INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

BECAUSE CARDIAC OUTPUT is an important parameter for estimation of the functional condition of the circulatory system, an opportunity is often desirable for its measurement in comparison with blood pressure and heart rate. The most favored method for this purpose is the thermodilution method, which can be repeated as often as necessary because the indicator totally disappears within a few seconds. In addition, this technique is easy to perform and does not require expensive equipment. However, a major problem with the thermodilution method is the loss of indicator during its transit from the venous side through the injection catheter (12) to the arterial detection side. Specifically, the larger the ratio of surface area to passing blood volume (which is inversely proportional to vessel diameter) and the longer the path between injection and detection sides, the greater the temperature exchange across vessel walls (11). In rabbits or larger species, the loss of indicator seems to be negligible, because simultaneous recordings of thermodilution curves in the pulmonary artery and the aorta after injection of the indicator near the right atrium resulted in similar values of cardiac output (4, 14). In contrast, in small animals, e.g., rats, measurements of cardiac output using the thermodilution method are highly dependent on the position of the thermistor probe and the injection site within the vascular system (11).

No loss of indicator occurs with the use of nondiffusive indicators, e.g., dyes, for determination of cardiac output. In contrast to the thermodilution method, the continuous measurement of intravascular dye concentrations requires expensive equipment and fiber-optical probes that cannot be produced in the necessary small size. The smallest available probes are suitable for rats (10). Although dilution curves of nondiffusive indicators can also be determined from timed blood samples withdrawn from a free-flowing arterial catheter (3, 26, 30), this procedure results in a loss of blood volume. Thus, especially in small animals such as mice, this will limit the number of measurements.

Similarly, the number of measurements of cardiac output is also limited with the use of the microsphere technique. Cardiac output determinations with this technique require the injection of microspheres (diameter ~15 µm) into the left ventricle simultaneously with the sampling of arterial blood for a defined time period with a constant flow rate. The cardiac output is then calculated from the amount of microspheres contained in the blood sample. All other injected microspheres become lodged in small arterioles. This leads to an increase in the peripheral flow resistance, which results in a breakdown of the circulation after a few measurements (13).

Determination of cardiac output in dogs from recordings of blood conductivity after intravenous bolus injection of hypertonic saline has been reported previously (6, 7, 23). The present study aimed to adapt this technique for use in small laboratory animals. After bolus injection of 5% glucose solution into the femoral veins of mice and rats, blood conductivity was measured between an electrode placed inside the aorta and a rectal reference electrode. In rats, the reliability of this method was tested by recording conductivity-dilution curves simultaneously with the time course of plasma glucose concentration. The values of cardiac output determined with the conductivity-dilution method were then compared with cardiac output determinations based on the plasma glucose concentration.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animal preparation and general proceedings. Experiments were performed on Sprague-Dawley rats (250-300 g) and BALB/c mice (14-28 g) in accordance with institutional guidelines. All animals were anesthetized with a gas mixture containing 1-1.5% halothane, 70% N₂O, and the remainder O₂. Body temperature was maintained at 37-37.6°C using a temperature-controlled heating pad. The right femoral artery was isolated to allow placement of a Teflon-coated platinum wire (World Precision Instruments, Berlin, Germany) inside the aorta. Previously, the insulation at the tip of the wire was removed at a length of 100-200 µm to achieve contact with the blood. The platinum at the tips of the electrodes was then rounded off and polished to prevent coagulation of blood (Fig. 1). After the intra-aortic electrodes were prepared in this manner, they were connected with the use of a soldering point to an extension lead (copper) and plugged into the preamplifier of the apparatus used for measurement of blood conductivity. The resistance of the electrodes was <3 . Inside the rectum, a silver-coated copper wire 2 mm in diameter and 25 mm in length in mice and 60 mm in length in rats was placed as a reference electrode. The apparatus for measuring blood conductivity was assembled by the electrical engineering laboratory of the University of Heidelberg (Heidelberg, Germany). According to Worley et al. (31), who show a comparable circuit diagram, a constant alternating current of 300 µA was applied at 10 kHz between the electrode inside the aorta and that inside the rectum. These parameters were identical for rats and mice. The left femoral vein was cannulated for the application of 5% glucose solution. The left femoral artery was cannulated for 1) monitoring of arterial blood pressure, 2) sampling of arterial blood to determine arterial blood gases (AVL 990; AVL, Bad Homburg, Germany) and plasma glucose concentrations (Beckman Glukose-Analysator 2; Beckman Instruments, Munich, Germany), and 3) in rats, determining the time course of arterial concentrations of glucose from timed droplets collected from the free-flowing femoral arterial catheter (5-7 droplets/s). All droplets passed a light barrier, and a signal was recorded and dropped on an aluminum foil-covered board that was moved below the end of the catheter. This procedure resulted in separate placing of the droplets on the foil. Each droplet was taken up into a hematocrit tube, and the plasma glucose concentration of each droplet was measured using the Beckman Glukose-Analysator 2 in a 5-µl plasma sample after centrifugation.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Design and dimensions of intra-aortic conductivity probes. A: rat electrode. B: mouse electrode.

