Main

Synthetic catecholamines, such as isoproterenol, can have a part in the treatment of children with left-to-right shunts and circulatory congestion when systemic blood flow and oxygen delivery remain insufficient after removal of excessive interstitial fluid by diuretics, afterload reducing agents, and correction of Hb concentration(1). The therapeutic effect of isoproterenol results from a dual mechanism of action. A peripheral vasodilatation redistributes left ventricular output from the pulmonary to the systemic circulation. Furthermore, the chronotropic effect may enhance systemic oxygen delivery, especially when cardiac function is depressed by extracardiac causes, because heart rate is an important modulator of cardiac output in neonates(2). The inotropic effect of isoproterenol is probably of less importance because agents with a predominantly inotropic effect have not been demonstrated to improve systemic blood flow in patients with left-to-right shunts(3, 4).

As a consequence of the chronotropic and inotropic effects of isoproterenol, myocardial work and oxygen consumption will increase. In a previous study we demonstrated that resting myocardial oxygen consumption is already increased in lambs with aortopulmonary left-to-right shunts as a consequence of the left ventricular volume load(5). Because myocardial oxygen extraction reserve is very limited, the increase in myocardial oxygen consumption must be realized by an increase in myocardial blood flow so that myocardial oxygen supply/demand ratio remains unaltered. This mechanism was also observed in chronicaly hypoxemic lambs(6). Because the maximum achievable myocardial blood flow will at best be equal to that of control lambs (in case of adequate capillary hyperplasia), the ability to increase myocardial blood flow, or coronary blood flow reserve, is decreased in shunt compared with control lambs. Moreover, the decreased endo-to-epicardial blood flow ratio of the shunt lambs indicates that their coronary blood flow reserve is decreased(7, 8). Any factor that further increases myocardial oxygen consumption thus constitutes a potential threat of inducing an imbalance between myocardial oxygen demand and supply.

Apart from increasing oxygen demand, catecholamines, through aβ3-adrenergic effect on adipose tissue(9), increase arterial FFA concentration, an important determinant of myocardial substrate uptake(10). Furthermore, they enhance lipoprotein lipase activity, situated on the myocardial endothelium, which enhances myocardial fatty acid uptake(11). Both these properties of catecholamines may thus induce an increased myocardial fatty acid utilization. Oxidation of fatty acids requires more oxygen per mol of ATP produced during oxidative phosphorylation, in comparison with the oxidation of glucose, a so-called oxygen-wasting effect(12). Manipulation of myocardial energy substrate utilization has indeed been shown to influence myocardial oxygen consumption. An increased contribution of fatty acids to myocardial energy metabolism was accompanied by an increased oxygen consumption(13), whereas a shift toward glucose utilization was accompanied by a decrease in myocardial oxygen consumption(14).

We hypothesize that isoproterenol increases myocardial FFA concentration and consequently myocardial FFA uptake. Such a finding would be relevant to the treatment of infants with left-to-right shunts because an increased myocardial fatty acid utilization could enhance myocardial oxygen uptake and, therefore, coronary blood flow. Considering their high resting coronary blood flow induced by the left ventricular volume load, any added increase in myocardial oxygen demand increases the risk for a mismatch between myocardial oxygen demand and supply with possible negative effects on long-term cardiac function.

METHODS

We studied 19 7-wk-old lambs of mixed breed with documented dates of birth. These lambs had also been included in previous studies from our laboratory on resting myocardial metabolism(15) and myocardial oxygen consumption during exercise(16). They were assigned to two groups: 10 lambs with an aortopulmonary left-to-right shunt and 9 lambs without a shunt. Until the day of study each lamb remained with its mother.

