INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

RAT MODELS have been used extensively in cardiovascular research and invasive hemodynamic techniques have been employed most frequently (4, 10, 11, 18). However, these invasive techniques mandate the loss of the animal after the procedure, thereby preventing long-term longitudinal follow-up in a single rat. Therefore, newer methods permitting repeated determinations of hemodynamic variables have been actively pursued. In addition to techniques involving chronic implantation of sensing devices, echocardiographic and Doppler advances have provided valid means for studying cardiovascular hemodynamics in small animals (1, 5, 6, 9, 12, 13). Echocardiographic and Doppler techniques were validated clinically many years ago (3).

The objective of this study was to validate the echocardiographic measurement of cardiac output in rats by comparing it to the standard thermodilution method. To this end 10 different echocardiographic (transthoracic approach) means of measurement were employed to determine cardiac output. Furthermore, changes in intravascular volume were also produced so that the effect of volumic variations on echocardiographically determined structural (geometric) properties of the left ventricle and the aortic, pulmonary, and mitral orifices could also be assessed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eleven male Wistar rats (384 ± 20 g) were obtained from Charles River Breeding Laboratories (Wilmington, MA) for this study. They were housed in temperature- and humidity-controlled rooms and were permitted free access to standard rat chow (PMI Nutrition International; St. Louis, MO) and tap water. The Institutional Animal Care and Use Committee approved the experimental protocol. Studies were performed with rats under pentobarbital anesthesia (50 mg/kg ip). A jugular vein was cannulated, and the polyethylene catheter (PE-50) was advanced into the right atrium. A thermistor probe (Physitemp Instruments; Clifton, NJ), connected to a thermodilution-recording device (Cardiotherm 500, Columbus Instruments; Columbus, OH), was placed into the ascending aorta through the right internal carotid artery. To determine cardiac output, 150 µl of 1% saline solution at room temperature was injected into the venous catheter (14). Arterial pressure was recorded using a catheter placed into the descending aorta through the left femoral artery.

Transthoracic echocardiographic (TTE) determinations were performed in the left lateral decubitus position using a commercially available echocardiographic system (Sonos 4500 with an 6- to 12-MHz transducer, Agilent Technologies). A preliminary study in five rats demonstrated that cardiac output could not be measured at the level of the tricuspid valve or annulus, mitral annulus, pulmonary valve, or using the measurement of left ventricular volume with the Simpson method. This was due to a very poor precision of the echocardiographic measurements with a large difference between two different measurements. Thus cardiac output was measured at the level of aortic annulus, aortic valve, thoracic aorta, pulmonary annulus, mitral valve, and by the measurements of systolic and diastolic left ventricular diameter. With the use of parasternal positioning, two-dimensional and M-mode guided images of the long axis of the left ventricle were obtained. With the use of the M-mode image, left ventricular diastolic (LVDD) and systolic diameters (LVSD) were measured using American Society of Echocardiography (ASE) guidelines. Aortic annulus diameter (D_AoA) and the distance between the valves (D_Aov) were measured during systole from the two-dimensional images (Fig. 1). The largest diastolic distance between the mitral leaflets (MD) was also measured using two-dimensional images (Fig. 2). Maximal diameter of pulmonary annulus (PD), and velocity time integral of pulsed and continuous wave Doppler of the pulmonary flow (VTI_pp and VTI_pc, respectively) were obtained from the short-axis view. From an apical five-chamber view, aortic flow (at the annulus level) was recorded using pulsed and continuous Doppler. Mitral flow was recorded using pulsed Doppler at the level of the tip of the mitral valves. Velocity time integrals were measured for all three flows (VTI_Aop, VTI_Aoc, VTI_m) (Fig. 2). Horizontal thoracic aortic diameter (D_Ao) was measured from a suprasternal view (average of systolic and diastolic diameter), and ascending aorta flow was recorded using pulsed and continuous wave Doppler thus permitting measurement of the velocity time integrals (VTI_thap, VTI_thac). With the use of these measurements, stroke volume could be calculated by using the following formula: (diameter)² × 3.14 × velocity time integral (VTI)/4 at the level of mitral valve (MD and VTI_m), aortic annulus (D_AoA, VTI_Aop, or VTI_Aoc), aortic valve (D_Aov, VTI_Aop, or VTI_Aoc), thoracic aorta (D_Ao, VTI_thap, or VTI_thac), or pulmonary annulus (PD, VTI_pp, or VTI_pc). Stroke volume was also calculated using measurements of left ventricular systolic and diastolic diameter (Fig. 3): LVDD³  LVSD³. To obtain cardiac output, stroke volume was multiplied by heart rate.

