BJA Advance Access originally published online on January 11, 2007
British Journal of Anaesthesia 2007 98(2):176-182; doi:10.1093/bja/ael341
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Semi-invasive monitoring of cardiac output by a new device using arterial pressure waveform analysis: a comparison with intermittent pulmonary artery thermodilution in patients undergoing cardiac surgery
Department of Anaesthesiology and Intensive Care Medicine, Klinikum Ludwigshafen, Bremserstr. 79, 67063 Ludwigshafen, Germany
* Corresponding author. E-mail: j-mayer{at}gmx.de
Accepted for publication November 21, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
BACKGROUND: Thermodilution technique using a pulmonary artery catheter (PAC) is a widely used method to determine cardiac output (CO). It is increasingly criticized because of its invasiveness and its unclear risk–benefit ratio. Thus, less invasive techniques for measuring CO are highly desirable. We compared a new, semi-invasive device (FloTrac/VigileoTM) using arterial pressure waveform analysis for CO measurement in patients undergoing cardiac surgery with bolus thermodilution measurements.
METHODS: Forty patients undergoing coronary artery bypass grafting or valve repair were enrolled. A PAC was inserted and routine radial arterial access was used for semi-invasive determination of CO with the Vigileo. CO was measured simultaneously by bolus thermodilution and the Vigileo technique after induction of anaesthesia (T1), before cardiopulmonary bypass (CPB) (T2), after CPB (T3), after sternal closure (T4), on arrival in the intensive care unit (ICU) (T5), and 4 h (T6), 8 h (T7), and 24 h after surgery (T8). CO was indexed to the body surface area (cardiac index, CI).
RESULTS: A total of 244 pairs of CI measurements were analysed. Bias and precision (1.96 SD of the bias) were 0.46 litre min–1 m–2 and ± 1.15 litre min–1 m–2 (r = 0.53) resulting in an overall percentage error of 46%. Subgroup analysis revealed a percentage error of 51% for data pairs obtained intraoperatively (T1–T4), 42% in ICU (T5–T8), and 56% for values obtained during low CI (T1–T8).
CONCLUSIONS: In cardiac surgery patients, CO measured by a new semi-invasive arterial pressure waveform analysis device showed only moderate agreement with intermittent pulmonary artery thermodilution measurement.
Keywords: measurement techniques; thermodilution; monitoring; cardiopulmonary; surgery; cardiovascular
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Assessment of cardiac performance is of high importance in the management of patients undergoing cardiac surgery. Intermittent thermodilution bolus technique using a pulmonary artery catheter (PAC) is an established method to determine cardiac output (CO). However, the use of PAC has been increasingly criticized because it is an invasive technique and because of its unclear risk–benefit ratio.1–3
In recent years, less invasive techniques such as transthoracic bioimpedance, pulse dye densitometry, and Doppler techniques have been developed for measuring CO. The results of these techniques with regard to validity, practicability, and accuracy are not uniform.4 Pulse wave analysis calculating CO from peripheral arterial pressure waveform is already available, but needs to be initially calibrated by either transpulmonary thermodilution or pulmonary artery thermodilution to compensate for inter-individual differences in arterial compliance.56 A recently introduced device (FloTrac/VigileoTM, Edwards Lifesciences, Irvine, CA, USA) calculates continuous CO on arterial pressure waveform characteristics but does not require external calibration. Individual demographic data, including height, weight, age, gender, and the real-time arterial pressure waveform analysis, are used to estimate arterial compliance.
The aim of this study was to determine the accuracy of the new arterial pressure waveform device compared with intermittent thermodilution technique for measuring CO in patients undergoing elective cardiac surgical procedures.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Forty patients of American Society of Anaesthesiologists (ASA) physical status III aged 41 to 86 scheduled for elective cardiac surgery [coronary artery bypass grafting (CABG), mitral or aortic valve repair] were studied after IRB approval and obtaining written informed consent. Patients with pacemakers, history of cardiac arrhythmias, severe peripheral vascular disease, cardiac support (intra-aortic balloon pump), and persisting mitral or aortic dysfunction after surgery were excluded.
