JOURNAL
OF APPLIED
Vol. 39, No. 2, August
Shunt
PHYSIOLOGY
1975.
Printed
in U.S.A.
dynamics
in experimental
atria1 septal defects JAMES A. ALEXANDER, JUDITH C. REMBERT, WILL C. SEALY, AND JOSEPH C. GREENFIELD, JR. Departments of Surgery and Medicine, Duke University Medical Center, Durham 27710; Veterans Administration Hospital, Durham, North Carolina 27705
ALEXANDER, JAMES A., JUDITH C. REMBERT, WILL AND JOSEPH C. GREENFIELD, JR. Shunt dynamics in atrial septal defects. J. Appl. Physiol. 39(Z) : 281-286.
C. SEALY, experimental
1975.-In order to study the hemodynamic variables involving the magnitude, direction, and timing of phasic shunt flow, both the interatria1 pressure gradient and blood flow along with other pertinent hemodynamic variables were measured instantaneously across a surgically created atria1 septal defect (ASD) in seven awake dogs. Atria1 and ventricular pacing and infusion of phenylephrine and isoproterenol were used to alter hemodynamic conditions. The wave form of phasic ASD flow was similar both in configuration and timing to the interatrial pressure gradient. During the cardiac cycle, both left-to-right (L-R) and right-to-left (R-L) shunting occurred: atria1 contraction augmented L-R flow; the onset of ventricular contraction was associated with R-L flow; during the latter part of ventricular contraction, flow returned to L-R with the maximum L-R shunting occurring in early diastole. Tachycardia, infusion of phenylephrine and isoproterenol did not alter the phasic flow pattern. Both spontaneous and positive pressure respiration decreased net L-R shunting. interatrial
pressure gradient;
respiration;
blood flow
A NUMBER of cardiovascular physiologists have investigated the complex hemodynamic relationships that influence the direction, timing, and magnitude of blood flow across an atria1 septal defect in both experimental animals (5, 13) and patients (Z-4, 7, 9, 11). However, a precise description of the nature of shunt flow has been limited since it has not been possible previously to measure the phasic flow directly. It is the purpose of this report to describe the contour of pulsatile flow that occurs in a surgically created atria1 septal defect and to relate the characteristics of the flow to the interatrial pressure gradient and to several hemodynamic variables that might influence flow. METHODS
Adult mongrel dogs weighing 25-30 kg were anesthetized with sodium pentobarbital (30 mg/kg), intubated, ventilated with a respirator (Mark VII, Bird Corp., Richmond, Calif.), and underwent a right thoracotomy via the fourth intercostal space. The ascending aorta was dissected from the surrounding tissue, and an electromagnetic flowmeter (EMF; TTQ-type probe, model
and
M-4000 flowmeter, Statham Instruments, Inc., Oxnard, Calif.) probe of appropriate size was fitted around the aorta as described previously ( 1). Bipolar epicardial pacing electrodes were sutured to regions near the sinus node and right ventricular apex. A purse-string suture was placed around the right atria1 appendage and a clamp positioned just below this suture. Beginning at its junction with the inferior vena cava, a 4-cm section of the right atrium was clamped and the atrium incised. Traction sutures were placed on each side of the cut atria1 edge. Heparin (2 “g/kg) was administered intravenously as a bolus. The heart emptied following occlusion of the venae cavae, and ventricular fibrillation was induced. The clamps were removed-from both the atria1 appendage and atrium. An atria1 septal defect was created, and a specially designed l.O-cm (inside diam) EMF probe (model RC1000, Micron Instruments, Inc., Los Angeles, Calif.) was sutured to the cut edges of the septum with interrupted silk sutures. The probe wires were brought out through the right atria1 appendage and the purse-string suture ‘secured. The atria were flooded with saline, then reclamped and the venae cavae released. The heart was defibrillated, and the atriotomy was closed. Total occlusion time varied from 2.5 to 3.5 min. Two identical Z-mm (inside diam) especially made catheters 30 cm long having two lateral pressure openings at the end were inserted either through the pulmonary vein from the superior segment of the right upper lobe into the left atrium or through the azygos vein into the right atrium. The tips were positioned so that they were separated by approximately 5 cm. A # 8 Lehman catheter (U.S. Catheter and Instrument Corp., Glens Falls, N. Y.) was passed into the left ventricle from the left femoral artery. Another no. 8 Lehman catheter was placed in the right ventricle from the right femoral vein. A suction tube was placed in the seventh intercostal space and under 10 cm water suction. The thoracotomy was closed. After a recovery period of 2-3 h, the animals were awake, breathing spontaneously, and having no apparent difficulties. The catheters in the left and right atria were connected to matched high-fidelity transducer (model P23Db, Statham Instruments, Inc.)-amplifier (model 350- 1500, Hewlett-Packard Co., Waltham, Mass.) systems. In order to measure the interarterial pressure gradient, the outputs of these systems were fed into an analog computer (model
281
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
ALEXANDER
282 3400, Systron-Donner Corp., Concord, Calif.) subtraction circuit and the pressure differences obtained. The recording system was statically balanced to within +O. 1 % for O-300 cmH 20. The dynamic characteristics were evaluated before and after use with a sinusoidal generator, the unbalance being no greater than 5 % through 30 Hz. These recording specifications have been shown to yield a valid pressure gradient for studies of aortic flow (6). Left and right ventricular pressures were obtained by connecting the catheters to transducers (model P23Db, Inc.). The respiratory cycle was Statham Instruments, monitored with a pneumotachygraph. Ventricular pacing was carried out by connecting the epicardial pacing electrodes to a stimulator (model SR4, Grass Instrument Co., Quincy, Mass.) that d e 1ivered square wave pulses of 3-ms duration 10 % above threshold voltage through the isolation unit. During the study, flow across the atria1 septal defect, aortic flow, left ventricular pressure, right ventricular pressure, left atria1 pressure, right atria1 pressure, respiration, the interatrial pressure gradient, and lead II of the electrocardiogram were continuously recorded on two direct-writing oscillographs (model 958- 100, series 7700, Hewlett-Packard Co.) with the first seven parameters also directly recorded on FM magnetic tape (model 3955-D, Hewlett-Packard Co.). Control measurements with the animal breathing spontaneously in normal sinus rhythm were recorded. Pacing was begun at a rate of 150 beats/min and continued for 3 min at each site. The rate was increased by approximately 30-beats/min increments in a random sequence through 240 beats/min for each atria1 and ventricular site allowing at least a 5-min interval to elapse between each new rate. Pacing was discontinued and the animal allowed to return to sinus rhythm. A lOO-pg/min infusion of phenylephrine was given intravenously in order to increase the mean arterial pressure. After data were recorded, the infusion of phenylephrine was discontinued and an infusion of isoproterenol (1 pgjmin) was begun. When all parameters returned to the initial control levels, the animal was connected to a respirator to accomplish positive pressure breathing (+ 20 cmH 20). Succinylcholine administered as a 60-mg bolus was given as needed during this portion of the study. With respirators interrupted for a 10-20-s period, the changes in septal defect flow produced by respiration could be evaluated separately. In addition, the effects produced by increasing the intrabronchial pressure to 40 cmHz0 for a 30-s period were studied. At the conclusion of the studies the animals were killed with a lethal dose of pentobarbital and the defect examined to be certain that the probe covered the orifice adequately. A low-profile ( 10 mm, ID) especially designed EMF probe (model RC-1000, Micron Instruments, Inc.) coated with liquid silicone rubber and having external Teflon sewing rings was used to measure shunt flow instantaneously. Stroke volume also was measured instantaneously with an EMF probe (TTQ-type probe, model M-4000 flowmeter, Statham Instruments, Inc.) on the ascending aorta, assuming the end of diastole as zero flow. All probes were calibrated previously both in vitro and in vivo by allowing varying known amounts of saline or blood to flow through
Zero
Zero
ET
AL.