Control experiments. Calculation of cardiac output using time-conductivity curves is influenced by the relation of changes in the output voltage of the instrument on a percentage basis to changes in the conductivity on a percentage basis. If, for example, blood conductivity is decreased by 10% and, as a result, the output voltage of the instrument is decreased by >10%, then cardiac output will be underestimated. This relationship between output voltage of the instrument and blood conductivity was measured in vitro using the rat as well as the mouse electrode. While the conductivity was measured, incremental amounts of 5% glucose solution were added to 20 ml of rat blood, which was heparinized, warmed to 37°C, and stirred continuously. The slope value s of the linear regression lines between the decrease of the output voltage on a percentage basis and the calculated decrease of blood conductivity on a percentage basis was used to correct the size of the integral below the conductivity-dilution curve. Furthermore, it is essential for the method presented here that s is independent of the distance between the intra-aortic and rectal reference electrodes. To verify this, the output voltage of the instrument was measured continuously while 10 ml of distilled water were added 10 times to 200 ml of 0.9% saline, which was stirred continuously. This procedure was repeated several times at different distances between the electrodes. The slope value s of the linear regression lines between the decrease of the output voltage on a percentage basis and the calculated decrease of the conductivity on a percentage basis was correlated to the distance between both electrodes.

Data analysis. The integral below the conductivity-dilution curves was determined using an image analyzing system (MCID; Imaging Research, St. Catherines, Ontario, Canada). For calculation of cardiac output, the mean of the integrals of the outer and inner shapes of the dilution curves was taken. Because recirculation effects are more relevant in small animals (11), the integrals of the thermodilution and conductivity-dilution curves were corrected according to Schorer (21). The downslopes of the glucose-dilution curves were plotted in a semilogarithmic manner, and the recirculation was eliminated by linear extrapolation of the first linear part of the downslope to the zero line. Formulas were used for the calculation of cardiac output. The formula for measurement of the time course of plasma glucose was

(1)

where CO is cardiac output (ml/min); V_i is injectate volume (ml); C_i and C_p are the concentrations of glucose (mg/ml) in the injectate or plasma, respectively; and Hct is hematocrit. From each value of C_p over time [C_p(t)], the baseline glucose concentration measured during 1 s before the appearance of the glucose-dilution curve was subtracted. Because glucose is distributed during the time period between injection and detection only within plasma (22), a term was added regarding the Hct to correct for blood cell volume.