View larger version (78K): [in this window] [in a new window]   Fig. 1.   Transthoracic echocardiographic (TTE) bidimensional long-axis view of the left ventricle (LV). A: measurement of (left double arrow) aortic annulus diameter and (right double arrow) systolic distance between aortic valves. B: measurement of diastolic (D) and systolic (S) thoracic aortic (Ao) diameter on M-mode image. LA, left atrium.

View larger version (60K): [in this window] [in a new window]   Fig. 2.   A: TTE bidimensional long-axis view of the LV. Double arrow, measurement of diastolic distance between the mitral leaflets. B: TTE bidimensional short-axis view at the pulmonary artery (PA) level. Measurement of pulmonary annulus.

View larger version (120K): [in this window] [in a new window]   Fig. 3.   Left ventricular M-mode image. Measurement of left ventricular diastolic (LVDD, left double arrow) and systolic (LVSD, right double arrow) diameters.

Experimental protocol. Cardiac output was measured simultaneously using Doppler or echocardiographic techniques and thermodilution at different levels of volemia. For this, diameter and VTI at the level of aortic annulus were recorded and immediately followed by the injection of saline solution for thermodilution measurement of cardiac output. We performed the same method at thoracic aortic, pulmonary, and mitral levels and for the determination of cardiac output from the left ventricular diameters. After basal measurements were obtained, they were repeated after blood withdrawal and then again after transfusion with blood obtained from a donor rat. A 15-min stabilization period was allowed for stabilization after each blood withdrawal or transfusion before cardiac output measurement was repeated.

Statistical analysis. Bias and precision were calculated as described by Bland and Altman (2). Simultaneous thermodilution and echocardiographic measurements of cardiac output provided a mean for obtaining correlation coefficients and linear regression analysis. Correlation between blood volume and echocardiographic and Doppler measurement changes were also obtained. Intraobserver reproducibility was assessed in seven rats as the mean percent error (absolute difference divided by average of the two observations).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cardiac output measurements were performed in 11 rats at 43 different levels of intravascular volume (average 2.5 ± 1.2 for each rat). Basal systolic and diastolic pressure, heart rate, LVDD, and LVSD were 107 ± 3 mmHg, 68 ± 1 mmHg, 349 ± 10 beats/min, 7.70 ± 0.17 mm, and 4.23 ± 0.01 mm, respectively. Other basal measurements are detailed in Table 1.

At baseline, the echogenicity of the rats was excellent, and all the echocardiographic indexes were obtained. Echogenicity diminished during hypovolemia, and we failed to obtain thoracic aortic diameter and M-mode left ventricular measurements on three occasions (17%). During blood transfusion, echogenicity increased, and all measurements were accomplished. Aortic annulus and thoracic aortic diameters were always measured with a good reproducibility, but mitral and particularly pulmonary diameters were measured without certainty. Aortic, pulmonary, and mitral flows were recorded all of the time. Positive correlation was found between rat body weight and aortic annulus (r = 0.80; P < 0.0001), pulmonary annulus (r = 0.38; P < 0.01), diastolic mitral valve diameter (r = 0.38; P < 0002), and left ventricular end-diastolic diameter (r = 0.34; P < 0.03); however, this was not significant with thoracic aortic or aortic valve diameters.