All patients received midazolam 0.1 mg kg–1 orally for premedication. Induction of anaesthesia was performed with midazolam 0.07 mg kg–1, sufentanil 1 µg kg–1 and pancuronium 0.1 mg kg–1 were injected thereafter to facilitate tracheal intubation. Ventilation was controlled to maintain normocapnia (expiratory PCO2 32–38 mmHg) using a constant fresh gas flow of 1 litre min–1 (50% air in oxygen) in a semi-closed circle system. A standard set for arterial cannulation was used for arterial access via right or left radial artery and the FloTrac sensor was placed and connected to a Vigileo monitor (software version 1.0) for semi-invasive determination of continuous CO. A balloon tipped, flow-directed PAC (7.5 F, EFV/OTD catheter; Edwards, Irvine, CA, USA), which is part of standard monitoring in our cardiac surgery patients, was placed via the right internal jugular vein and the correct position was confirmed by pressure tracings and by routine chest radiograph immediately after admission to the intensive care unit (ICU). All intravascular pressure measurements were referenced to the mid-chest level.
A Sarns 9000 heart–lung machine (Sarns Inc., Ann Arbor, MI, USA) with non-pulsatile flow mode and a hollow-fibre membrane oxygenator (Cobe Optima XPTM; Cobe Laboratories, Planegg-Martinsried, Germany) were used to provide extracorporal circulation [cardiopulmonary bypass (CPB)]. The circuit was primed with 1000 ml of lactated Ringer's solution and 500 ml of hydroxyethyl starch. Temperature was kept at mild hypothermia (bladder temperature > 33°C), and a CPB flow rate of 2.4 litre min–1 m–2 was used.
Thermodilution CO measurements with the PAC were performed with fast and constant injections of 10 ml of ice-cold 0.9% saline. At least four consecutive measurements over the entire respiratory cycle were obtained and the plausibility of every temperature curve was judged visually on the attached monitor. A difference of less than 10% between the measurements was considered appropriate. The mean thermodilution CO was calculated from four measurements and compared with the mean CO value derived from the Vigileo monitor over the same period. If arrhythmias occurred during the measurements, the results were discarded and measurements were repeated. The measured CO values were indexed to the body surface area (cardiac index, CI) using standard formula.
Ventilator settings and vasopressor therapy/inotropic support, if necessary, remained unchanged during CO measurements. All patients were admitted to the ICU and controlled mechanical ventilation was continued during the following 6 h at least. Patients were sedated with propofol (0.5–1.5 mg kg–1 h–1), but not paralysed, and their tracheas remained intubated and ventilated using a pressure-controlled ventilation mode. Peak airway pressures were adjusted to deliver a constant tidal volume of 10 ml kg–1 body weight; inspiratory/expiratory time ratio was set to 1:2, and a positive end-expiratory pressure of 5 mm Hg was used. Tracheal extubation was performed when body temperature was > 36°C, the patients breathed spontaneously with adequate blood gas variables, and haemodynamics were stable.
CI was measured simultaneously by thermodilution (TDCI) and the VigileoTM monitor (APCI) after induction of anaesthesia (T1), before (T2) and after (T3) CPB, after sternal closure (T4), on arrival in ICU (T5), and 4 h (T6), after 8 h (T7), and 24 h after surgery (T8, all patients spontaneously breathing). Haemodynamic management was carried out according to the standard practice protocol in our institution using CI data derived by the thermodilution technique. Thermodilution CO and CO measurements using APCI were obtained by two anaesthesiologists who were blinded to the corresponding CO measurement of the other method.
Statistical analysis was performed using the method described by Bland and Altman.7 Bias was defined as the mean difference between pulmonary artery thermodilution and arterial pressure waveform analysis. Precision was represented by the upper and lower limits of agreement. The limits of agreement were calculated as the bias (1.96 SD) defining the range in which 95% of the differences between the two methods were expected to lie.
The percentage error was calculated as 2 SD of the bias mean CI–1 · 100, as proposed by Critchley and Critchley.8 The mean CI values derived from thermodilution VS arterial waveform analysis were compared using a paired Student t-test.