(2)

Statistics. Differences in cardiac output between awake and halothane-anesthetized mice or rats and between hypervolemia and different degrees of hypovolemia were tested by an analysis of variance and Student's t-test with Bonferroni correction. Slopes of regression lines were tested for their differences from zero. The levels of significance are given in the legends of Figs. 4-6.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Table 1 shows the physiological variables measured in anesthetized mice before bleeding, after withdrawal, after reinfusion of 0.6 ml blood, and after infusion of 0.6 ml saline. After infusion of saline, a mild combined acidosis was found.

The effects of halothane anesthesia on cardiac index (cardiac output/kg body wt) are summarized in Fig. 2. In both species the cardiac index was ~12% lower during halothane anesthesia than in awake animals. However, these differences were not significant.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Cardiac index (cardiac output/kg body wt) measured in conscious (solid bars) and halothane-anesthetized (open bars) rats and mice. Values are means ± SE of 6 rats or mice, respectively. Cardiac index was measured at least 6 times in each animal. In both species, cardiac index was slightly but not significantly lower during halothane anesthesia.

In six mice cardiac output was measured during hypovolemia and hypervolemia. Figure 3 shows a typical experiment. After blood was withdrawn, cardiac output was reduced and tended to increase spontaneously with time in parallel with blood pressure. Reinfusion of the first half of the shed blood immediately normalized cardiac output and blood pressure. Infusion of the second half of the withdrawn blood as well as the infusion of 0.6 ml of saline did not significantly affect the cardiac index (Fig. 4). During final bleeding, cardiac index was gradually reduced, with a highly significant correlation to the withdrawn blood volume (Fig. 5).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Recording of blood pressure and heart rate in parallel with cardiac output measurements during hypovolemia and hypervolemia in a mouse (body wt 25 g). Top: arterial blood pressure (thick track) and heart rate (thin line). Arrowheads indicate measurement of arterial acid-base status (cf. Table 1). Middle: experimental setup for changes in blood volume. Bottom: cardiac output. Withdrawal of blood resulted in a drop in cardiac output, followed by a spontaneous increase parallel with the recovery of blood pressure. Reinfusion of blood restored cardiac output and blood pressure to initial values, whereas infusion of saline had no effect.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Changes in cardiac index in 6 mice during bleeding (), reinfusion of shed blood (shaded squares), and infusion of saline (). Values are means ± SE of 3-4 determinations of cardiac index in 6 mice each. * P < 0.05: withdrawal of 0.3 ml blood resulted in a significant drop in cardiac index. The drop in cardiac index after removal of an additional 0.3 ml blood was not significant. ** P < 0.01: reinfusion of 0.3 ml of shed blood immediately normalized cardiac index. Further infusions of 0.3 ml blood or 0.6 ml saline (in 2 portions of 0.3 ml each) had no significant effects.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Dependency of cardiac index in mice on withdrawn blood volume. Values are means ± SE of 3-4 measurements of cardiac index in n mice for each respective value. Cardiac index is gradually decreased with increasing loss of blood volume (P < 0.001).

The reliability of cardiac output measurements using recordings of blood conductivity after bolus injection of 5% glucose solution was tested in rats. Cardiac output was calculated from simultaneously recorded conductivity-dilution curves and time courses of plasma glucose concentration. The correlation between the values of cardiac output determined simultaneously with these two methods is shown in Fig. 6. The highly significant regression line (P < 0.001) of Fig. 6 correlates with the line of identity.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Correlation between values of cardiac output calculated from simultaneously recorded conductivity-dilution curves (CD cardiac output) and time courses of plasma glucose concentration (GC cardiac output) after bolus injection of a 5% glucose solution in rats. Regression line (solid line) is highly significant and correlates with line of identity (dashed line) (P < 0.001).

In addition, I tested whether the slope of the regression line of the correlation between output voltage of the instrument and conductivity depends on the distance between the intra-aortic and rectal reference electrodes. Figure 7 demonstrates that no changes in the slopes of output voltage versus conductivity plots occurred when the distance between the electrodes was varied.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Slopes of regression lines of output voltage vs. conductivity plots (both parameters in percentages of initial values) correlated to distance between conductivity probe and reference electrode. Slopes are independent of distance between electrodes.