Excellent correlation was demonstrated between thermodilution and all echocardiographic measurements of cardiac output (Table 2). The best correlation (r = 0.93; P < 0.0001), smallest bias (3 ml/min), and best precision (16 ml/min) were obtained at the aortic annulus level using pulsed Doppler (Fig. 4). Good correlations and precision and small bias were obtained at the level of aortic valve using either pulsed or continuous wave Doppler as well as at aortic annulus level using continuous wave Doppler (Fig. 5).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Bland and Altman (2) representation of the comparison between thermodilution cardiac output and aortic cardiac output at the level of aortic annulus using pulsed (A) or continuous Doppler (B).

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Bland and Altman (2) representation of the comparison between thermodilution cardiac output and aortic cardiac output at the level of aortic valves using pulsed (A) or continuous Doppler (B). 139 and 116 ml/min are missing in A.

Correlation between Doppler and thermodilution measurements of cardiac output was good at the thoracic aorta level and bias close to zero; however, precision of the measurements was not good, reaching as high as 69 ml/min (Fig. 6). Overestimation of the cardiac output was demonstrated at the pulmonary annulus level, and this was greater as cardiac output increased (Fig. 7). A fair correlation was obtained between the measurements of cardiac output at the mitral valve level (Fig. 8A), but precision was low (50 ml/min). Basal cardiac output measurements using left ventricular systolic and diastolic diameters (Teichholz formula) were closely correlated (r = 0.84; P < 0.0002) to the thermodilution measurements with least bias and good precision. However, when all measurements were included, the correlation coefficient was r = 0.61 (P < 0.0001), and a large overestimation and a poor precision were demonstrated (Table 2, Fig. 8B).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Bland and Altman (2) representation of the comparison between thermodilution cardiac output and aortic cardiac output at the level of thoracic aorta using pulsed (A) or continuous Doppler (B).

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Bland and Altman (2) representation of the comparison between thermodilution cardiac output and pulmonary cardiac output at the level of pulmonary annulus using pulsed (A) or continuous Doppler (B).

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Bland and Altman (2) representation of the comparison between thermodilution cardiac output and cardiac output (A) at the level of mitral valve using pulsed Doppler or (B) calculated from left ventricular systolic and diastolic diameters. In B, Y-axis scale is different.

Reproducibility of the measurement of aortic annulus (4.1 ± 3%) and valve (5.5 ± 4.7%), mitral valve (6.2 ± 5.8%), left ventricular diastolic (7.6 ± 5.8%) diameters, and aortic (6.7 ± 0.3% pulsed Doppler, 7.1 ± 0.3% continuous Doppler), mitral (8.4 ± 4.5%), and pulmonary (9.4 ± 8.1% pulsed Doppler, 7.5 ± 8.1% continuous Doppler) VTI was excellent. In contrast, we obtained a poor reproducibility for the measurements of left ventricular systolic (14 ± 11%), pulmonary (14 ± 10%), and thoracic aortic (17 ± 17%) diameters and thoracic aortic VTI (22 ± 14% pulsed Doppler, 14 ± 10% continuous Doppler).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study is the first to systematically evaluate different transthoracic echocardiographic methods to determine cardiac output in rats. The results provide validation for the longitudinal measurement of cardiac output using echocardiographic techniques with high correlation with the thermodilution technique obtained at the aortic annulus level with the use of pulsed Doppler. This method was also shown to be best for echocardiographic cardiac output determination clinically (3). Recently, Bjornerheim et al. (1) measured cardiac output (at the aortic level) in 14 rats using echocardiography and the Doppler technique, providing a good correlation (r = 0.85) with the ultrasound transit time technique without bias and with good precision (20 ml/min). These authors measured the diameter at the level of left ventricular outflow tract, and no correlation was demonstrated between animal body weight and this diameter. In this study, we measured the aortic annulus diameter, which demonstrated a good relationship with body weight. With volume variations, aortic diameter remained constant, and this was in accordance with older mechanical studies of aortic annulus (15, 17).