Zero
1. Recording obtained during calibration with the probe mounted in a calibrating device and blood allowed to flow through the probe. Since there is no vessel in the lumen of the probe, this situation is analogous to that obtained during measurement of shunt flow in the atria1 septal defect. Note the exact similarity between the true zero obtained by stopping the flow and that obtained with the electrical zero. FIG.
remained in a known length of time. The calibrations within a standard deviation of =t 1.7 % for the septal defect probes and =t3.7 % for the aortic probes ; both were linear within zt2 %. The phase lag produced by the septal defect flowmeter was linear with frequency and had a phase lag of approximately 0.01 s. Since the two probes are energized differently, no electrical interference was encountered during the study. Pressure and flow data were analyzed either directly from the oscillographic recordings or from tape using a digital computer (model 1130, IBM Corp., Armonk, N. 76.) in conjunction with an analog-to-digital converter (model 663, Redcor Corp., Canoga Park, Calif.). The directional and total shunt flow was computed directly from the flowmeter sutured to the defect. Pulmonary flow (QP) was calculated from the aortic flow (QS) by adding the left-to-right shunt flow and subtracting the right-to-left shunt flow. Modeling functions of the septal defect flow were carried out with an analog computer (Simulators, Inc., Northbrook, Ill.). To measure the phasic shunt flow it was essential that the electrical zero reference be precise. Thus the flowmeter electrical zero was tested on multiple occasions in vitro during the course of routine calibration procedures by obtaining electrical zero and then by comparison to mechanical zero flow as illustrated in Fig. 1. In this situation, electrical zero and mechanical zero flow were essentially the same. The possibility that spurious flow signals might be induced by movement of the probe with the atria1 septum during cardiac contraction was studied also. The probe was placed in a container of saline and oscillated at approximately ZOO-ms intervals over a distance necessary to produce a signal of approximately the same magnitude as that found with flow in the septal defect. It was calculated that in order to produce a spurious flow signal equivalent to this peak flow signal, the probe must move approximately 15 cm/s. Thus the movement necessary to produce this magnitude of flow during systole, i.e., 200 ms, was 3.0 cm. In three dogs the EMF probe was attached to a rigid bar and sutured into the atria1 septal defect as described above. The probe was kept motionless with the bar, and no change in the contour of the flow signal was noted. Silver clips were implanted on the atria1 septum in the other dogs and their movement studied with cinefluorography. No significant movement of either the probe or clips was noted. These data substantiate the fact that movement of the atria1 septum to a degree necessary to give erroneous data is unlikely.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
EXPERIMENTAL
Erpimtion Reumotach
ASD Shunt Flow
ATRIAL
SEPTAL
DEFECTS
283
FIG. 2. Recordings of various hemodynamic parameters during the control state during spontaneous respiration. Note the marked similarity between the
F.e..----
1 Inqimttm L-R
configuration of the atria1 septal defect shunt flow and the pressure gradient. The increase in the direction of a right-to-left shunt during inspiration is obvious. The phasic contour of the shunt flow is typical of those obtained in other dogs studied.
;z;;E./ R+L
RESULTS
ECG
-F---
Complete hemodynamic studies were available from seven dogs. The contour of the phasic flow through the atria1 septal defect measured with the electromagnetic L-+R flowmeter probe was correlated with the pressure gradient. PRESSURE ‘O cm “a”p ,I RJL4,&&/‘~, GRADIENT IOcmH20 The wave form of the shunt flow was markedly similar R-+L both in configuration and timing to the pressure gradient in ail situations studied. The configuration of the flow illustrated in Fig. 2 appeared to be quite consistent in the COMPUTED SHUNT FLOW seven dogs studied. During atria1 contraction the direction of shunting was enhanced from left to right. Immediately following the onset of ventricular systole, the direction of L-tR ASD 50cm)/se;[ the shunt rapidly changed, usually resulting in a net rightto-left shunt for this period. During the latter part of SHUNT FLOW SOc,qsec A-it ventricular systole the shunt returned to a left-to-right ’ direction so that the majority of the left-to-right shunt ’ ISEC occurred in early diastole. FIG. 3. Relationship of the shunt flow computed using Eq. I The configuration of the phasic flow illustrated in Fig. 2 can be compared to the true shunt flow measured with the electromagnetic flowmeter. Note the marked similarity in both the configurawas similar regardless of whether the net shunt was righttion and timing. In A the dog was unpaced. In R, obtained during to-left or left-to-right, i.e., the configuration remained the rapid atria1 pacing, there was a marked pulsus alternans of both the same but the base line shifted. In all situations studied right and left ventricle. The shunt flow also varied with the alternans, there was a period of both left-to-right and right-to-left the decreased shunting occurring with the weak beat. Even with these marked changes in hemodynamics, the computed flow is almost shunting during the cardiac cycle. In the control state the identical to the true flow. The output filter on the electromagnetic magnitude of the peak left-to-right and right-to-left shunt flowmeter will cause a delay of approximately 0.01 s. When this is was similar and in the range of f50 cm3/s. During the taken into account, no phasic delay between the computed shunt high heart rates produced by pacing, a similar contour flow and the true shunt flow can be determined at any time during was observed except that the shunt flow during diastole the cardiac cycle. from left to right was constant with no obvious increase in flow during atria1 systole (Fig. 3). This configuration was Table 1. During the control state the QP/QS ratio had an not altered by ventricular pacing except for the absence average value of 1.55 f 0.14. During this time the preof the shunt related to atria1 systole. During both atria1 dominant shunt was left-to-right; however, a consistent and ventricular pacing no significant change in average right-to-left shunt occurred in all dogs and averaged 53 shunt flow could be demonstrated from rates of 150-240 f 8% of the left-to-right shunt. Following the marked beats/min, although there appeared to be a trend to inincrease in mean arterial blood pressure produced by crease the mean left-to-right shunt flow slightly at moderate phenylephrine, the magnitude of the left-to-right shunt heart rates. In general, high pacing rates were not associincreased so that the average QP/QS ratio increased to ated with greater peak-to-peak flows. 2.72 f 0.42. The right-to-left shunt decreased in all dogs The magnitude of the bidirectional shunt flow during and was 18 f 5 % of the left-to-right shunt. The phasic normal sinus rhythm in the control state and during incontour of the flow was unaltered at this time but appeared fusion of phenylephrine and isoproterenol are given in only to shift the entire flow in the direction of a more Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
284
ALEXANDER
TABLE
1. Bidirectional
shunt flow during normal sinus rhythm Hemodynamic
Animal No.
