman plot between estimated continuous cardiac output (Pulse wave transit time) and cardiac output by thermodilution

Click here to view
Table 2: Surgical indication of included patients

Click here to view

   Discussion Top

Our study found that the bias between the values was only 0.13 (2.27). Although this value is very small but the clinical significance can only be interpreted by understanding the limitations of statistics here. Bias in simple terms means mathematical difference of the measured value, CO measured by the new method (esCCO) from the mean of the two methods, CO by TD and by esCCO. It is important to note the values we obtained from esCCO method were both higher and lower than what we considered the 'gold standard' TD method. Thus 0.13 , actually is a result of both these higher and lower values cancelling out to produce a very small value. The mean CO throughout the study (as measured by TD was 6.49 ± 2.09 L/min) and if 95% CI of bias is now evaluated being - 4.32-4.58 (range = 8.90), the variation is even more than the actual value of CO. In other words, esCCO can estimate higher or lower values by error almost equal to 3/4 th of actual CO. In a clinical scenario where intervention are to be made on the basis of CO values (e.g. in patients with shock/hypotension planned with the goal directed fluid therapy, whether to administer fluids in low CO or vasopressors in shock with high CO), such degree of error is likely to guide completely toward a wrong therapeutic intervention. Due to this almost 70% error, if an actual CO of 6 L/min is erroneously reported to be as low as 2 L/min (error of around 66% by esCCO) over transfusion of fluids can precipitate complications like pulmonary edema. As per our findings, the utility of CO values obtained from esCCO method for making therapeutic decisions is thus questionable. Similar to our findings, Ball and colleagues in patients undergoing cardiac surgery reported mean bias value of 0.80 L/min with 95% CI of - 2.00-3.61 L/min. They also, similarly challenged the clinical utility of esCCO method because of large variations in the level of agreement of both the tests. [4] Studying both postoperative and patients in ICU, Yamada et al. evaluated esCCO in comparison to TD in 587 data points reported a bias of 0.13 L/min but with a much smaller CI of only 0.04-0.22 L/min, that is, more clinically acceptable limits of agreement. [3]

One of the most unique advantages as proposed by the esCCO protocol is its complete non-invasive nature. Unlike previously developed methods of measuring CO, it does not require an arterial line or even a previous calibration. [5-7] Ishihara and colleagues based on previously available data from 7 Japanese universities used an updated model for esCCO. They mainly focused on esCCO calibration and an exclusion algorithm for patients based upon high error rates found in previous data. This led to the elimination of patients who underwent surgeries on cardiopulmonary bypass. They found that linear correlation values between esCCO and TD method " r" values improved from 0.57 to 0.64. Despite this, the percentage of error remained high and they concluded that ability of the new esCCO system is not clinically acceptable based on percentage error (being 69.6%) and polar plots analysis, even though its trending ability is comparable with currently available arterial waveform analysis methods. [8]

Comparisons of esCCO with other non-invasive, reliable methods like transthoracic echocardiography (TTE) have also shown limitations of esCCO system. Recently, Fischer et al. showed that changes in CO values as measured by esCCO before and after a fluid bolus did not correlate well with those measured with TTE. They reported that Bias, precision, and limits of agreement were 0.25 L/min, 2.4 L/min, and - 4.4-4.9 L/min, respectively. Our results show that with the present software, esCCO has an error of nearly 80% in its measurement. Thus its clinical utility with such a large degree of error is debatable. [9] In a similar study Bataille et al. found an error of 49% and also changes in esCCO values failed to follow the changes in the CO as seen with TTE. Thus, they also challenged esCCO's actual utility in clinical decision making. [10] Tsutsui et al. compared esCCO (Edwards Lifesciences) with TD method. They reported CCO and esCCO showed a high correlation with a correlation coefficient of 0.84 in 496 total data points, and 95% limits of agreement between these two methods were - 2.49-2.35 L/min. These values suggest that both methods are fairly comparable in estimating continuous CO. [11] Although above results with esCCO have not been very promising and at least for the present time, TD remains the gold standard; however, TD method is known to show an error of (actually difference of CO) 15-20% when measurements are taken during inspiration followed by expiration. [12,13] Critchley and Critchley in their meta-analysis proposed a statistical method to compare TD to other non-invasive methods and estimated that newer devices must have accuracy within 30% of CO measured by TD to be considered equivalent. If the accuracy of the newer method falls beyond 30% (that is more than intrinsic error in TD) the method must be rejected. [14] The error reported in our study along with above-cited trials probably highlight the need for further improvement in esCCO protocols. Another possible reason for such contrasting results seen in multiple studies can be demographic profile differences in studied population. Most studies that have shown higher degree of accuracy for esCCO originate from Japan whereas studies from other parts of the world seem to not validate the reported accuracy. The basic design of esCCO system, in fact, relies upon the time for pulse wave to travel from the heart to peripheral vessel where the sensor is used. Any anatomical (body habitus) differences are obviously likely to effect the duration for pulse travel time. Most patients participating in our study were either Caucasians or African-Americans, who are significantly different in their body habitus from the Japanese counterpart population. Other possible limitations that exist with the present measurement protocol are that a stenotic valve can significantly lower the pulse generation speed, whereas the regurgitant lesion generates fast but small volume pulse. These are also the key areas where esCCO method has scope for improvement.

Our results show limitation of accuracy of esCCO and highlight the need for improvement in protocols that are used in calculations. Also, a possibility of the need for better calibration accounting for body habitus of various population subgroups is also required. In various valvular lesions, the pulse velocity is also likely to be affected.

In conclusion, the idea and technique used in esCCO method are presently the most convenient and the only truly non-invasive method for estimating continuous CO. However, utilizing the current algorithms produces an unacceptable degree of error and its utility for clinical/therapeutic decision-making is questionable.