Surgical procedure. Surgical preparation, catheter care, and antibiotic administration were performed as described previously(17). In brief, after induction of halothane anesthesia we performed a thoracotomy through the 4th intercostal space and sutured a Goretex conduit (inside diameter, 6 mm; W. L. Gore and Assoc. Inc., Flagstaff, AZ) between the descending aorta and the main pulmonary artery, at the level of the fibrotic string of the ductus arteriosus. Precalibrated electromagnetic flow transducers (inside diameter, 10-15 mm; Skalar Medical, Delft, The Netherlands) were placed around the ascending aorta just above the coronary arteries and around the pulmonary artery proximal to the conduit. Polyvinyl catheters (inner diameter, 1.0 mm; outer diameter, 1.5 mm) were placed in the ascending aorta, coronary sinus, right ventricle, pulmonary artery, and left atrium. The chest wall was closed in layers, and the catheters and flow probe cables were led through a subdermal tunnel to a cloth pouch, sewn to the left flank of the lamb. For the control lambs surgical instrumentation was the same, except for the conduit, the flow transducer around the pulmonary artery, and the position of the catheter for mixed venous blood sampling for which the pulmonary arterial one was used, so that the right ventricular one could be omitted.

Experimental protocol. On the day of study, 12 ± 1(range 8-18) days after surgery, the lamb was weighed and brought to the experimental room where it was allowed to stand freely. To prevent interference with FFA metabolism, the heparin solution was carefully aspirated from the catheters. Measurements were started after a 2-3-h habituation period, but only when the lamb stood quietly. Throughout the experiment and the 2-3 h preceding, the lamb was allowed no food. Systemic and pulmonary blood flows, as well as aortic and left atrial pressures, were measured every 5 min over a 30-min period. At 15 and 30 min blood samples (0.7 mL) were withdrawn with a dry heparinized syringe from the aortic, coronary sinus, and mixedvenous catheters, which was the right ventricular one in the shunt lambs and the pulmonary arterial catheter in the control lambs. O2 saturation was determined in all samples; Hb concentration, hematocrit, pH, Pco2, Po2, and plasma HCO-3 concentrations only in the sample from the aorta. At 30 min, blood samples (9 mL) were drawn from the aortic and the coronary sinus catheter to determine the myocardial energy substrate concentrations in triplicate. Immediately after collecting these blood samples, radioactive microspheres labeled with either 141Ce,51 Cr, 103Ru, or 95Nb (NEN-Trac™, Dupont, Biotechnology Systems, Wilmington, DE) were injected into the left atrium, while simultaneously a reference sample was withdrawn with a pump (Harvard Apparatus Company, Millis, MA) from the aortic catheter into a preweighed, heparinized syringe for 1.25 min at a rate of 6 mL/min(18).

Isoproterenol was infused into the right atrial catheter at a rate of 0.1μg·kg-1·min-1 after the resting (baseline) period. The dose was based on a previous study(19) and is similar to the dose used in the treatment of infants with circulatory congestion. At 10 min after the start of the infusion, blood flows and pressures were measured, and blood samples withdrawn for determination of O2 saturation, Hb concentration, hematocrit, pH, Pco2, Po2, and plasma HCO-3 concentrations as described for the resting period. At 15 min, blood flow and pressure measurements and blood sampling were repeated. This was followed by the withdrawal of samples for the determination of the energy substrates and the injection of microspheres labeled with another isotope than the one used during the resting period. Because it takes at least 5 min to reach a steady state(19), we did not take measurements before 10 min after the start of the infusion.