The thoracic aortic diameter has been analyzed in rats to determine aortic compliance and the elastic modulus (7, 16). In these studies, the diastolic thoracic aortic diameter was measured at the ascending aortic level and was slightly smaller than in our report in which the aortic diameter was derived from the average of systolic and diastolic diameters at the horizontal thoracic aorta (7). No relationship was found between this diameter and the animal body weight.

Cardiac output at the mitral valve level was measured assuming that the mitral diameter was constant during diastole and that its shape was circular. Unfortunately, these assumptions were false, demonstrating a significant variation of the mitral valve diameter throughout the diastole, and the mitral valve shape was found to be ellipsoid in its short axis. This observation was demonstrated previously in patients (3). Our findings are, therefore, in accordance with clinical studies in which cardiac output measured at the mitral valve or annulus level demonstrated poor correlation with the thermodilution method (3).

As previously reported (1) with rats as well as patients, measurement at the pulmonary annulus level is inaccurate because the pulmonary orifice is positioned parallel to the ultrasound beam. Therefore, to measure the pulmonary annulus, we must use the lateral resolution of the echocardiographic image, far less precise than the axial one. Before any volume changes were produced, M-mode determination of cardiac output was accurate. Similarly, Nakamura et al. (8) reported a good correlation between stroke volume measured simultaneously using M-mode echocardiography and pulsed Doppler flowmeter placed around the ascending aorta. However, when all measurements were included in our study, M-mode measurement overestimated cardiac output. Calculation of left ventricular volumes assumes that the length of the long axis is two times greater than the short axis of the left ventricle. During blood transfusion the short axis may increase more than the long axis, and this could explain, at least in part, the overestimation of M-mode measurement.

Left ventricular, thoracic aorta, pulmonary annulus, mitral valve diameters, and aortic velocity time integral were found to be modified by intravascular volemic changes. Therefore, stroke volume changes were mainly due to flow modifications at the aortic annulus level and to cross-sectional changes of the orifice at the level of thoracic aorta, pulmonary annulus, and mitral valve.

Clearly, this study is not devoid of limitations. The main limitation of this method is the determination of the annulus diameter. Because this diameter is squared for the stroke volume calculation, even a small error in this measurement can lead to an important variation of the stroke volume. In clinical studies the standard deviation of differences between repeated measurements of cardiac output at the level of aortic annulus is ~5-8%, which corresponds to a precision of the measurement of aortic annulus above 0.5 mm. With the use of a 2.5-MHz transducer, this is close to the theoretical axial resolution (half of wavelength) of 0.3 mm. In rats the annulus diameter is 2.7 mm, but with the use of 12 MHz, the theoretical axial resolution is 0.06 mm, which is exactly the precision that we found (4.1%, which corresponds to 0.06 mm). Because the duration of the experiment was long, we decided to modify preload but not contractile function. Therefore, we wonder whether our method could be used in studies in which contractile function is modified. Because only one researcher (M. Slama) performed all the measurements, we did not analyze the interobserver variability or reproducibility. Even if the echocardiographic and Doppler technique is easy to perform, it is necessary to learn the technique for few weeks before being accurate. Another limitation is that systolic variations of aortic and pulmonary annulus were not assessed. Furthermore, aortic, mitral, and pulmonary diameters (measured echocardiographically) were not compared with the respective diameters calculated from stroke volume determined by thermodilution and velocity time integrals of the flow at the same level. This comparison could not be done because we did not compare the recorded velocity time integral to a flow measurement using standard methods. Therefore, considering these limitations, our conclusions concerning the diameter changes of those valvular orifices during volumic changes should be taken with caution.

In conclusion, cardiac output can be accurately assessed using transthoracic echocardiography and pulsed Doppler at the level of aortic annulus in the rat. This annulus remained stable, but pulmonary annulus and mitral valve diameters were found to be modified during volumic changes as well as left ventricular diastolic diameter.