Bias and limits of agreement and linear regression were calculated for the entire data set, separately for intraoperative and postoperative data and for all TDCI derived values in the low quartile with a CI < 2.0 litre min–1 m–2. Delta CI (
CI) was calculated for both methods to investigate changes of CI regardless of absolute accuracy. Data are presented as mean (SD), unless otherwise stated. P < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
From the 40 patients included, 27 patients were analysed intraoperatively and 33 patients were analysed in ICU, resulting in 244 data pairs of CI measurements. Thirteen patients had to be excluded from intraoperative measurement, six during the first 24 h ICU stay because of permanent arrhythmias, one patient needed invasive cardiac support in ICU (intra-aortic balloon pump). Basic demographic data and data from anaesthesia and surgery are listed in Table 1. Haemodynamics during CO determination are shown in Table 2. The TDCI values ranged from 1.3 to 4.8 litre min–1 m–2 [mean 2.3 (0.54) litre min–1 m–2], whereas the values for APCI ranged from 1.6 to 4.9 litre min–1 m–2 [mean 2.8 (0.65) litre min–1 m–2] (r = 0.53, P < 0.0001). Bias and precision were 0.46 litre min–1 m–2 and ± 1.15 litre min–1 m–2 for all CI data pairs (Fig. 1). The percentage error between all TDCI and APCI measurements was 46%. Subgroup analysis revealed a correlation coefficient of r = 0.33 for CI (P < 0.0001) and bias and precision of 0.47 litre min–1 m–2 and ± 1.22 litre min–1 m–2 for data pairs obtained intraoperatively (T1–T4, Fig. 2). On the ICU (T5–T8), linear regression resulted in a correlation coefficient of 0.57 (P < 0.0001). Bias and precision were 0.45 litre min–1 m–2 and ± 1.14 litre min–1 m–2, respectively (Fig. 3). The percentage error intraoperatively and in ICU was 51% and 42%, respectively. Bias and precision for every data point (T1–8) are shown in Table 3. Analysis of data pairs in the low quartile (n = 61, CI < 2.0 litre min–1 m–2) showed a percentage error of 56% with a mean CI difference of 0.67 litre min–1 m–2 and a precision of ± 1.16 litre min–1 m–2 (r = 0.28, P < 0.0001, Fig. 4).
CI was calculated separately for each method and data comparison revealed a bias of – 0.04 litre min–1 m–2 and a precision of ± 1.50 litre min–1 m–2 (r = 0.40, P < 0.0001, Fig. 5).
|
|
|
|
|
|
|
Average dose, expressed as mean [range]
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The concept of arterial pressure waveform analysis to measure CO has already been known for a number of years. Arterial pulse wave analysis using transpulmonary thermodilution for calibration (PiCCO, Pulsion Medical Systems, Munich, Germany) has been validated in various clinical settings and shows overall good correlation compared with pulmonary artery thermodilution.6–9 Invasive central venous and arterial access and manual calibration are required with this technique.10 There have been attempts to obtain CO from arterial pressure waveform analysis without the need of invasive calibration: Hirschl and colleagues11 reported a method of calculating the CO based upon the waveform of a finger artery with only fair agreement compared with thermodilution measurements. A subsequent re-evaluation by the same authors revealed an even higher inconsistency of this technique.12 Mukkamala and colleagues13 demonstrated a promising approach to estimate continuous CO from the peripheral arterial pulse wave without invasive calibration resulting in an overall CO error of 14.6% compared with aortic blood flow measurement in an animal model.
The clinical relevant differences of techniques for measuring CO remained a matter of debate. Inconsistent cut-off points in assessing accuracy and clinical inter-changeability and varying methods in calculating the difference between different methods have been reported. Critchley and Critchley8 demonstrated in a meta-analysis that errors of both test and reference method should be combined when assessing comparative cardiac output, which results in a percentage error of 30% to be clinically acceptable. In the present study, we fulfilled the criteria they defined in their meta-analysis for comparative CO measurement studies, including calculating mean CI, bias, limits of agreement (precision), and the percentage error.
