|Home | About | Journals | Submit | Contact Us | Français|
Pulmonary mean-transit-time volume (PBV) of 12 anesthetized dogs was measured repeatedly by using pairs of dye-dilution curves recorded from the pulmonary artery (PA) and left atrium (LA) after right atrial injections of indocyanine green. In 345 pairs, the areas of the extrapolated LA curves exceeded those of the PA curves by 8% (sd, 14%), indicating that delayed color development of the dye-blood mixture is not a serious source of error in standard dye-dilution techniques. In eight dogs, measurement every 3–5 min before and after 2- or 5-min infusions of morphine (1 mg/kg) revealed no change in PBV, although arterial pressure decreased and cardiac output increased transiently. Prolonged (40 min) intra-arterial infusions of ATP, acetylcholine, or norepinephrine into four other dogs caused large changes in systemic and intracardiac pressures and in heart rate but not in PBV.
The pulmonary vasculature may act as a reservoir which can supply blood to the systemic arterial circulation (11). Whatever the details of the mechanisms of redistribution of blood volume in stressful situations, one might gain relevant information by exploring the relationships between pulmonary blood volume and flow, stroke output, and intravascular and intracardiac pressures, both under resting conditions and after the administration of pharmacologic agents to alter these physiologic variables. Morphine is of particular interest because of its dramatic effect in alleviating the distressful symptoms accompanying acute pulmonary edema. The commonest explanation (12, 30) of this effect has been that morphine decreases left ventricular end-diastolic and left atrial pressures, possibly associated with decreased pulmonary blood volume. It also would be of interest to study the effect of vasoactive agents such as adenosine triphosphate, norepinephrine, and acetylcholine.
Various methods of estimating pulmonary blood volume have been used for such studies. For the present study, the method of injecting into the right atrium with simultaneous sampling at the pulmonary artery and the left atrium in the intact dog was chosen because only one indicator is required and mixing of the dye with the blood is likely to be complete upstream from the pulmonary arterial sampling site. The contained volume between pulmonary artery and left atrium was calculated from the product of the flow, calculated from the two dilution curves, and the difference between the two mean transit times. This method is thought to have greater accuracy than other methods—for example, that used by Levinson and associates (14), of injecting into the pulmonary artery while sampling from the left atrium.
Because binding of the indocyanine green by plasma proteins is not instantaneous, when the injection site is in close proximity to the sampling sites there might not be time for the blood-dye mixture to achieve its equilibrium absorption spectrum. Earlier experiments (1) led to the conclusion that this was not a problem; however, Saunders and associates (24) suggested that, with the dye batches being supplied in 1967 and 1968, the dye-blood mixture reaches maximal light absorption at 800 mµ only after several seconds. Therefore, we investigated the possibility of delayed color development in the course of these experiments. These studies, described and discussed in a separate section before RESULTS, suggest that, although there may be some delay in color development, the effect is small and does not invalidate the dye-dilution technique used in this study.
The experiments were on mongrel dogs weighing between 15.7 and 26 kg. In the eight dogs used for studying the effects of morphine, anesthesia was achieved by intravenous injection of pentobarbital sodium (15–20 mg/kg) and α-chloralose (60 mg/kg). In four additional dogs studied during intra-arterial infusion of the other drugs, anesthesia was with pentobarbital and morphine. In both groups, supplementary doses of pentobarbital (60–120 mg) were given intravenously when necessary. The animals were intubated with a cuffed endotracheal tube and ventilated with room air via a Bird respirator (Bird Oxygen Breathing Equipment, Inc., Palm Springs, Calif.) at rates of 20–30 breaths/min. Peak positive pressure was maintained at 15–20 cm H2O. All measurements were obtained with the animal lying on its left side.
Catheters were positioned using fluoroscopy and pressure monitoring; a 100-cm 6-F bird’s-eye catheter was introduced percutaneously via the left jugular vein into the right atrium for dye injections; 70-cm Teflon catheters (0.93 ml total volume and 1.3 mm internal diameter) were passed percutaneously via a jugular vein to the pulmonary artery and, by means of transseptal puncture, to the left atrium. Polyethylene tubes (PE-90) were placed in the femoral vein for infusion of morphine, in the femoral artery for continuous pressure monitoring, and, via the other femoral artery, into the aortic arch for intra-arterial infusion of vasoactive agents. Catheters and densitometers were flushed with isotonic saline containing heparin sulfate (Fellows-Testagar) (10 mg/liter). Pressures in both atria and pulmonary and femoral arteries were simultaneously recorded using Statham (P23Db) strain-gauge transducers. The electrocardiogram, respiration rates, intracardiac and intravascular pressures, indicator concentrations, injection syringe travel, and time coder were recorded on an analog magnetic tape and a photokymographic recorder.