Measurements and calculations. The precalibrated electromagnetic flow transducers were connected to Skalar MDL 400 flow m(Skalar Medical, Delft, The Netherlands). Systemic blood flow of the shunt lambs was obtained from the pulmonary arterial flow transducer which is situated proximal to the anastomosis of the conduit and the pulmonary artery and thus measures systemic venous return (see Fig. 1 inRef. 20). Pulmonary blood flow of the shunt lambs was obtained from the aortic flow transducer which is situated proximal to the anastomosis of the conduit and the aorta and thus measures pulmonary venous return. Systemic blood flow of the control lambs was obtained from the aortic flow transducer. The position of the aortic flow transducer in both groups was distal to the origin of the coronary arteries. To obtain total left ventricular output, the flow signal was later modified by adding coronary blood flow obtained with the microspheres(15). Heart rate was obtained from the blood flow signal. Aortic and left atrial pressures were measured with Gould P23 ID pressure transducers (Spectramed Inc., Oxnard, CA) referenced to atmospheric pressure with zero obtained with the pressure transducer at the right atrial level. All variables were recorded on an Elema Mingograf 800 ink-jet recorder (Siemens-Elema AB, Solna, Sweden). O2 saturation was determined in duplicate with an OSM2 hemoxymeter (Radiometer, Copenhagen, Denmark). Hb concentration was determined with the methemoglobincyanide method(21). Aortic hematocrit was determined in duplicate by the microcapillary method. pH, Pco2, Po2, and plasma HCO-3 concentrations were determined with an ABL-2 blood gas analyzer (Radiometer, Copenhagen, Denmark). Left-to-right shunt flow was obtained by subtracting systemic from pulmonary blood flow. Left-to-right shunt fraction was calculated by dividing left-to-right shunt flow by pulmonary blood flow. Blood O2 concentration was calculated as the product of O2 saturation, Hb concentration, and a Hb binding capacity of 1.36 mL/g(22).

Figure 1
figure 1

Oxygen extraction ratios of the various myocardial substrates for control and shunt lambs at rest and during isoproterenol infusion. § = p < 0.05, rest vs isoproterenol.

Full size image

Blood flow to the myocardium at rest and during isoproterenol infusion was determined with two of the four radionuclide-labeled 15-μm microspheres injected in random order. Each injection contained 1.5·106 microspheres for control lambs and twice as many for shunt lambs. After the experiment the lamb was killed with an overdose of pentobarbital i.v. Then the heart, cerebral hemispheres, and kidneys were removed and weighed. The heart was cleared of its pericardium, great vessels, chordae, and epicardial fat. To be able to study regional myocardial blood flows, the heart was fixed in formalin 8% for a week and then was reweighed for correction of mass changes during fixation. The left ventricular free wall was separated from the rest of the heart and further divided into three layers of equal thickness, parallel to the ventricular surface: subendocardial, middle, and subepicardial layers. All parts of the heart were subsequently weighed with an accuracy of 1 mg, and the results were multiplied by a correction factor for the weight changes during formalin fixation. Radioactivity in the left ventricular free wall, the rest of the myocardium, hemispheres, kidneys, and reference samples was determined in a Beckman 9000 gamma counter (Beckman Instruments, Fullerton, CA). Organ blood flows were calculated from the radioactivity counts in the tissue samples and arterial reference samples with the aid of a special computer program(23). Blood flows were expressed in mL min-1·100 g-1 wet weight. Adequate mixing of the microspheres in each lamb was confirmed by ascertaining that blood flow per 100 g of tissue to the two cerebral hemispheres, as well as to the two kidneys, did not differ by more than 10%(18).

Concentrations of glucose, pyruvate, lactate, β-OH-butyrate, and acetoacetate were determined in triplicate in whole blood by enzymatic methods(15). Immediately after sample collection, the blood(4.5 mL) was transferred to a tube containing a small quantity of NaF and carefully mixed to prevent glycolysis. Then the blood was deproteinized with perchloric acid. After subsequent neutralization to pH 7 with KOH and morpholinopropansulfonic acid solution, the mixture was centrifuged, and the supernatant was removed to determine the substrate concentrations(24).

Concentrations of FFA, total glycerol, and free glycerol were determined in triplicate in plasma. Immediately after sample collection, the blood (4.5 mL) was transferred to a chilled tube and centrifuged. The plasma was removed and stored without delay at -70°C, pending determination of the substrate concentrations. FFA concentration was determined enzymatically with a commercial kit (NEFAC, Wako Chemicals GmbH, Neuss, Germany). Total glycerol concentration was determined similarly (Test-Combination Triglyceride, Boehringer Mannheim GmbH, Mannheim, Germany). Free glycerol was also determined enzymatically(24). Plasma concentrations were converted to blood concentrations through multiplication by: (100 - hematocrit (%)/100. The triglyceride kit measures total glycerol concentration after hydrolysis of triglycerides. To obtain the triglyceride concentration we subtracted free glycerol from total glycerol concentration.