The effect of frequent injections of 5% glucose solution on plasma glucose concentration in rats and mice is demonstrated in Fig. 8. In rats, a time lag of 1 min between injections of 0.2 ml 5% glucose solution produced an increase of plasma glucose level up to the tenth injection. Thereafter, the plasma glucose level did not change any more despite further injections (Fig. 8A). This increase of the plasma glucose level was lower when the time lag between the injections was prolonged to 2 min and disappeared when it was 4 min (Fig. 8, C and E). Ten minutes after the last injection, the plasma glucose concentration was comparable in all rats to that at the beginning of the injection series. In mice, a time lag of 1 min between injections of 0.02 ml 5% glucose solution resulted in an increase of the plasma glucose level without reaching a steady state (Fig. 8B). Even 40 min after the last injection, the glucose value was not restored to that measured at the beginning of the experiment. Similar to that in rats, this increase of the plasma glucose level was lower when a time lag of 2 min was allowed between the injections. Twenty minutes after the last injection, the plasma glucose level was comparable to that at the beginning of the experiment (Fig. 8D). The plasma glucose concentration did not change during the whole experiment and during the 40-min observation period after the last injection, when the time lag between the injections was prolonged to 4 min (Fig. 8F). These re- sults demonstrate that measurements of cardiac out- put with the conductivity-dilution technique can be per- formed in rats as well as in mice every fourth minute without changes in the plasma glucose level. It is important to note that the hematocrit did not change in either species when the time lag between the injections of the 5% glucose solution was 4 min (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Effect of frequent intravenous injections (arrows) of 0.2 ml (rats) or 0.02 ml (mice) 5% glucose solution on plasma glucose concentration of rats and mice. A-F: time course of plasma glucose concentration in 1 rat or 1 mouse. A and B: time lag between each injection is 1 min. In the first rat (285 g; A), plasma glucose increased during the first 10 min and stayed constant thereafter. Ten minutes after the last injection, plasma glucose concentration was comparable to that at beginning of experiment. In the first mouse (29.2 g; B), plasma glucose increased continuously during whole experiment. Even 40 min after the last injection, plasma glucose level was not restored to that at beginning of experiment. C and D: time lag between each injection is 2 min. Plasma glucose increased slightly in a second rat (280 g; C) as well as in a second mouse (28.5 g; D) during the first 10 min and stayed constant thereafter. Ten (rat) or twenty (mouse) minutes after the last injection, plasma glucose concentration was comparable to that at beginning of experiment. E and F: time lag between each injection is 4 min. Neither in this rat (290 g; E) nor this mouse (26.2 g; F) did plasma glucose concentration change during experiment or observation period after the last injection.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Cardiac output can be quantified frequently in mice and rats with recordings of blood conductivity after bolus injections of 5% glucose solution. Compared with conscious animals, halothane-anesthetized animals of both species had lower values of cardiac index. However, these differences were not significant. In mice, cardiac index was gradually decreased during hypovolemia (58 ml/kg body wt), whereas hypervolemia (23 ml/kg body wt) had no significant effect. In rats, simultaneously obtained conductivity-dilution curves and time courses of plasma glucose concentration yielded closely corresponding values of cardiac output.