1 2 3 4 5 6 7 Mean &SEM 1 2 3 4 56 7
Heart Rate, beats/ min
Arterial Blood Pressure, (mean>, -Hi?
135 88 165 150 150 90 143 132 zk12
95 130 115 140 105 85 105 110 *7
135 110 150 160 150 100
125 185 230 180 250 215
Data
Cardiac ($~$ min
Mean Left-toright, cms/min
1520 1150 1050 950 1180 1160 1110 1160 =t70 B. 2400 940 660 950 1720 1380
Shunt
QRl{($os
RightLeft/ LeftRight, %
1200 160 170
1.79 1.14 1.16
24 86 72
930 1070 320 640
1.98 1.91 1.27 1.57
46 41 67 34
640 ~160
1.55 ho.14
53 zt8
2080 1190 1470 2980 1340 900
1.87 2.25 3.23 4.13 1.78 1.65
240
2480
4.15
9
340 rt80 0.01
1780 zk280 0.004
2.72 ztO.42 0.03
18 z/z5 0.006
1.56 1.07
75 12
Ri,::-tocms/min
A. Control 380 1580 1100 940 610 440 1720 790 1810 740 970 650 970 330 1250 610 A170 It90 Pheny Zej hrine 2370 290 1770 580 1490 20 3150 170 1940 600 1400 500
Net flow cms/min’
80
230
790
126 =tll NS
200 zk16 0.003
1260 =t230 NS
21
133 122
150
1630 3460
1050 1000
130 750
3
174
190
1320
50
1030
4 56
184 138 156
200 220 115
2650 2580 2860
2110 1190 810
160 530 120
1950 660 690
1.74 1.25 1.24
Mean &SEM P
7 Mean & SEM P
2720
Flow
2120 &250 0.006 C. Zso~roterenoZ
920 250 -980
12 33 1 5 31 36
0.26 8 45 15
176
180
3950
660
1020
- 360
0.91
154 zt12 NS
172 &14 0.003
2640 A350 0.008
980 zk240 NS
510 zt150 NS
450 &360 NS
1.15 =tO. 18 0.02
P values listed below B and C compare these mean values during the control state given in A. NS signifies that P > 0.05.
with
those
obtained
marked left-to-right shunt. Following isoproterenol infusion, the systemic flow increased markedly and, at the same time, the left-to-right shunt in most of the animals was at a level similar to that found during the control state. The effects on the right-to-left shunt were variable such that in five dogs the right-to-left shunt decreased below control values. However, in two dogs this flow increased rnarkedly so that the net shunt was right-to-left. During this time the phasic flow contour was not altered and the magnitude of the peak-to-peak flow appeared to be similar to that during the control conditions. Note that these drugs were not given to specifically test a drug effect but rather to change the hemodynamic conditions so that these effects on the phasic shunt flow and its relation to the \ pressure gradient could be evaluated. It was difhcult to determine precisely the effects of respiration on shunt flow because of the constant influence of intracardiac hemodynamics. In general, inspiration was accompanied by either a decrease in the magnitude of the left-to-right shunt or by an augmentation of the rightto-left shunt depending on when inspiration occurred during the cardiac cycle (Fig. 2). Spontaneous respiration decreased the net left-to-right shunt flow in the seven animals by an average of 300 cm3/min when compared to flow in the same animal when respiration was temporarily stopped. Positive pressure respiration also was associated with a decrease in left-to-right shunting and appeared to be
ET AL.
similar to that produced by spontaneous respiration. The effects of respiration were similar during all hemodynamic conditions studied. Sustained intrabronchial pressure of 40 cmHx0 reduced the mean shunt flow to* essentially zero and also markedly decreased the amplitude of the phasic flow. In attempting to evaluate the determinants of the pressure-flow characteristics across the atria1 septal defect, a model similar to that assumed to exist in the ascending assumptions, the aorta was used. With certain simplifying Navier-Stokes equations of fluid motion can be reduced to -
dP -= dZ
LZ+Rq
where z is the axial coordinate, p is the pressure, q is the instantaneous blood flow, and t is time. The terms L and R are the inertial and frictional components, respectively (6). This equation was solved for flow continuously with an analog computer using the measured pressure gradient as the forcing function (Fig. 3). These studies were carried out using the data that had been recorded on tape. The coefficients of the resistance and inertial components were modified to obtain the best empirical fit between the actual measured flow and the computed flow. The dominant factor in the computed flow was found to be the resistance term. It appeared that a linear relationship existed between the pressure gradient and the amount of flow through the orifice. The inertial component was quite small since, as can be seen in Fig. 3, there is virtually no phase shift between the observed flow and the computed flow. U
DISCUSSION
A number of hypotheses have been formulated to account for the magnitude and timing of blood flow across an atria1 septal defect (8). In anesthetized dogs, Weldon (13) found that as soon as an atria1 septal defect was created, shunting occurred from the left to the right atrium, and he attributed this finding to a differential ventricular distensibility. Dexter (4) advocated the concept that leftto-right shunting was due to the relative resistance to ventricular filling. Uhley (12) stated that the left-to-right shunt was related to the more cephalad position of the left atrium as compared to the right, presenting the gravitational factor as being the major mechanism. Cournand et al. (3) noted that different dynamic conditions in each atrium appeared to be contributing factors to the shunting mechanism. Wiggers (14) suggested that the up-and-down movements of the ventricular base produced greater fluctuations in the left atria1 pressure than in the right and probably accounted for a transient right-to-left shunt following the onset of ventricular contraction. In 1947 Cournand et al. (3) found that the mean pressure was higher in the left than the right atrium, providing a gradient for shunting the blood in a left-to-right direction. Levin et al. (9) measured simultaneous left and right atria1 pressures and computed the pressure differences in children with 1.5-3-cm diameter atria1 septal defects. Biplane cineangiography was used to evaluate timing of shunts in various phases of the cardiac cycle. They found that left-to-right shunting across the atria1 septal defect
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
EXPERIMENTAL
ATRIAL
SEPTAL
285
DEFECTS