We found a bias of 0.47 litre min–1 m–2 and a percentage error of 46% for all data pairs of CI measurement, exceeding the 30% limit of acceptance. Thus, APCI using the first generation software has to be judged as clinically not comparable with TDCI, although there was a weak, but positive correlation to thermodilution measurements (r = 0.53). Similar results were found by Costa and colleagues,14 who compared the intermittent thermodilution method with the FloTrac/Vigileo technique in 14 patients in hyperdynamic condition after liver-transplantation with a bias of 0.48 litre min–1 m–2 and precision of (1.4) litre min–1 m–2. They concluded that APCI provided ‘acceptable measurements of cardiac output in hyperdynamic conditions’. Unfortunately, they presented no percentage error. Our subgroup analysis revealed an even higher discrepancy of 51% for the CI data obtained intraoperatively. Comparative measurement of TDCI and APCI in ICU resulted in a percentage error of 42%, which is slightly better than the intraoperative error, but is also too inaccurate to be acceptable. Separate analysis of data pairs in the low quartile (CI < 2.0 litre min–1 m–2) showed a mean CI difference of 0.67 litre min–1 m–2 and a percentage error of 56%. This weak agreement may result in misleading therapeutic approaches in patients who particularly need adequate haemodynamic treatment.
There is a variety of reasons for the moderate agreement between the two techniques. The algorithms of arterial waveform analysis are based on properties of the arterial system such as impedance, peripheral vascular, resistance, and compliance.15 The aortic impedance, which is crucial to calculate the stroke volume,16 differs from patient to patient. Former approaches to arterial pulse wave analysis excluded this source of error by performing an initial invasive calibration.56 Inter-individual differences in aortic impedance may contribute to inaccuracies in calculating the cardiac output, as ‘calibration’ is performed by demographic data. This approach is supported by the results of CI analysis, where data comparison between TDCI and APCI showed a bias of only – 0.04 litre min–1 m–2. It demonstrates the ability of the FloTrac/VigileoTM device to precisely track changes of CI in individual patients regardless of the absolute accuracy.
It has been shown that the contour of the arterial pulse wave changes significantly as it traverses through the arterial tree,17 causing mismatches because of vessel tapering, bifurcations, and calibre changes. Thus, the arising wave reflections may corrupt the peripheral arterial pulse wave signal, which consequently leads to difficulties to compensate the signal error by the system software of the APCI monitor without prior calibration.
Peripheral arterial pulse wave represents interaction between the left ventricular output and the capacitance of the vascular tree.15 The fact that vasoactive treatment causes changes in the systemic vascular system may adversely affect the accuracy of APCI to a greater extent than the accuracy of TDCI. This would explain the higher percentage error obtained during intraoperative measurements, as the incidence of conditions affecting the systemic vascular system (hypothermia, fluid exchange) was higher intraoperatively and significant higher doses of vasoactive drugs were needed than subsequently in the ICU (Table 2).
The thermodilution technique, although still regarded as the clinical ‘gold-standard’, has its own limitations. Apart from increased risks for the development of arrhythmias, valvular lesions, and rupture of the pulmonary artery,18 the accuracy of thermodilution measurements can be influenced by factors such as timing of the injection within the respiratory cycle, temperature of the injectate, speed of injection, and placement of the catheter.1920 We diminished this source of methodical bias of all comparative CO studies using PAC as the reference method by following the criteria of Critchley and Critchley8 for data comparison. All participating anaesthetists were well trained to handle PAC measurements, the thermodilution indicator had a standardized volume of 10 ml and a temperature of 4°C, and no stressful interventions for the patients were undertaken during the measurements. As measuring TDCI during cardiac arrhythmias is not recommended by the manufacturer, CI data measured during arrhythmias were not included in our analysis.
One objection of the present study is that we used a FloTrac/Vigileo device that was equipped with the first software generation that is apparently no longer available.