Indocyanine green injectate (1.25 mg/ml) was prepared by dissolving 50 mg of dye in 5 ml of triple-distilled water and then adding 33 ml of isotonic saline and, for dye stability, 2 ml of heparinized blood from the dog. With a pneumatically driven syringe, the dye solution was injected into the right atrial catheter, which was previously filled with dye solution. The dye-dilution curves were obtained by simultaneous sampling, via the identical catheters in the pulmonary artery and left atrium (1), through identical densitometers (XC100A, Waters Corp., Rochester, Minn.). A third densitometer was usually connected in series to that sampling from the main pulmonary artery, in order to record the optical density of the same dye-blood mixture after a mean delay of about 1 sec (see Discussion of methods). The sampling system volume (catheter tip to lumen of first densitometer) was 1.15 ml; thus, with a withdrawal rate of 24.7 ml/min (Harvard withdrawal syringe, Harvard Apparatus Co., Millis, Mass.), the mean transit time of each system was 2.8 sec.
At the beginning and end of the experiment, the densitometers were calibrated in series by using dye concentrations of 0, 5, 10, 15, and 20 mg/liter of blood. The mixtures were continuously stirred during this calibration.
The injection syringe was calibrated by recording, on both the analog magnetic tape and the photokymographic recorder, the syringe travel for measured volumes delivered.
After suitable control measurements of pressures and dye-dilution curves, the test drug was infused. Morphine sulfate in isotonic saline (1 mg/ml per kg of animal weight) was infused into the femoral vein over a period of 2 or 5 min in eight dogs. During the infusion and for the subsequent 30 min, dye-dilution curves were inscribed every 2.5–3 min while pressures from the different catheter sites were recorded. The dosages of the drugs given by intra-arterial infusion in four other dogs were: adenosine triphosphate (ATP), 1.23 mg/min; acetylcholine (ACh), 0.06–0.25 mg/min; and norepinephrine (NE), 5–20 µg/min. Measurements were made every 2–5 min for the duration of the intra-arterial infusion (40 min).
The following on-line analysis of the data was performed. The analog signals of the dye syringe travel (representing the amount of dye injected and also the zero time for the dye-dilution curve) and of the output from the densitometer were converted into digital form (15 samples/sec) by a CDC 3200 digital computer and a time-sharing monitor. The decay of the dye curve was extrapolated to base line by a single exponential continuance of the downslope established from 70 % to 30 % of the peak. The calculated mean transit times and flows were displayed, numerically, in the laboratory on an oscilloscope (Tektronix 564, Tektronix, Inc., Beaverton, Ore.).
The pulmonary blood volume was calculated from the product of the difference in mean transit times between the left atrial and the pulmonary dye curves (in seconds) and the flow from the left atrial dye curves (in milliliters per second) (1). Flows and volumes are expressed per kilogram of body weight.
A total of 1,033 dye-dilution curves were recorded from the pulmonary artery (PA) and left atrium (LA) after 387 dye injections in 12 dogs. For 259 of these injections, tandem sampling was used to compare the areas of the two curves obtained from the same site via different densitometers labeled Dl, D2, and D3. The mean transit time () between the tandem densitometers (calculated from the recorded dye-dilution curves) averaged 0.99 sec (sd = 0.16) from PA (Fig. 1), and 1.44 seconds (sd = 0.41) from LA curves. The respective appearance times (ta) and mean transit times for the proximal densitometers were 2.12 (sd = 0.26) and 4.63 (sd = 0.77) for the PA, and 3.72 (sd = 0.52) and 9.13 (sd = 1.42) sec for LA. There were no statistically significant differences in the areas of these paired curves. The densitometer positions were exchanged every few curves to avoid any bias due to errors in densitometer calibration. The positions of the densitometers and the site of sampling for the paired comparisons made no difference.
From 345 pairs of simultaneous dye curves sampled from the left atrium and pulmonary artery, using any combination of densitometers, the area of the former exceeded that of the latter by an average of 8 % of the mean of the two areas (sd = 14%); this difference was statistically significant (P < 0.001). When the ratios of the areas of the simultaneous dye curves from left atrium and pulmonary artery were plotted against the difference in mean transit times, difference in appearance time, or difference in time of injection to peak of dye curves, the scatter of the points was such that no statistical relationship could be observed between the ratios of the areas and these time differences. Hence, increasing the time interval between the two sampling sites did not show increasing disparity between their areas when the ta at the proximal densitometer was about 2 sec, tp, the time from injection to the peak of the curve, was about 3.4 sec, and was less than 5 sec. Arbitrary division of the differences in ta into three groups of 0.9–2, 2–3.5, and >3.5 sec and comparison of the ratio of areas of LA and PA curves did not show a significant difference among the three groups.