half of ventricular occurred primarily over the latter systole with the peak occurring at the apex of the V wave during atria1 systole. The and a second accentuation pressure cgradient wave form and timing of shunting observed by Levin et al. (9) is very similar to the observations Our study shows that shunt in our animal preparation. flow through an atria1 septal defect is pulsatile in nature and follows exactly the pressure gradient generated between the two atria. The determinants of the interatrial pressure gradient may well be explained by the anatomic variation of the right atri UIII and the nat ure of th e inflow svs‘ten1 to ea ch of the atria and ventricles. The left a trium has a th .icker wall and is therefore less deformable than the right. The venous reservoirs have smaller capacity, i.e., the pulmonary veins being shorter and their diameter smaller than the superior and inferior venae cavae. The effect of the muscular development of the ventricles is different, i.e., the left ventricle being thicker and less compliant than that of the right. Thus the effective atrioventricular orifice areas, ventricular compliance, and end-diastolic pressures appear to affect the timing and duration of phasic shunting. Right-to-left shunt flow occurs at a time when the atrioventricular valves are closed and during the upstroke of the ventricular pressure tracing. This coincides with the atria1 C wave, thought to be caused by the bulging of the atrioventricular valves. The tricuspid valve, having 50-75 % more area than the mitral valve, may bulge more and may be the explanation of right atria1 pressure being higher than left atria1 pressure at this time. During the latter half of systole, corresponding to venous inflow into the atria, left atria1 pressure increases and the shunt flow direction is changed to left-to-right. The lower left atria1 and pulmonary venous compliance may well account for the shift in the pressure gradients. The diastolic phase shows flow and gradients A less compliant left ventricle in a left-to-right direction. with a smaller mitral valve orifice size and higher enddiastolic pressure is the most likely explanation for this finding. The augmentation of left-to-right flow during atria1 systole may be explained also by this difference in compliance. The magnitude of the right-to-left shunts which were measured in these animals appears to be considerably greater than that usuallv found in patients with atria1 septal defects. A major portion of thii difference may be attributable to such factors as acute experimental preparation, anesthesia, and surgical trauma. Obviously the hemodynamic status of these animals is not identical to a patient with a congenital atria1 septal defect. However,
differences in the methods used to measure the shunts also may account for some of this discrepancy. Although the QP/QS ratios obtained in this study are correct, the values reported for the shunt flows represent the total volume of blood that moves across the defect. Since the primary right-to-left shunt occurs during early systole when the mitral valve is closed, much of the blood shunted into the left atrium may be returned to the right atrium during the latter part of systole when the shunt becomes left-toright. Thus the magnitude of the right-to-left shunt measured by the flowmeter cannot be compared to right-to-left shunts detected by either indicator dilution or Fick techniques. That this is the case is supported by the data reported by Levin et al. (9) in patients with atria1 septal defects. Every one of their patients was found to have rightto-left shunting during the onset of left ventricular contraction, but none were found to have a right-to-left shunt which could be detected by indicator dilution techniques. The effects of respiration on shunting in atria1 septal defects have not been previously studied directly. Osher (10) demonstrated by phonocardiography that right atria1 murmurs were inconstant, extending from midsystole to late diastole or being confined to diastole, intensifying with expiration, and either diminishing or disappearing with inspiration, which suggested that the shunting in atria1 septal defects was influenced by respiration. In our studies, flow was influenced by respiration. Inspiration either decreased the left-to-right shunt or augmented the right-toleft shunt, depending on when it occurred during the cardiac cycle. The studies of the relation between the pressure gradient that resistance and the flow using Eq. I demonstrated was the primary determinant of the pressure gradient. It would be extremely useful if this relationship could be used to compute the shunt flow from the pressure gradient. Unfortunately, in order to accurately quantitate the precise geometry of the defect, the exact relation between the interatrial pressure catheters must be known. Thus its application to clinical studies will be extremely difficult. The authors express their appreciation to Drs. Frank Starmer and Philip McHale, computer modeling and data analysis; Mr. William Joyner and Mr. Kirby Cooper, surgical technicians; Mrs. Rosa B. Ethridge and Mrs. Brenda Haley, typists; and the Department of Medical Illustration of the Veterans Administration Hospital. This work supported in part by Grants HL-09711 and HL-01782 from the Public Health Service, and is designated Veterans Administration Program No. 3330-03. J. C. Greenfield, Jr., is recipient of Research Career Development Award 1-K3-HL-28112 from the Public Health Service. Received
for
publication
18 December
1974.
REFERENCES 1. ALEXANDER, J. A., W. C. SEALY, AND J. C. GREENFIELD, JR. Improved technique for implanting electromagnetic flowmeter probes on the coronary artery. J. A@Z. Physiol. 27: 139-140, 1969. 2. BRANNON, E. S., H. S. WEEMS, AND J. V. WARREN. Atria1 septal defect: study of hemodynamics by the techniques of right heart catheterization. Am. .I. Med. Sci. 2 10 : 489-491, 1945. 3. COURNAND, A., H. L. MOTLEY, D. D. HIMMELSTEIN, AND J. BALDWIN. Recording of blood pressure from the left auricle and the pulmonary veins in human subjects with interauricular septal defect. Am. J. Physiol. 150 : 267-271, 1947. 4. DEXTER, L. Atria1 septal defect. Brit. Heart J. 18 : 209-225, 1956.
5. DOUGLAS, J. E., J. C. REMBERT, W. C. SEALY, AND J. C. GREENFIELD, JR. Factors affecting shunting in experimental atria1 septal defects in dogs. Circulation Res. 24: 493-505, 1969. 6. GREENFIELD, J. C., JR. Pressure gradient technic. In: Methods in iMedical Research, edited by R. F. Rushmer. Chicago: Year Book, 1966, p. 83-93. 7. HICKHAM, J. B. Atria1 septal defect. Study of intercardiac shunts, ventricular outputs, and pulmonary pressure gradient. Am. Heart J. 38: 801-812, 1949. 8. HULL, E. The cause and effects of flow through defects of the atria1 septum. Am. Heart J. 38: 350-360, 1949. 9. LEVIN, A. R., M. S. SPACH, J. P. BOINEAU, R. V. CANENT, JR.,
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
286
10. 11.
12.
ALEXANDER M. P. CAPP, AND P. H. JEWETT. Atria1 pressure-flow dynamics in atria1 septal defects (secundum type). Circulation 37: 476-488, 1968. OSHER, H. L. The diagnostic value of intracardiac phonocardiography. Heart Bull. 16 : 32-37, 1967. SHAFFER, A. B., E. M. SILBER, AND I,. N. KATZ. Observations on the interatrial pressure gradients in man. Circulation 10 : 527-535, 1954. UHLEY, M. H. Lutembacher’s syndrome and a new concept of
13. 14.
ET
AL.
the dynamics of interatrial septal defect. Am. Heart J. 24 : 315-328, 1942. WELDON, C. S. Hemodynamics in acute atria1 septal defect. Arch. Surg. 93 : 724-729, 1966. WIGGERS, C. J. The sequence of cardiodynamic events as established by pressure pulses from the auricles, ventricles, and aortae. In : The Pressure Pulses in the Cardiovascular System, edited by C. L. Evans and A. V. Hill. New York : Longmans, Green, 1928, p. 47-64.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
OF APPLIED
Vol. 39, No. 2, August
Shunt
PHYSIOLOGY
1975.