Although less invasive and continuous methods of CO measurement would be highly desirable, pulmonary artery thermodilution provides data about cardiac filling pressures such as central venous pressure, pulmonary artery pressure, pulmonary artery occlusion pressure, and mixed venous oxygen saturation. This additional information could be a valuable tool in assessing the haemodynamic situation particularly in haemodynamically unstable patients.
In conclusion, semi-invasive arterial pressure waveform analysis using the first generation FloTrac/VigileoTM device does not appear to adequately agree with the invasive pulmonary artery thermodilution technique for determination of CO in patients undergoing major cardiac surgery. Despite its ability to track changes of CO in individual patients with adequate accuracy, an overall percentage error of 45.9% compared with the bolus thermodilution technique does not recommend routine use as an alternative to invasive CO measurement at present.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Connors AF Jr, Speroff T, Dawson NV, et al. (1996) The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 276:889–97.[Abstract]
2 Sandham JD, Hull RD, Brant RF, et al. (2003) A randomized, controlled trial of the use of pulmonary-artery catheters in high-risk surgical patients. N Engl J Med 348:5–14.
3 Harvey S, Harrison DA, Singer M, et al. (2005) Assessment of the clinical effectiveness of pulmonary artery catheters in management of patients in intensive care (PAC-Man): a randomized controlled trial. Lancet 366:472–77.[CrossRef][ISI][Medline]
4 Baulig W, Bernhard EO, Bettex D, et al. (2005) Cardiac output measurement by pulse dye densitometry in cardiac surgery. Anaesthesia 60:968–73.[CrossRef][ISI][Medline]
5 Yamashita K, Nishiyama T, Yokoyama T, et al. (2005) Cardiac output by PulseCOTM is not interchangeable with thermodilution in patients undergoing OPCAB. Can J Anesth 52:530–4.
6 Rodig G, Prasser C, Keyl C, et al. (1999) Continuous cardiac output measurement: pulse contour analysis vs. thermodilution technique in cardiac surgical patients. Br J Anaesth 82:525–30.
7 Bland JM and Altman DG. (1986) Statistical method for assessing agreement between two methods of clinical measurements. Lancet 1:307–10.[CrossRef][ISI][Medline]
8 Critchley LAH and Critchley JAJH. (1999) A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput 15:85–91.[Medline]
9 Goedje O, Hoecke K, Goetz AE, et al. (2002) Reliability of a new algorithm for continuous cardiac output determination by pulse contour analysis during hemodynamic instability. Crit Care Med 30:52–8.[CrossRef][ISI][Medline]
10 Orme LE, Pigott DW, Mihm FG. (2004) Measurement of cardiac output by transpulmonary arterial thermodilution using long radial artery catheter. A comparison with intermittent pulmonary artery thermodilution. Anaesthesia 59:590–4.[CrossRef][ISI][Medline]
11 Hirschl MM, Binder M, Gwechenberger M, et al. (1997) Noninvasive assessment of cardiac output in critically ill patients by analysis of the finger blood pressure waveform. Crit Care Med 25:1909–14.[CrossRef][ISI][Medline]
12 Hirschl MM, Kittler H, Woisetschläger C, et al. (2000) Simultaneous comparison of thoracic bioimpedance and arterial pulse waveform-derived cardiac output with thermodilution measurement. Crit Care Med 28:1798–802.[CrossRef][ISI][Medline]
13 Mukkamala R, Reisner AT, Hojman HM, et al. (2006) Continuous cardiac output monitoring by peripheral blood pressure waveform analysis. IEEE Trans Biomed Eng 53:459–67.[CrossRef][ISI][Medline]
14 Costa MG, et al. (2006) Continuous arterial pulse cardiac output validation in hyperdynamic conditions. ISICEM 26th (Brussels, Belgium).
15 Sujatha P, Mehta P, Dhar A, et al. (2006) Comparison of cardiac output in OPCAB: Bolus thermodilution technique versus pulse contour analysis. Ann Card Anaesth 9:44–8.[Medline]
16 Wesseling R, de Witt B, Ty Smith N. (1983) A simple device for the continuous measurement of cardiac output. Adv Cardiovas Phys 5:16–52.