That this was not due to the method of extrapolation of the downslopes of the curves was shown by another experiment in which the extrapolation played a minor role in defining the areas of the curves. After injections of dye into the right ventricle, dye curves were recorded from the pulmonary artery by three densitometers in series. The appearance times were 1.5–2.4 sec at the proximal and 3.1–4.4 sec at the distal densitometers; the mean transit times to the proximal densitometer were 3.2–4.9 sec and the mean delays between the proximal and distal densitometers were 2.3–3.8 sec. As shown in Fig. 2, the areas of the curves from the distal densitometer all were larger than those from the proximal one; the differences in areas divided by the mean of the two areas averaged 8.6 % (range, 3–17%).
These comparisons of simultaneously obtained dye curves suggested that neither the time difference between injection and inscription of the curves nor the mean transit time difference between the simultaneous curves played the predominant role in the disparity between the values of the compared areas. If this was the predominant factor, then a statistically significant relationship should exist between the time differences (Δta, Δtp, and Δ) and the differences in areas; that is, as the time difference becomes less, the ratio of areas should approach 1.
We still have no explanation for the significantly smaller areas of PA curves compared to LA curves, a difference of about 8%. Similar differences, but of lesser magnitude, were found in previous experiments (1) when areas of simultaneous curves with ta values of 1–2 sec were compared to those with ta values of 10–20 sec. However, in that experiment the dye injectate was prepared in accordance with the manufacturer’s instructions and without the addition of albumin, and also it was reported that the standard deviation of the difference was most marked when the injection site was very near the proximal sampling site (injection, RV; sampling, PA and femoral artery). The conclusion reached was that the light absorption of this dye becomes stabilized in vivo within 1–2 sec after its injection into the central circulation of anesthetized dogs. An increasing deviation of the calculated from the measured flow with increasing flows observed by Saunders and associates (24), which would be seen as a decreasing ratio of PA/LA areas, was not apparent either in this study or in the earlier one (1).
In contrast, on the basis of a careful study using dog right-heart bypass preparation and several models, Saunders and associates (24) showed that flow calculations can be overestimated by 20–60% of the actually measured flow. Their conclusions were that the error was due to delayed development of maximal light absorption because of aggregation of particles of molecules of the indocyanine green. This aggregation was made worse by the addition of plasma or human serum albumin or by increasing the dye concentration. Similar disparity between measured and calculated flow was not seen using Evans blue dye. The presence or absence of blood to stabilize the dye solution prior to injection cannot be the explanation for the difference between their results and those of the present report and of Bassingthwaighte et al. (1), since Bassingthwaighte et al. did not use blood to stabilize their dye either. The effect of the temperature was stressed by Saunders et al. (24). They showed that the overestimation of flow using indocyanine green was decreased from 32 % (sd = 8) to 14% (sd = 10) when the temperature was increased from 25 to 37.5 C. Their model experiments usually were done at about 25 C and in the experiments on dogs with right-heart bypass, the temperature of the sampled blood was probably less than 32 C.
Errors in extrapolation of the downslope, resulting in overestimates of area, are more likely with LA curves than with PA curves; also, they are more likely with a PA curve if it has been dispersed by withdrawal through a sampling system of large volume. Such errors are in the appropriate direction to explain our observed results, but we doubt that extrapolation errors would be so large.
Another possible explanation is that the error in area is due to time-averaged sampling instead of the required volume-averaged sampling when flow is unsteady. At sites where the fluctuation of flow is greatest and where the passage time of the dye is least, there is maximal likelihood of overestimation of flow (2). The areas can be overestimated or underestimated but, with large numbers of curves (as in these experiments), the mean ratios of area of the PA curve to area of the LA curve will be significantly less than 1.0, and a ratio of 0.9 would not be unlikely.
In conclusion, our data do not support Saunders and co-workers’ (24) contention that the primary cause of the overestimation of flow calculations is that at least 6 sec are required for the indocyanine green-protein complex to develop. It is suggested that the <10% difference in the areas of PA and LA dye curves are more likely due to a combination of poor mixing, extrapolation errors, time-averaged sampling, and perhaps some small effect of delayed color development. It also is suggested that, if the temperature of the experiments for the two studies were closely similar, the results would be less divergent.