Because coronary sinus blood of lambs consists predominantly of venous blood from the left ventricular free wall, we calculated myocardial oxygen consumption of the left ventricular free wall as the product of aortocoronary sinus oxygen concentration difference and blood flow to the left ventricular free wall(25), obtained with the radioactive labeled microspheres. Similarly, we calculated myocardial substrate uptake as the product of aortocoronary sinus substrate concentration difference and blood flow to the left ventricular free wall. The substrate extraction ratio was calculated by dividing the aortocoronary sinus difference by the arterial substrate concentration. The myocardial oxygen supply/demand ratio was obtained by dividing the product of myocardial blood flow and arterial oxygen concentration by myocardial oxygen consumption. This ratio is the reverse of myocardial oxygen extraction.

The pressure-work index was obtained using the following formula(26): MVO2PWI = (K1·(SBP × HR) + K2·(0.8·SBP + 0.2·DBP) × HR × SV + K3)·K4, where MVO2PWI is myocardial oxygen consumption(μmol·min-1·100 g-1), SBP is systolic and DBP is diastolic aortic pressure (mm Hg), HR is heart rate (beats/min), SV is stroke volume (mL/kg), K1 = 4.08·10-4, K2 = 3.25·10-4, K3 = 1.43, and K4 = 44.6.

To assess the relative participation of each substrate in myocardial oxidative metabolism, we calculated the oxygen extraction ratio (OER)(12, 27). It was calculated by assuming that during a steady state all of the energy substrate taken up by the myocardium is oxidized, using the equation: OER = (ACSsubstrate × N)/(ACSoxygen), where ACS is the aortocoronary sinus difference and N represents the amount of oxygen required for the complete oxidation of 1 mol of substrate. Nglucose = 6, Npyruvate = 2.5, Nlactate = 3, Nβ-OH-butyrate = 4.5, Nacetoacetate = 4, NFFA = 25, and Ntriglycerides = 75 mol/mol(12, 27).

Statistical analysis. Data are presented as means ± SEM. The differences between the shunt and the control lambs, at rest or during isoproterenol infusion were evaluated using the Mann-Whitney test(28). The differences between the resting and isoproterenol infusion period were evaluated using the Wilcoxon rank test(28). A p value ≤ 0.05 was considered significant.

RESULTS

At rest. There were no differences in age (50 ± 1versus 47 ± 1 d) and weight (11.8 ± 0.8versus 12.8 ± 1.0 kg) between the control and the shunt lambs.

The left-to-right shunt led to significant hemodynamic differences between shunt and control lambs at rest (Table 1) which were similar to those reported from our laboratory previously(5). The aortopulmonary conduit gave rise to a shunt flow amounting to 57% of left ventricular output. Pulmonary blood flow, mean pulmonary arterial, and left atrial pressures, left ventricular stroke volume, and left ventricular free wall mass (29.4 ± 1.7 versus 21.7± 0.8 g; p < 0.05) increased substantially.

Table 1 Hemodynamic data and blood gases, myocardial blood flow, and oxygen consumption during isoproterenol infusion at a dose of 0.1 μg · kg-1 · min-1
Full size table

Resting myocardial blood flow, oxygen consumption, and pressure work index were higher in shunt than in control lambs, mainly as a result of the increased heart rate, because myocardial oxygen consumption per beat was not significantly different between the control and shunt lambs. The myocardial oxygen supply/demand ratio was similar in the two groups but the endo-to-epicardial blood flow ratio was lower in the shunt lambs. There were no significant differences in the arterial concentrations or arteriocoronary sinus differences of the myocardial energy substrates between the two groups(Table 2). The larger arteriocoronary sinus difference of lactate in the shunt lambs did not quite reach statistical significance (0.05< p < 0.10). In a previous study from our laboratory in which the number of measurements was larger, it did reach statistical significance(15). Myocardial substrate uptake in the two groups followed pari passu the myocardial blood flow and tended to be higher in the shunt lambs. There were no significant differences in oxygen extraction ratios between the shunt and the control lambs(Fig. 1).