   References Top

1.Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.  Back to cited text no. 1
2.Bland JM, Altman DG. Agreement between methods of measurement with multiple observations per individual. J Biopharm Stat 2007;17:571-82.  Back to cited text no. 2
3.Yamada T, Tsutsui M, Sugo Y, Sato T, Akazawa T, Sato N, et al. Multicenter study verifying a method of noninvasive continuous cardiac output measurement using pulse wave transit time: A comparison with intermittent bolus thermodilution cardiac output. Anesth Analg 2012;115:82-7.  Back to cited text no. 3
4.Ball TR, Tricinella AP, Kimbrough BA, Luna S, Gloyna DF, Villamaria FJ, et al. Accuracy of noninvasive Estimated Continuous Cardiac Output (esCCO) compared to thermodilution cardiac output: A pilot study in cardiac patients. J Cardiothorac Vasc Anesth 2013;27:1128-32.  Back to cited text no. 4
5.Guarracino F, Stefani M, Lapolla F, Cariello C, Doroni L, Danella A, et al. Monitoring cardiac output with Flo Trac Vigileo TM . Br J Anaesth 2007;99:142.  Back to cited text no. 5
6.Alhashemi JA, Cecconi M, Hofer CK. Cardiac output monitoring: An integrative perspective. Crit Care 2011;15:214.  Back to cited text no. 6
7.Sugo Y, Ukawa T, Takeda S, Ishihara H, Kazama T, Takeda J. A novel continuous cardiac output monitor based on pulse wave transit time. Conf Proc IEEE Eng Med Biol Soc 2010;2010:2853-6.  Back to cited text no. 7
8.Ishihara H, Sugo Y, Tsutsui M, Yamada T, Sato T, Akazawa T, et al. The ability of a new continuous cardiac output monitor to measure trends in cardiac output following implementation of a patient information calibration and an automated exclusion algorithm. J Clin Monit Comput 2012;26:465-71.  Back to cited text no. 8
9.Fischer MO, Balaire X, Le Mauff de Kergal C, Boisselier C, Gérard JL, Hanouz JL, et al. The diagnostic accuracy of estimated continuous cardiac output compared with transthoracic echocardiography. Can J Anaesth 2014;61:19-26.  Back to cited text no. 9
10.Bataille B, Bertuit M, Mora M, Mazerolles M, Cocquet P, Masson B, et al. Comparison of esCCO and transthoracic echocardiography for non-invasive measurement of cardiac output intensive care. Br J Anaesth 2012;109:879-86.  Back to cited text no. 10
11.Tsutsui M, Yamada T, Sugo Y, Sato T, Akazawa T, Sato N, et al. Comparison of continuous cardiac output measurement methods: Non-invasive estimated CCO using pulse wave transit time and CCO using thermodilution. Masui 2012;61:1011-7.  Back to cited text no. 11
12.Elkayam U, Berkley R, Azen S, Weber L, Geva B, Henry WL. Cardiac output by thermodilution technique. Effect of injectate's volume and temperature on accuracy and reproducibility in the critically Ill patient. Chest 1983;84:418-22.  Back to cited text no. 12
13.Nelson LD, Anderson HB. Patient selection for iced versus room temperature injectate for thermodilution cardiac output determinations. Crit Care Med 1985;13:182-4.  Back to cited text no. 13
14.Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput 1999;15:85-91.  Back to cited text no. 14

Correspondence Address:
Ashish C Sinha
Department of Anaesthesia, Drexel University College of Medicine, Philadelphia, PA
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/0971-9784.142059

Rights and Permissions


  [Figure 1], [Figure 2]

  [Table 1], [Table 2]

This article has been cited by
1 Central Noninvasive Hemodynamics in Combined Spinal Epidural Anesthesia with Expansion of the Epidural Space in Obese Patients
N. V. Davydov, I. G. Trukhanova, ?. D. Gureev, Yu. G. Kutyreva
Messenger of ANESTHESIOLOGY AND RESUSCITATION. 2022; 18(6): 90
[Pubmed] | [DOI]
2 Individualized Hemodynamic Management in Sepsis
Marcell Virág, Tamas Leiner, Mate Rottler, Klementina Ocskay, Zsolt Molnar
Journal of Personalized Medicine. 2021; 11(2): 157
[Pubmed] | [DOI]
3 Continuous Estimation of Cardiac Output in Critical Care: A Noninvasive Method Based on Pulse Wave Transit Time Compared with Transpulmonary Thermodilution
Ulrike Ehlers, Rolf Erlebach, Giovanna Brandi, Federica Stretti, Richard Valek, Stephanie Klinzing, Reto Schuepbach
Critical Care Research and Practice. 2020; 2020: 1
[Pubmed] | [DOI]
4 Estimated continuous cardiac output based on pulse wave transit time in off-pump coronary artery bypass grafting: a comparison with transpulmonary thermodilution
Alexey A. Smetkin,Ayyaz Hussain,Evgenia V. Fot,Viktor I. Zakharov,Natalia N. Izotova,Angelika S. Yudina,Zinaida A. Dityateva,Yanina V. Gromova,Vsevolod V. Kuzkov,Lars J. Bjertnæs,Mikhail Y. Kirov
Journal of Clinical Monitoring and Computing. 2016;
[Pubmed] | [DOI]
5 The Ability of esCCO™ and ECOM™ Monitors to Measure Trends in Cardiac Output During Alveolar Recruitment Maneuver After Cardiac Surgery
Magalie Thonnerieux,Brenton Alexander,Catherine Binet,Jean-François Obadia,Olivier Bastien,Olivier Desebbe
Anesthesia & Analgesia. 2015; 121(2): 383
[Pubmed] | [DOI]