Printed
in U.S.A.
dynamics
in experimental
atria1 septal defects JAMES A. ALEXANDER, JUDITH C. REMBERT, WILL C. SEALY, AND JOSEPH C. GREENFIELD, JR. Departments of Surgery and Medicine, Duke University Medical Center, Durham 27710; Veterans Administration Hospital, Durham, North Carolina 27705
ALEXANDER, JAMES A., JUDITH C. REMBERT, WILL AND JOSEPH C. GREENFIELD, JR. Shunt dynamics in atrial septal defects. J. Appl. Physiol. 39(Z) : 281-286.
C. SEALY, experimental
1975.-In order to study the hemodynamic variables involving the magnitude, direction, and timing of phasic shunt flow, both the interatria1 pressure gradient and blood flow along with other pertinent hemodynamic variables were measured instantaneously across a surgically created atria1 septal defect (ASD) in seven awake dogs. Atria1 and ventricular pacing and infusion of phenylephrine and isoproterenol were used to alter hemodynamic conditions. The wave form of phasic ASD flow was similar both in configuration and timing to the interatrial pressure gradient. During the cardiac cycle, both left-to-right (L-R) and right-to-left (R-L) shunting occurred: atria1 contraction augmented L-R flow; the onset of ventricular contraction was associated with R-L flow; during the latter part of ventricular contraction, flow returned to L-R with the maximum L-R shunting occurring in early diastole. Tachycardia, infusion of phenylephrine and isoproterenol did not alter the phasic flow pattern. Both spontaneous and positive pressure respiration decreased net L-R shunting. interatrial
pressure gradient;
respiration;
blood flow
A NUMBER of cardiovascular physiologists have investigated the complex hemodynamic relationships that influence the direction, timing, and magnitude of blood flow across an atria1 septal defect in both experimental animals (5, 13) and patients (Z-4, 7, 9, 11). However, a precise description of the nature of shunt flow has been limited since it has not been possible previously to measure the phasic flow directly. It is the purpose of this report to describe the contour of pulsatile flow that occurs in a surgically created atria1 septal defect and to relate the characteristics of the flow to the interatrial pressure gradient and to several hemodynamic variables that might influence flow. METHODS
Adult mongrel dogs weighing 25-30 kg were anesthetized with sodium pentobarbital (30 mg/kg), intubated, ventilated with a respirator (Mark VII, Bird Corp., Richmond, Calif.), and underwent a right thoracotomy via the fourth intercostal space. The ascending aorta was dissected from the surrounding tissue, and an electromagnetic flowmeter (EMF; TTQ-type probe, model
and
M-4000 flowmeter, Statham Instruments, Inc., Oxnard, Calif.) probe of appropriate size was fitted around the aorta as described previously ( 1). Bipolar epicardial pacing electrodes were sutured to regions near the sinus node and right ventricular apex. A purse-string suture was placed around the right atria1 appendage and a clamp positioned just below this suture. Beginning at its junction with the inferior vena cava, a 4-cm section of the right atrium was clamped and the atrium incised. Traction sutures were placed on each side of the cut atria1 edge. Heparin (2 “g/kg) was administered intravenously as a bolus. The heart emptied following occlusion of the venae cavae, and ventricular fibrillation was induced. The clamps were removed-from both the atria1 appendage and atrium. An atria1 septal defect was created, and a specially designed l.O-cm (inside diam) EMF probe (model RC1000, Micron Instruments, Inc., Los Angeles, Calif.) was sutured to the cut edges of the septum with interrupted silk sutures. The probe wires were brought out through the right atria1 appendage and the purse-string suture ‘secured. The atria were flooded with saline, then reclamped and the venae cavae released. The heart was defibrillated, and the atriotomy was closed. Total occlusion time varied from 2.5 to 3.5 min. Two identical Z-mm (inside diam) especially made catheters 30 cm long having two lateral pressure openings at the end were inserted either through the pulmonary vein from the superior segment of the right upper lobe into the left atrium or through the azygos vein into the right atrium. The tips were positioned so that they were separated by approximately 5 cm. A # 8 Lehman catheter (U.S. Catheter and Instrument Corp., Glens Falls, N. Y.) was passed into the left ventricle from the left femoral artery. Another no. 8 Lehman catheter was placed in the right ventricle from the right femoral vein. A suction tube was placed in the seventh intercostal space and under 10 cm water suction. The thoracotomy was closed. After a recovery period of 2-3 h, the animals were awake, breathing spontaneously, and having no apparent difficulties. The catheters in the left and right atria were connected to matched high-fidelity transducer (model P23Db, Statham Instruments, Inc.)-amplifier (model 350- 1500, Hewlett-Packard Co., Waltham, Mass.) systems. In order to measure the interarterial pressure gradient, the outputs of these systems were fed into an analog computer (model
281
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
ALEXANDER
282 3400, Systron-Donner Corp., Concord, Calif.) subtraction circuit and the pressure differences obtained. The recording system was statically balanced to within +O. 1 % for O-300 cmH 20. The dynamic characteristics were evaluated before and after use with a sinusoidal generator, the unbalance being no greater than 5 % through 30 Hz. These recording specifications have been shown to yield a valid pressure gradient for studies of aortic flow (6). Left and right ventricular pressures were obtained by connecting the catheters to transducers (model P23Db, Inc.). The respiratory cycle was Statham Instruments, monitored with a pneumotachygraph. Ventricular pacing was carried out by connecting the epicardial pacing electrodes to a stimulator (model SR4, Grass Instrument Co., Quincy, Mass.) that d e 1ivered square wave pulses of 3-ms duration 10 % above threshold voltage through the isolation unit. During the study, flow across the atria1 septal defect, aortic flow, left ventricular pressure, right ventricular pressure, left atria1 pressure, right atria1 pressure, respiration, the interatrial pressure gradient, and lead II of the electrocardiogram were continuously recorded on two direct-writing oscillographs (model 958- 100, series 7700, Hewlett-Packard Co.) with the first seven parameters also directly recorded on FM magnetic tape (model 3955-D, Hewlett-Packard Co.). Control measurements with the animal breathing spontaneously in normal sinus rhythm were recorded. Pacing was begun at a rate of 150 beats/min and continued for 3 min at each site. The rate was increased by approximately 30-beats/min increments in a random sequence through 240 beats/min for each atria1 and ventricular site allowing at least a 5-min interval to elapse between each new rate. Pacing was discontinued and the animal allowed to return to sinus rhythm. A lOO-pg/min infusion of phenylephrine was given intravenously in order to increase the mean arterial pressure. After data were recorded, the infusion of phenylephrine was discontinued and an infusion of isoproterenol (1 pgjmin) was begun. When all parameters returned to the initial control levels, the animal was connected to a respirator to accomplish positive pressure breathing (+ 20 cmH 20). Succinylcholine administered as a 60-mg bolus was given as needed during this portion of the study. With respirators interrupted for a 10-20-s period, the changes in septal defect flow produced by respiration could be evaluated separately. In addition, the effects produced by increasing the intrabronchial pressure to 40 cmHz0 for a 30-s period were studied. At the conclusion of the studies the animals were killed with a lethal dose of pentobarbital and the defect examined to be certain that the probe covered the orifice adequately. A low-profile ( 10 mm, ID) especially designed EMF probe (model RC-1000, Micron Instruments, Inc.) coated with liquid silicone rubber and having external Teflon sewing rings was used to measure shunt flow instantaneously. Stroke volume also was measured instantaneously with an EMF probe (TTQ-type probe, model M-4000 flowmeter, Statham Instruments, Inc.) on the ascending aorta, assuming the end of diastole as zero flow. All probes were calibrated previously both in vitro and in vivo by allowing varying known amounts of saline or blood to flow through
Zero
Zero
ET
AL.