Table 2 Arterial O2 and substrate concentrations and arteriovenous difference across and uptake by the left ventricular free wall at rest and during isoproterenol infusion at a dose of 0.1 μg · kg-1 · min-1
Full size table

Isoproterenol infusion. Isoproterenol infusion increased heart rate substantially in both groups, but to a lesser extent in shunt than in control lambs, so that there was no longer a significant difference between the two groups (Table 1). Pulmonary and systemic blood flows increased similarly in shunt and control lambs. Aortic blood pressures did not change significantly in the two groups. Left atrial pressure decreased by 3-4 mm Hg in both groups and remained substantially higher in the shunt lambs. Stroke volume decreased in the control lambs but not in the shunt lambs.

Myocardial oxygen consumption and pressure work index increased in both groups but less in the shunt lambs so that there was no significant difference between the two groups (Table 2). There was no significant difference in the increase of myocardial oxygen consumption and pressure-work index between the two groups: control lambs: 122 ± 28% and 102 ± 14%, respectively, and shunt lambs 49 ± 11% and 44± 5%, respectively. The increase in myocardial blood flow paralleled the increase in myocardial oxygen consumption, whereas the arteriocoronary sinus oxygen difference did not change. The myocardial oxygen supply/demand ratio increased significantly in the shunt lambs. The endo-to-epicardial blood flow ratio decreased significantly in both groups but was no longer significantly different between the two groups.

The arterial concentrations of all substrates increased, except for triglycerides, which decreased, and glucose, which did not change(Table 2). The most substantial increase concerned FFA; it was 3-fold. The increase could not be attributed to lipolysis of plasma triglycerides alone, which would have accounted for an increase in FFA concentration of approximately 150 μmol/L. The arteriocoronary sinus substrate difference of FFA did not change significantly. The same was true for the other substrates, except for triglycerides and β-OH-butyrate, which decreased. The increase in the arteriocoronary sinus difference of lactate in the control lambs did not reach statistical significance, but the change was larger than in shunt lambs.

Myocardial uptake of FFA followed the changes in myocardial blood flow because the arteriovenous difference did not change significantly. The increase in uptake did, in contrast to that of oxygen, not reach statistical significance, probably because of the scatter. Myocardial lactate uptake increased significantly in the control lambs but not in the shunt lambs. The release of free glycerol by the myocardium did not change(Table 2). The oxygen extraction ratio of triglycerides and β-OH-butyrate decreased significantly, whereas that of the other substrates did not change significantly (Fig. 1).

DISCUSSION

Despite a 3-fold increase in arterial concentration, an important determinant of myocardial substrate uptake(29), there was no increase in the arteriocoronary sinus difference of FFA during the infusion of isoproterenol. The relative contribution of FFA to oxidative energy metabolism as measured by the oxygen extraction ratio did not change from resting values. The metabolic response is similar in shunt and control lambs.

The finding that the aortocoronary sinus difference of FFA does not increase during isoproterenol infusion is in contrast with the present view on myocardial fatty acid utilization(10). The arterial FFA concentration, the myocardial workload, and alternating substrates are considered to be the major determinants of myocardial FFA extraction(10). During isoproterenol infusion, arterial concentration increased 3-fold, and myocardial work load, as reflected by myocardial oxygen consumption, increased 1.5-2-fold, but the aortocoronary sinus difference of FFA did not change. A similar discord with the present theory, but in opposite direction, has been observed in adult dogs(30). Plasma FFA concentration was lower in dogs 20 min after propanolol (0.15 mg/kg) injection than in control dogs (300 ± 45versus 490 ± 70 μmol/L, p < 0.05), but this was not accompanied by a significantly decreased myocardial fatty acid uptake(3.8 ± 0.5 versus 4.2 ± 0.3% of injected131 I-heptadecanoic acid). The present data, strengthened by those of Van der Wall et al.(30), suggest that, although positive correlations between arterial concentration and aortocoronary sinus difference have been found in human volunteers(31, 32), myocardial fatty acid extraction is not a simple function of arterial concentration under changing conditions; more complicated mechanisms for the regulation of fatty acid utilization must exist at the myocardial level.