Until now, no methods were available which enabled easy, frequent, and, at the same time, reliable measurements of cardiac output in mice. The most favored method for determinations of cardiac output, the thermodilution method, appears to be susceptible to incorrect measurements in small animals. In rats, Hayes at al. (11) demonstrated that values of cardiac output determined with the thermodilution method in the pulmonary artery were about 85% higher than those measured simultaneously in the thoracic aorta. This observation was ascribed to heat loss caused by the large ratio of surface area to passing blood volume, which becomes greater as vessel diameter gets smaller. In addition, the temperature exchange through vessel walls can be expected to be proportionately greater with decreasing flow rate at low-output states (16). Thus, particularly in mice, the thermodilution method appears to be unsuitable. Radioisotope and dye dilution methods or the injection of microspheres (diameter ~15 µm) into the left ventricle while sampling the arterial blood at a constant flow rate for a defined time period avoids the problem of indicator loss. In mice, all these methods require the sampling of relatively large amounts of arterial blood compared with the total blood volume, which may limit the number of measurements in these animals. An additional problem arises when the microsphere technique is used for measurements of cardiac output. Microspheres occlude vessels, which leads to a breakdown of the circulation after about five measurements of cardiac output (13). Besides, all these methods do not enable on-line computation of cardiac output because of the time-consuming analysis of the blood samples in contrast to intravascular measurements of temperature changes or dye concentrations (10, 21). In principle, recordings of blood conductivity can also be used for on-line computation of cardiac output.

Measurements of cardiac output using the time course of blood conductivity have been performed previously in dogs with the use of hypertonic saline as an indicator and a calibrated flow-through conductivity cell (16), an extra-arterial conductivity cell (23), or a tetrapolar pulmonary catheter (7). The large probes used in these studies are not suitable for small laboratory animals such as rats or mice because of size considerations. In contrast, the probes used in the present study can be produced in almost any size, which enables measurements of cardiac output with a simple, reliable indicator-dilution method in very small animals.

However, inserting a probe into the aorta of rats or mice may influence cardiac output due to the partial obstruction of the lumen. The rat electrode used in the present study had an outer diameter of 0.6 mm, whereas that of the mouse electrode was 0.2 mm (Fig. 1). With regard to the inner diameter of the aortas of both species [rats, 2.8 mm (1); mice, 1.2 mm (9)], the electrodes have reduced the cross-sectional area of the aorta in rats by <5% and in mice by <3%. It appears unlikely that this extent of obstruction can influence cardiac output because, for arterial stenoses <40%, no significantly decreased volume flow or increased pressure drop can be measured (5, 32).

The reliability of the conductivity-dilution method was tested in rats and complemented with in vitro control experiments. In rats, simultaneously recorded conductivity-dilution curves and time courses of plasma glucose concentration yielded clearly corresponding values of cardiac output (Fig. 6). To the knowledge of the author, measurements of cardiac output based on time courses of plasma glucose concentration determined from arterial microsamples are not described in literature. Despite this fact, a 5% glucose solution appears to be suitable for cardiac output measurements. An indicator that is appropriate for cardiac output measurements based on the Fick principle should be nontoxic and enable the exact determination of its blood concentration. These requirements are met undoubtedly by glucose. In addition, the indicator should not quantitatively leave the plasma during the time period between intravenous injection and measurement of its plasma concentration. This also holds true for glucose (22). Microsampling of arterial blood for the determination of cardiac output has been performed previously in rats as well as in mice (3, 15, 26). In these studies the indicators were either radioisotopes (3, 15) or Evans blue (26). However, these indicators are not suitable for measurements of cardiac output in parallel with the conductivity-dilution method. Radioisotopes and Evans blue are salts, and in water they are dissociated into ions, which results in a conductivity of the injectate that is not zero. It follows from Ohm's law that, in solutions of salts, the conductivity depends on the distance between the electrodes. In the present study, the distance between the electrodes was undefined and variable. On the other hand, the change in the conductivity of a test solution on a percentage basis is always the same when the added volume of an indicator solution that has no conductivity (e.g., 5% glucose solution) is in a constant ratio to the volume of the test solution. This is independent of the initial value of conductivity of the test solution. When 5% glucose solution is used as an indicator, the change in the output voltage of the instrument on a percentage basis is therefore always in the same proportion to changes in blood conductivity on a percentage basis. This relationship holds true as long as this proportion is not affected by the distance between the electrodes. This was verified in vitro (Fig. 7). Cardiac output can then be calculated using the value that describes the exact relationship between changes in the output voltage and changes in blood conductivity. This value is the slope of the output voltage versus blood conductivity plots. For the final computation of cardiac output, the slopes of output voltage versus conductivity plots were determined in vitro in rat blood for each electrode used.