Zero
1. Recording obtained during calibration with the probe mounted in a calibrating device and blood allowed to flow through the probe. Since there is no vessel in the lumen of the probe, this situation is analogous to that obtained during measurement of shunt flow in the atria1 septal defect. Note the exact similarity between the true zero obtained by stopping the flow and that obtained with the electrical zero. FIG.
remained in a known length of time. The calibrations within a standard deviation of =t 1.7 % for the septal defect probes and =t3.7 % for the aortic probes ; both were linear within zt2 %. The phase lag produced by the septal defect flowmeter was linear with frequency and had a phase lag of approximately 0.01 s. Since the two probes are energized differently, no electrical interference was encountered during the study. Pressure and flow data were analyzed either directly from the oscillographic recordings or from tape using a digital computer (model 1130, IBM Corp., Armonk, N. 76.) in conjunction with an analog-to-digital converter (model 663, Redcor Corp., Canoga Park, Calif.). The directional and total shunt flow was computed directly from the flowmeter sutured to the defect. Pulmonary flow (QP) was calculated from the aortic flow (QS) by adding the left-to-right shunt flow and subtracting the right-to-left shunt flow. Modeling functions of the septal defect flow were carried out with an analog computer (Simulators, Inc., Northbrook, Ill.). To measure the phasic shunt flow it was essential that the electrical zero reference be precise. Thus the flowmeter electrical zero was tested on multiple occasions in vitro during the course of routine calibration procedures by obtaining electrical zero and then by comparison to mechanical zero flow as illustrated in Fig. 1. In this situation, electrical zero and mechanical zero flow were essentially the same. The possibility that spurious flow signals might be induced by movement of the probe with the atria1 septum during cardiac contraction was studied also. The probe was placed in a container of saline and oscillated at approximately ZOO-ms intervals over a distance necessary to produce a signal of approximately the same magnitude as that found with flow in the septal defect. It was calculated that in order to produce a spurious flow signal equivalent to this peak flow signal, the probe must move approximately 15 cm/s. Thus the movement necessary to produce this magnitude of flow during systole, i.e., 200 ms, was 3.0 cm. In three dogs the EMF probe was attached to a rigid bar and sutured into the atria1 septal defect as described above. The probe was kept motionless with the bar, and no change in the contour of the flow signal was noted. Silver clips were implanted on the atria1 septum in the other dogs and their movement studied with cinefluorography. No significant movement of either the probe or clips was noted. These data substantiate the fact that movement of the atria1 septum to a degree necessary to give erroneous data is unlikely.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
EXPERIMENTAL
Erpimtion Reumotach
ASD Shunt Flow
ATRIAL
SEPTAL
DEFECTS
283
FIG. 2. Recordings of various hemodynamic parameters during the control state during spontaneous respiration. Note the marked similarity between the
F.e..----
1 Inqimttm L-R
configuration of the atria1 septal defect shunt flow and the pressure gradient. The increase in the direction of a right-to-left shunt during inspiration is obvious. The phasic contour of the shunt flow is typical of those obtained in other dogs studied.
;z;;E./ R+L
RESULTS
ECG
-F---
Complete hemodynamic studies were available from seven dogs. The contour of the phasic flow through the atria1 septal defect measured with the electromagnetic L-+R flowmeter probe was correlated with the pressure gradient. PRESSURE ‘O cm “a”p ,I RJL4,&&/‘~, GRADIENT IOcmH20 The wave form of the shunt flow was markedly similar R-+L both in configuration and timing to the pressure gradient in ail situations studied. The configuration of the flow illustrated in Fig. 2 appeared to be quite consistent in the COMPUTED SHUNT FLOW seven dogs studied. During atria1 contraction the direction of shunting was enhanced from left to right. Immediately following the onset of ventricular systole, the direction of L-tR ASD 50cm)/se;[ the shunt rapidly changed, usually resulting in a net rightto-left shunt for this period. During the latter part of SHUNT FLOW SOc,qsec A-it ventricular systole the shunt returned to a left-to-right ’ direction so that the majority of the left-to-right shunt ’ ISEC occurred in early diastole. FIG. 3. Relationship of the shunt flow computed using Eq. I The configuration of the phasic flow illustrated in Fig. 2 can be compared to the true shunt flow measured with the electromagnetic flowmeter. Note the marked similarity in both the configurawas similar regardless of whether the net shunt was righttion and timing. In A the dog was unpaced. In R, obtained during to-left or left-to-right, i.e., the configuration remained the rapid atria1 pacing, there was a marked pulsus alternans of both the same but the base line shifted. In all situations studied right and left ventricle. The shunt flow also varied with the alternans, there was a period of both left-to-right and right-to-left the decreased shunting occurring with the weak beat. Even with these marked changes in hemodynamics, the computed flow is almost shunting during the cardiac cycle. In the control state the identical to the true flow. The output filter on the electromagnetic magnitude of the peak left-to-right and right-to-left shunt flowmeter will cause a delay of approximately 0.01 s. When this is was similar and in the range of f50 cm3/s. During the taken into account, no phasic delay between the computed shunt high heart rates produced by pacing, a similar contour flow and the true shunt flow can be determined at any time during was observed except that the shunt flow during diastole the cardiac cycle. from left to right was constant with no obvious increase in flow during atria1 systole (Fig. 3). This configuration was Table 1. During the control state the QP/QS ratio had an not altered by ventricular pacing except for the absence average value of 1.55 f 0.14. During this time the preof the shunt related to atria1 systole. During both atria1 dominant shunt was left-to-right; however, a consistent and ventricular pacing no significant change in average right-to-left shunt occurred in all dogs and averaged 53 shunt flow could be demonstrated from rates of 150-240 f 8% of the left-to-right shunt. Following the marked beats/min, although there appeared to be a trend to inincrease in mean arterial blood pressure produced by crease the mean left-to-right shunt flow slightly at moderate phenylephrine, the magnitude of the left-to-right shunt heart rates. In general, high pacing rates were not associincreased so that the average QP/QS ratio increased to ated with greater peak-to-peak flows. 2.72 f 0.42. The right-to-left shunt decreased in all dogs The magnitude of the bidirectional shunt flow during and was 18 f 5 % of the left-to-right shunt. The phasic normal sinus rhythm in the control state and during incontour of the flow was unaltered at this time but appeared fusion of phenylephrine and isoproterenol are given in only to shift the entire flow in the direction of a more Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
284
ALEXANDER
TABLE
1. Bidirectional
shunt flow during normal sinus rhythm Hemodynamic
Animal No.