Data on eight dogs before, during, and after morphine administration are shown in Tables 1 and and22 and Figs. 3 and and4.4. Prior to morphine infusion, the control values for flow, pulmonary blood volume, heart rate, and mean arterial pressure were relatively stable in each dog. Morphine infusion did not cause a consistent pattern of changes except for mean systemic blood pressure, which consistently decreased during the infusion and tended to remain decreased during the subsequent half hour. Arrhythmias were not observed.
Although one might expect diminution in pulmonary blood volume to occur with decreases in pulmonary artery or left atrial pressure, no consistent associations were observed. In one dog (20.0 kg), mean pulmonary artery pressure increased from 14 to 23 mm Hg during infusion but pulmonary blood volume and flow were unchanged. In two dogs (22.5 and 23.5 kg) exhibiting high initial intracardiac pressures (Pla = 14 and 16 mm Hg; Pra = 9 and 10 mm Hg; Ppa = 22 and 24 mm Hg), morphine infusion caused pressure decreases of less than 2 mm Hg in these chambers, did not change cardiac output, and was followed in only one dog (23.5 kg) by a change in pulmonary blood volume 20 min later. In the one dog (15.7 kg) which showed a persistent small decrease in pulmonary blood volume after morphine infusion, pulmonary artery pressure decreased from 16 to 11 mm Hg, heart rate from 100 to 70/min, and mean systemic pressure from 180 to 120 mm Hg, but left atrial and right atrial pressures remained stable. In the dog which showed a delayed diminution in pulmonary blood volume (23.5 kg), there were no significant changes from the normal initial intracardiac pressures. In two dogs (17.5 and 21.0 kg), intracardiac pressure recordings were not obtained.
Figure 4 shows that morphine had no discernible effect on the ill-defined relationship between pulmonary blood volume and either cardiac output or stroke volume. Table 2 shows that, over wide ranges of the variables, pulmonary blood volume (range, 5.5–16.4 ml/kg) correlated only poorly with flow (60–256 ml/kg per min), stroke output (0.2–1.8 ml/kg per beat), heart rate (70–260 beats/min), and mean arterial pressure (60–178 mm Hg). It was not apparent that regression relationships other than linear would offer any better correlation.
These experiments initially were designed to study the influences of changes in cardiac output on the transport functions (the probability density functions of transit times) from pulmonary artery to left atrium (13). These drugs caused large changes in pressures (systemic and intracardiac) and in cardiac output and heart rate, but the effects on pulmonary blood volume were, at the most, modest (Fig. 5 and Table 3). A small increase in relative dispersion (standard deviation of the transport function divided by its mean transit time) occurred with longer mean transit times (lower cardiac outputs), and there was a high correlation (r = −0.74, N = 9) between low left atrial pressure and larger relative dispersion (see DISCUSSION). There was also low linear correlation of pulmonary blood volume with flow (r = 0.49) and with stroke volume (r = 0.33).
Although heart rate ranged from 52 to 210 beats/min and mean systemic arterial pressure ranged from 27 to 165 mm Hg, their influences on pulmonary blood volume were not statistically significant (Table 3).
The method of estimation of pulmonary mean-transit-time volumes used in this study is slightly more complicated than the classic technique of injecting into the inflow and sampling from the outflow (10, 26). While the general principles are the same, pulmonary artery injection is suspected to introduce a systematic error because it does not produce a flow-proportional labeling of the blood. Instead, it more likely produces a cross-sectional labeling, which means that an excessively large fraction of the injectate mixes into the slower moving blood along the vessel wall or in the sinuses above the pulmonary valve and prolongs the mean transit time, giving overestimates of volume.
However, other sources of error are common to both methods. First, exclusion of recirculating indicator by extrapolation of the downslope of the recorded curves is often unsatisfactory, but the likelihood of a systematic error is thought to be decreased by sampling at inflow and outflow. Second, error occurs because flow varies with cardiac and respiratory cycles, so that continuous sampling through a densitometer produces time-averaging of the blood-dye mixture at the catheter tip rather than the required volume-averaged (or flow-proportional) sampling. Fortunately, this results in relatively less error in than in F, but errors in estimated volume may be very large, perhaps up to 50 % (2). The left atrial curves were used for calculation of cardiac output, this arbitrary choice being made because of the longer time available for full color development and because the curves encompass more cardiac cycles and thus reduce the error due to unsteady flow. However, in the present report, the range of accuracy of blood volume measurement was not investigated. The control values prior to perturbation only give the range of repeatability but not the sensitivity of the method.