17 O'Rourke NF and Yaginuma T. (1984) Wave reflections and the arterial pulse. Arch Intern Med 144:366–71.[Abstract]
18 Schuerholz T, Cobas Meyer M, Friedrich L, et al. (2006) Reliability of continuous cardiac output determination by pulse-contour analysis on porcine septic shock. Acta Anaesthesiol Scand 50:407–13.[CrossRef][ISI][Medline]
19 Pinsky MR. (2001) A rose by any other name: cardiac output. Crit Care Med 29:2021–22.[CrossRef][ISI][Medline]
20 Morris AH, Chapman RH, Gardner RM. (1984) Frequency of technical problems encountered in the measurement of pulmonary artery wedge pressure. Crit Care Med 12:164–70.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. W. L. Scheeren, C. Wiesenack, F. D. Compton, B. Zukunft, C. Hoffmann, W. Zidek, and J.-H. Schaefer Performance of a minimally invasive cardiac output monitoring system (Flotrac/Vigileo) Br. J. Anaesth., August 1, 2008; 101(2): 279 - 280. [Full Text] [PDF]
|
|
|
|
 |
|
 B. Matthieu, N.-G. Karine, C. Vincent, V. Alain, C. J. Francois, R. Philippe, and S. Francois Cardiac Output Measurement in Patients Undergoing Liver Transplantation: Pulmonary Artery Catheter Versus Uncalibrated Arterial Pressure Waveform Analysis Anesth. Analg., May 1, 2008; 106(5): 1480 - 1486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. D. Compton, B. Zukunft, C. Hoffmann, W. Zidek, and J.-H. Schaefer Performance of a minimally invasive uncalibrated cardiac output monitoring system (FlotracTM/VigileoTM) in haemodynamically unstable patients Br. J. Anaesth., April 1, 2008; 100(4): 451 - 456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Davison and C. Junker Advances in Critical Care for the Nephrologist: Hemodynamic Monitoring and Volume Management Clin. J. Am. Soc. Nephrol., March 1, 2008; 3(2): 554 - 561. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Mayer, J. Boldt, M. W. Wolf, J. Lang, and S. Suttner Cardiac Output Derived from Arterial Pressure Waveform Analysis in Patients Undergoing Cardiac Surgery: Validity of a Second Generation Device Anesth. Analg., March 1, 2008; 106(3): 867 - 872. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Staier, C. Wiesenack, L. Gunkel, and C. Keyl Cardiac output determination by thermodilution and arterial pulse waveform analysis in patients undergoing aortic valve replacement: [Determination du debit cardiaque par thermodilution et par analyse du contour de pression arterielle chez des patients subissant un remplacement de la valve aortique] Can J Anesth, January 1, 2008; 55(1): 22 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Button, L. Weibel, O. Reuthebuch, M. Genoni, A. Zollinger, and C. K. Hofer Clinical evaluation of the FloTrac/VigileoTM system and two established continuous cardiac output monitoring devices in patients undergoing cardiac surgery Br. J. Anaesth., September 1, 2007; 99(3): 329 - 336. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Guarracino, M. Stefani, F. Lapolla, C. Cariello, L. Doroni, A. Danella, R. Baldassarri, J. Mayer, J. Boldt, T. Schollhorn, et al. Monitoring cardiac output with Flo Trac VigileoTM Br. J. Anaesth., July 1, 2007; 99(1): 142 - 143. [Full Text] [PDF]
|
|
E-letters:
Read all E-letters
- Monitoring cardiac output with Flo Trac VigileoTM
- Fabio Guarracino, et al.
- British Journal of Anaesthesia, 13 Mar 2007 [Full text]
- Author reply: Monitoring cardiac output with Flo Trac VigileoTM
- Jochen Mayer, et al.
- British Journal of Anaesthesia, 30 Mar 2007 [Full text]
- Arterial pressure wave form and cardiac output in dilated cardiomyopathy
- Usha Kiran, et al.
- British Journal of Anaesthesia, 6 Jun 2007 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||