1 2 3 4 5 6 7 Mean &SEM 1 2 3 4 56 7
Heart Rate, beats/ min
Arterial Blood Pressure, (mean>, -Hi?
135 88 165 150 150 90 143 132 zk12
95 130 115 140 105 85 105 110 *7
135 110 150 160 150 100
125 185 230 180 250 215
Data
Cardiac ($~$ min
Mean Left-toright, cms/min
1520 1150 1050 950 1180 1160 1110 1160 =t70 B. 2400 940 660 950 1720 1380
Shunt
QRl{($os
RightLeft/ LeftRight, %
1200 160 170
1.79 1.14 1.16
24 86 72
930 1070 320 640
1.98 1.91 1.27 1.57
46 41 67 34
640 ~160
1.55 ho.14
53 zt8
2080 1190 1470 2980 1340 900
1.87 2.25 3.23 4.13 1.78 1.65
240
2480
4.15
9
340 rt80 0.01
1780 zk280 0.004
2.72 ztO.42 0.03
18 z/z5 0.006
1.56 1.07
75 12
Ri,::-tocms/min
A. Control 380 1580 1100 940 610 440 1720 790 1810 740 970 650 970 330 1250 610 A170 It90 Pheny Zej hrine 2370 290 1770 580 1490 20 3150 170 1940 600 1400 500
Net flow cms/min’
80
230
790
126 =tll NS
200 zk16 0.003
1260 =t230 NS
21
133 122
150
1630 3460
1050 1000
130 750
3
174
190
1320
50
1030
4 56
184 138 156
200 220 115
2650 2580 2860
2110 1190 810
160 530 120
1950 660 690
1.74 1.25 1.24
Mean &SEM P
7 Mean & SEM P
2720
Flow
2120 &250 0.006 C. Zso~roterenoZ
920 250 -980
12 33 1 5 31 36
0.26 8 45 15
176
180
3950
660
1020
- 360
0.91
154 zt12 NS
172 &14 0.003
2640 A350 0.008
980 zk240 NS
510 zt150 NS
450 &360 NS
1.15 =tO. 18 0.02
P values listed below B and C compare these mean values during the control state given in A. NS signifies that P > 0.05.
with
those
obtained
marked left-to-right shunt. Following isoproterenol infusion, the systemic flow increased markedly and, at the same time, the left-to-right shunt in most of the animals was at a level similar to that found during the control state. The effects on the right-to-left shunt were variable such that in five dogs the right-to-left shunt decreased below control values. However, in two dogs this flow increased rnarkedly so that the net shunt was right-to-left. During this time the phasic flow contour was not altered and the magnitude of the peak-to-peak flow appeared to be similar to that during the control conditions. Note that these drugs were not given to specifically test a drug effect but rather to change the hemodynamic conditions so that these effects on the phasic shunt flow and its relation to the \ pressure gradient could be evaluated. It was difhcult to determine precisely the effects of respiration on shunt flow because of the constant influence of intracardiac hemodynamics. In general, inspiration was accompanied by either a decrease in the magnitude of the left-to-right shunt or by an augmentation of the rightto-left shunt depending on when inspiration occurred during the cardiac cycle (Fig. 2). Spontaneous respiration decreased the net left-to-right shunt flow in the seven animals by an average of 300 cm3/min when compared to flow in the same animal when respiration was temporarily stopped. Positive pressure respiration also was associated with a decrease in left-to-right shunting and appeared to be
ET AL.
similar to that produced by spontaneous respiration. The effects of respiration were similar during all hemodynamic conditions studied. Sustained intrabronchial pressure of 40 cmHx0 reduced the mean shunt flow to* essentially zero and also markedly decreased the amplitude of the phasic flow. In attempting to evaluate the determinants of the pressure-flow characteristics across the atria1 septal defect, a model similar to that assumed to exist in the ascending assumptions, the aorta was used. With certain simplifying Navier-Stokes equations of fluid motion can be reduced to -
dP -= dZ
LZ+Rq
where z is the axial coordinate, p is the pressure, q is the instantaneous blood flow, and t is time. The terms L and R are the inertial and frictional components, respectively (6). This equation was solved for flow continuously with an analog computer using the measured pressure gradient as the forcing function (Fig. 3). These studies were carried out using the data that had been recorded on tape. The coefficients of the resistance and inertial components were modified to obtain the best empirical fit between the actual measured flow and the computed flow. The dominant factor in the computed flow was found to be the resistance term. It appeared that a linear relationship existed between the pressure gradient and the amount of flow through the orifice. The inertial component was quite small since, as can be seen in Fig. 3, there is virtually no phase shift between the observed flow and the computed flow. U
DISCUSSION
A number of hypotheses have been formulated to account for the magnitude and timing of blood flow across an atria1 septal defect (8). In anesthetized dogs, Weldon (13) found that as soon as an atria1 septal defect was created, shunting occurred from the left to the right atrium, and he attributed this finding to a differential ventricular distensibility. Dexter (4) advocated the concept that leftto-right shunting was due to the relative resistance to ventricular filling. Uhley (12) stated that the left-to-right shunt was related to the more cephalad position of the left atrium as compared to the right, presenting the gravitational factor as being the major mechanism. Cournand et al. (3) noted that different dynamic conditions in each atrium appeared to be contributing factors to the shunting mechanism. Wiggers (14) suggested that the up-and-down movements of the ventricular base produced greater fluctuations in the left atria1 pressure than in the right and probably accounted for a transient right-to-left shunt following the onset of ventricular contraction. In 1947 Cournand et al. (3) found that the mean pressure was higher in the left than the right atrium, providing a gradient for shunting the blood in a left-to-right direction. Levin et al. (9) measured simultaneous left and right atria1 pressures and computed the pressure differences in children with 1.5-3-cm diameter atria1 septal defects. Biplane cineangiography was used to evaluate timing of shunts in various phases of the cardiac cycle. They found that left-to-right shunting across the atria1 septal defect
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
EXPERIMENTAL
ATRIAL
SEPTAL
285
DEFECTS