Cardiac index was measured using the conductivity-dilution method in conscious as well as halothane-anesthetized rats and mice. During halothane anesthesia, cardiac index in both species was slightly reduced, which is in accordance with findings of other authors (27). In the present study, these differences did not reach a level of statistical significance. The absolute values of the cardiac index measured during these experiments are for both species within the range of the values found in literature. Table 2 shows a number of cardiac index values from control groups of several different studies measured in mice and rats with various methods. Compared with other methods that are not complicated by indicator loss, such as labeled microspheres or albumin, the conductivity-dilution method appears to yield higher values of cardiac index. It could be argued that one reason for these differences might be a loss of indicator as the bolus passes through the capillaries of the lung circulation. Because the 5% glucose solution is isotonic, the injected bolus cannot leave the plasma along osmotic gradients as long as glucose is not taken up quantitatively into cells. Concerning the different cells with which the bolus comes into contact, only the contact time with erythrocytes that travel together with the bolus between the injection and detection sides for ~3-4 s appears to be long enough to account for a significant uptake of the injected glucose. However, it could be shown that, at normal hematocrit values, the plasma glucose concentration decreases only ~5-6% per hour as a consequence of glucose uptake into erythrocytes (22). Therefore, it appears unlikely that an indicator loss between the injection and detection side could account for the higher mean values of cardiac index measured in the present study compared with other studies in which nondiffusive indicators have been used (cf. Table 2). On the other hand, the comparison of the present data obtained, for example, from conscious animals with other studies in which conscious rats or mice were investigated is limited by the fact that the mean body weight varies between the different studies. Vizek and Albrecht (26) demonstrated that for rats the cardiac index dropped ~25% between the 50th and 60th postnatal day, remained stable up to the 90th day, and then fell again gradually ~25-30% up to the 180th day and an additional 10% up to the 400th day. These changes in cardiac index were paralleled by an increase in body weight of the rats from ~200 to 350 g between the 50th and the 180th postnatal day and to ~500 g up to the 400th day. The data presented in Table 2 appear to be in accordance with these findings of Vizek and Albrecht (26). For mice, no comparable study could be found in literature. However, it can be assumed that on this point no major differences exist between rats and mice. Thus the higher mean values of cardiac index of the present study compared with those measured in the studies cited in Table 2 are most likely due to the rather low mean body weight of the animals investigated.

In addition, cardiac index was measured in mice during hypovolemia and hypervolemia. Cardiac index gradually decreased during hypovolemia. The decline of cardiac index measured during increasing hypovolemia in halothane-anesthetized mice is in accordance with the well-known circulatory effects of hypovolemia that have been reported previously (20) for larger conscious species. However, for mice no data could be found in literature which provide a slope of the regression line between shed blood volume and cardiac index. Compared with that in rats or dogs (17, 25), the effect of hypovolemia on cardiac index appears to be less pronounced in mice. An increase in hypervolemia of ~30% in the total blood volume did not significantly affect cardiac index in mice. Possibly, the extent of hypervolemia in the present experiments was not marked enough to yield increases of cardiac index.

The present study shows that cardiac output of rats and mice can be calculated from conductivity-dilution curves after intravenous bolus injection of 5% glucose solution. Simultaneous recordings of plasma glucose concentration and blood conductivity in rats yielded corresponding values of cardiac output. In rats as well as mice, cardiac index was slightly but not significantly lower during halothane anesthesia. In mice, cardiac index was found to be significantly and gradually decreased during hypovolemia, whereas hypervolemia had no significant effect. Measuring blood conductivity in the manner presented here appears to be a reliable, inexpensive, and simple method for frequent quantifications of cardiac output in mice and rats.