Reports differ with regard to the relationship between pulmonary blood volume and stroke volume, flow rate, and mean systemic arterial pressure (5–7, 14–16, 19, 22, 23, 29). Although some authors (29, 31) have observed no correlation between pulmonary blood volume and right atrial and pulmonary artery pressures in intact animals, in isolated perfused lungs Permutt and associates (17) found a close relationship between pulmonary blood volume and pulmonary artery pressure while left atrial pressure had little or no influence. In our experiments, left atrial pressure recordings were made at about 6- to 10-min intervals and there also was no apparent influence on pulmonary blood volume. The observation (13) that increased left atrial pressure resulted in diminished relative dispersion in transpulmonary transit times suggests that higher pulmonary venous pressures result in increased uniformity of regional pulmonary blood volume and flow. The present experiments do not provide a critical test of the influence of pulmonary artery or left atrial pressure either on pulmonary blood volume or on the variance of the transpulmonary transit times because the range of these pressures was small.
An increasing number of investigations have attempted to delineate the effect of morphine on the pulmonary vasculature, systemic circulation, and myocardium in both experimental animals and a variety of clinical situations (8, 12, 21, 25, 27, 30, 31).
Morphine causes an increased distensibility of venous capacitance vessels which is independent of the resistance vessels (12, 25, 31), although it is well recognized that morphine may cause significant systemic arterial hypotension (12, 21, 27, 28). Schmidt and Livingston (25) have shown that this pooling of blood occurs only in the limbs and skin but the mesenteric vascular volume is decreased. A positive inotropic effect has been found in dogs after administration of morphine, but this action could be blocked by adrenalectomy or beta-adrenergic blockade (30). Release of catecholamines after administration of morphine had been demonstrated by Bodo and associates (3) and, more recently, by Vassalle (32). Although increased blood levels of histamine have been observed after morphine administration (28), it is not known whether the amounts are sufficient to explain the hemodynamic changes observed with morphine.
The work of Vasko and associates (31) and Roy and associates (21) suggests that at least one mechanism of action of morphine in relief of pulmonary edema is a decrease of pulmonary blood volume. The lack of a significant decrease in the present study may be related to several factors. It has been demonstrated that positive-pressure ventilation may impair filling of the right ventricle with subsequent decrease in stroke volume and in flow of blood into the pulmonary circuit (4); it also may expel some blood from the thoracic cavity (9, 20). However, Permutt and associates (18), in their studies on isolated lung with careful control of lung inflation pressures, air volume of lung lobe, vascular pressures, and changes in vascular volume, came to the conclusion that the changes in vascular volume with respect to lung inflation depended on the inflow vascular pressure. At low vascular pressure, full lung inflation was associated with an increase in vascular volume; the opposite result occurred if the inflow vascular pressure was high. It should be noted that the vascular volume changes noted in these inflated lungs were of the order of about 2 ml for a transpulmonary pressure (airway pressure minus the equivalent of pleural pressure) of up to 30 cm H2O. Their studies were done in the left lower lobes of adult dogs. These different effects may tend to obscure any changes in pulmonary blood volume that might occur as a result of administration of morphine in dogs with normal pulmonary and left atrial pressures. It is not suggested that these reported negative findings can be extrapolated to spontaneously breathing dogs. It is also possible that in these apparently normal dogs the small changes in pulmonary blood volume cannot be detected because of the insensitivity of the method used.
The hemodynamic effects of acetylcholine, adenosine triphosphate, and norepinephrine are very large, even in these anesthetized, artificially ventilated dogs. They might be expected to cause some redistribution of blood volume between central and peripheral regions, particularly at the extremes of blood pressure changes, but such redistribution apparently did not occur. These experiments were of several hours’ duration and the trauma of repeated blood withdrawal and reinfusion and of blood dilution due to erythrocyte loss and hemolysis certainly did produce an unphysiologic state, which could mean that the neural and humoral mechanisms that can play a role in redistributing the circulating blood volume were depressed or otherwise ineffective. The correlation between cardiac output and pulmonary blood volume (r = 0.49) is not strong, nor is it clear that this is a physiologically useful adjustment. If an increase in pulmonary blood volume with exercise is an important mechanism in the development of the pulmonary edema that occurs frequently in untrained men climbing at high altitudes, then this correlation does not appear to be in our own best interests.
The authors are particularly grateful to the following for their assistance during these studies: Lucille Cronin, Joel Dunnette, Helen Parkinson, and Jane Irving.
Indocyanine green was kindly supplied by Hynson, Wescott & Dunning, Inc, Baltimore, Md.
This investigation was supported by Research Grants HE-9719 and FR-7 from the National Institutes of Health. T. Yipintsoi is a Special Fellow, National Institutes of Health. J. B. Bassingthwaighte is the recipient of a National Institutes of Health Career Development Award.