half of ventricular occurred primarily over the latter systole with the peak occurring at the apex of the V wave during atria1 systole. The and a second accentuation pressure cgradient wave form and timing of shunting observed by Levin et al. (9) is very similar to the observations Our study shows that shunt in our animal preparation. flow through an atria1 septal defect is pulsatile in nature and follows exactly the pressure gradient generated between the two atria. The determinants of the interatrial pressure gradient may well be explained by the anatomic variation of the right atri UIII and the nat ure of th e inflow svs‘ten1 to ea ch of the atria and ventricles. The left a trium has a th .icker wall and is therefore less deformable than the right. The venous reservoirs have smaller capacity, i.e., the pulmonary veins being shorter and their diameter smaller than the superior and inferior venae cavae. The effect of the muscular development of the ventricles is different, i.e., the left ventricle being thicker and less compliant than that of the right. Thus the effective atrioventricular orifice areas, ventricular compliance, and end-diastolic pressures appear to affect the timing and duration of phasic shunting. Right-to-left shunt flow occurs at a time when the atrioventricular valves are closed and during the upstroke of the ventricular pressure tracing. This coincides with the atria1 C wave, thought to be caused by the bulging of the atrioventricular valves. The tricuspid valve, having 50-75 % more area than the mitral valve, may bulge more and may be the explanation of right atria1 pressure being higher than left atria1 pressure at this time. During the latter half of systole, corresponding to venous inflow into the atria, left atria1 pressure increases and the shunt flow direction is changed to left-to-right. The lower left atria1 and pulmonary venous compliance may well account for the shift in the pressure gradients. The diastolic phase shows flow and gradients A less compliant left ventricle in a left-to-right direction. with a smaller mitral valve orifice size and higher enddiastolic pressure is the most likely explanation for this finding. The augmentation of left-to-right flow during atria1 systole may be explained also by this difference in compliance. The magnitude of the right-to-left shunts which were measured in these animals appears to be considerably greater than that usuallv found in patients with atria1 septal defects. A major portion of thii difference may be attributable to such factors as acute experimental preparation, anesthesia, and surgical trauma. Obviously the hemodynamic status of these animals is not identical to a patient with a congenital atria1 septal defect. However,
differences in the methods used to measure the shunts also may account for some of this discrepancy. Although the QP/QS ratios obtained in this study are correct, the values reported for the shunt flows represent the total volume of blood that moves across the defect. Since the primary right-to-left shunt occurs during early systole when the mitral valve is closed, much of the blood shunted into the left atrium may be returned to the right atrium during the latter part of systole when the shunt becomes left-toright. Thus the magnitude of the right-to-left shunt measured by the flowmeter cannot be compared to right-to-left shunts detected by either indicator dilution or Fick techniques. That this is the case is supported by the data reported by Levin et al. (9) in patients with atria1 septal defects. Every one of their patients was found to have rightto-left shunting during the onset of left ventricular contraction, but none were found to have a right-to-left shunt which could be detected by indicator dilution techniques. The effects of respiration on shunting in atria1 septal defects have not been previously studied directly. Osher (10) demonstrated by phonocardiography that right atria1 murmurs were inconstant, extending from midsystole to late diastole or being confined to diastole, intensifying with expiration, and either diminishing or disappearing with inspiration, which suggested that the shunting in atria1 septal defects was influenced by respiration. In our studies, flow was influenced by respiration. Inspiration either decreased the left-to-right shunt or augmented the right-toleft shunt, depending on when it occurred during the cardiac cycle. The studies of the relation between the pressure gradient that resistance and the flow using Eq. I demonstrated was the primary determinant of the pressure gradient. It would be extremely useful if this relationship could be used to compute the shunt flow from the pressure gradient. Unfortunately, in order to accurately quantitate the precise geometry of the defect, the exact relation between the interatrial pressure catheters must be known. Thus its application to clinical studies will be extremely difficult. The authors express their appreciation to Drs. Frank Starmer and Philip McHale, computer modeling and data analysis; Mr. William Joyner and Mr. Kirby Cooper, surgical technicians; Mrs. Rosa B. Ethridge and Mrs. Brenda Haley, typists; and the Department of Medical Illustration of the Veterans Administration Hospital. This work supported in part by Grants HL-09711 and HL-01782 from the Public Health Service, and is designated Veterans Administration Program No. 3330-03. J. C. Greenfield, Jr., is recipient of Research Career Development Award 1-K3-HL-28112 from the Public Health Service. Received
for
publication
18 December
1974.
REFERENCES 1. ALEXANDER, J. A., W. C. SEALY, AND J. C. GREENFIELD, JR. Improved technique for implanting electromagnetic flowmeter probes on the coronary artery. J. A@Z. Physiol. 27: 139-140, 1969. 2. BRANNON, E. S., H. S. WEEMS, AND J. V. WARREN. Atria1 septal defect: study of hemodynamics by the techniques of right heart catheterization. Am. .I. Med. Sci. 2 10 : 489-491, 1945. 3. COURNAND, A., H. L. MOTLEY, D. D. HIMMELSTEIN, AND J. BALDWIN. Recording of blood pressure from the left auricle and the pulmonary veins in human subjects with interauricular septal defect. Am. J. Physiol. 150 : 267-271, 1947. 4. DEXTER, L. Atria1 septal defect. Brit. Heart J. 18 : 209-225, 1956.
5. DOUGLAS, J. E., J. C. REMBERT, W. C. SEALY, AND J. C. GREENFIELD, JR. Factors affecting shunting in experimental atria1 septal defects in dogs. Circulation Res. 24: 493-505, 1969. 6. GREENFIELD, J. C., JR. Pressure gradient technic. In: Methods in iMedical Research, edited by R. F. Rushmer. Chicago: Year Book, 1966, p. 83-93. 7. HICKHAM, J. B. Atria1 septal defect. Study of intercardiac shunts, ventricular outputs, and pulmonary pressure gradient. Am. Heart J. 38: 801-812, 1949. 8. HULL, E. The cause and effects of flow through defects of the atria1 septum. Am. Heart J. 38: 350-360, 1949. 9. LEVIN, A. R., M. S. SPACH, J. P. BOINEAU, R. V. CANENT, JR.,
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
286
10. 11.
12.
ALEXANDER M. P. CAPP, AND P. H. JEWETT. Atria1 pressure-flow dynamics in atria1 septal defects (secundum type). Circulation 37: 476-488, 1968. OSHER, H. L. The diagnostic value of intracardiac phonocardiography. Heart Bull. 16 : 32-37, 1967. SHAFFER, A. B., E. M. SILBER, AND I,. N. KATZ. Observations on the interatrial pressure gradients in man. Circulation 10 : 527-535, 1954. UHLEY, M. H. Lutembacher’s syndrome and a new concept of
13. 14.
ET
AL.
the dynamics of interatrial septal defect. Am. Heart J. 24 : 315-328, 1942. WELDON, C. S. Hemodynamics in acute atria1 septal defect. Arch. Surg. 93 : 724-729, 1966. WIGGERS, C. J. The sequence of cardiodynamic events as established by pressure pulses from the auricles, ventricles, and aortae. In : The Pressure Pulses in the Cardiovascular System, edited by C. L. Evans and A. V. Hill. New York : Longmans, Green, 1928, p. 47-64.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (165.190.089.176) on September 25, 2018. Copyright © 1975 American Physiological Society. All rights reserved.