Myocardial blood flow reserve is not encroached upon during isoproterenol infusion. This can be deduced from the endo-to-epicardial blood flow ratio which is not significantly lower than in control lambs and still above 1.0(7, 8). In addition, the myocardial oxygen consumption does not lag behind the predicted myocardial oxygen consumption(pressure work index), and no lactate is produced by the myocardium indicating that myocardial blood flow meets the myocardial oxygen demand.

Myocardial blood flow reserve is not encroached upon in the shunt lambs because myocardial work increases less than in control lambs, which is mainly due to the smaller increase in heart rate owing to the high resting heart rate of the shunt lambs. In addition, myocardial oxygen consumption per beat does not increase significantly in the two groups despite the positive inotropic effects of isoproterenol. The hemodynamic changes must therefore lead to different loading conditions of the heart which decrease myocardial oxygen consumption opposing the inotropic effect. Isoproterenol decreases left atrial pressure which indicates that left ventricular end-diastolic volume probably decreases. A lower end-diastolic volume is accompanied by a lower wall stress, one of the major determinants of myocardial oxygen consumption(33). Systolic aortic pressure and left ventricular stroke volume, which influence left ventricular wall stress and external work, are not significantly affected in the shunt lambs during isoproterenol infusion. Oxygen supply thus readily matches the increased oxygen demand by an increase in myocardial blood flow. The oxygen supply/demand ratio even improves in the shunt lambs because the vasodilatory effect of isoproterenol on the coronary vascular bed exceeds the increase in myocardial work.

Lactate uptake by the myocardium was not significantly different between the two groups during isoproterenol infusion, in contrast to at rest. This was due to an increase in the arteriocoronary sinus lactate difference in the control lambs. In our previous study we hypothesized that the higher resting lactate uptake in the shunt lambs could be the result either of the production of fetal isomyosins or of the higher resting cardiac work load(15). This study, in which the measurements were made at a similar cardiac work load, suggests that the higher work load of the shunt lambs at rest is responsible for the increased myocardial lactate uptake.

Although the sum of the oxygen extraction ratios does not significantly change from rest to isoproterenol infusion, Figure 1 shows a decrease which is mainly the result of a decrease in the contributions of triglycerides and β-OH-butyrate. This suggests that during the onset of a steep increase in myocardial work load, energy substrates are derived from endogenous stores. Catecholamines can stimulate both glycogen and triglyceride breakdown(10, 11, 34). The increased pyruvate release from the myocardium may indeed reflect an enhanced glycogen hydrolysis, but the unchanged glycerol release does not point to an enhanced endogenous triglyceride usage during isoproterenol infusion. The abundant presence of FFA could be responsible for this because this has been shown to prevent lipolysis but not glycogenolysis in myocytes during epinephrine infusion(35–37). Studies using isotope-labeled substrates are, however, required to elucidate a possible use of endogenous substrate stores during isoproterenol infusion.

Another possibility for the unaccounted part of the oxidative metabolism is the utilization of other substrates than the ones we measured. Acetate could be such a substrate, as discussed previously, because it is liberally produced by fermentation in the ovine stomachs and is readily taken up by the myocardium as an energy substrate(15, 38, 39).

The Hb concentration increases by approximately 10% in both groups during isoproterenol infusion. This is the result of transfer of fluid from the intravascular to the interstitial compartment(40). It constitutes an additional beneficial effect on systemic and myocardial oxygen delivery in the treatment of children with insufficient systemic circulation.

In conclusion, the infusion of isoproterenol is not accompanied by an imbalance between oxygen supply and demand. A shift toward a proportionate larger FFA uptake, and thus a potential oxygen wasting effect, does not occur. In addition, myocardial work load of the shunt lambs increases less than in control lambs, and the loading conditions of the left ventricle change favorably.