Quantification of flow through an interatrial communication

Surgery for Congenital Hearl Disease

Quantification of flow through an interatrial communication Application to the partial Fontan procedure The partial Fontan procedure has become an accepted alternative for the high-risk candidate. Creation of a small right-to-Ieft shunt will lower the systemic venous pressure and improve systemic cardiac output while maintaining an acceptable systemic arterial saturation. However, because of variations in patient size and postoperative transpulmonary gradient, proper sizing of the residual defect is difficult. We have therefore conducted a series of experiments on a model that simulates the blood flow across interatrial defects of varying sizes at several pressure gradients. We used porcine blood to develop guidelines for the sizing of the residual defect. Our results demonstrate a linear relationship between flow and pressure gradient across all hole sizes tested. In addition, there was a linear relationship between atrial septal defect size and flow at each pressure gradient. Our data show that the Gorlin formula predictions overestimated flow by 10 % to 40%. It is evident from these data that relatively small changes in the size of the atrial septal defect or in the pressure gradient result in significant changes in flow. Therefore we advocate the use of an adjustable interatrial communication such as the snare-controlled adjustable atrial septal defect for patients undergoing partial Fontan procedures. (J THoRAc CARDIOVASC SURG 1992;104:1702-8)

Jeffrey M. Pearl, MD, Hillel Laks, MD, Steven W. Barthel, BS, Elias M. Kaczer, BS, Dana K. Loo, BS, Davis C. Drinkwater, MD, and Paul Chang, BS, Los Angeles, Calif.

L e Fontan procedure is inevitably accompanied by an increase in the systemic venous pressure. If this is excessive because of an increased pulmonary vascular resistance or poor systemic ventricular function, it may result in increased mortality and morbidity rates from low cardiac output and venous congestion. We have proposed that an adjustable residual atrial communication would allow controlled right-to-left shunting, thereby lowering From the Division of Cardiothoracic Surgery, University of California at Los Angeles Medical Center, Los Angeles, Calif. Received for publication July 3, 1991. Accepted for publication Feb. 17, 1992. Address for reprints: Hillel Laks, MD, Professor and Chief, Division of Cardiothoracic Surgery, UCLA Medical Center, CHS 62-151, 10833 LeConte Ave., Los Angeles, CA 90024.

12/1/37844

1702

the systemic venous pressure and improving systemic cardiac output.': 2 A right-to-left shunt diverting as much as one third of the systemic venous return would maintain a systemic arterial oxygen saturation of 85%. The partial Fontan procedure would have particular application for high-risk patients with increased pulmonary vascular resistance. This principle has been applied clinically in the form of the snare-controlled adjustable atrial septal defect (ASD )1,3,4 and also in the form of the fenestration left in the polytetrafluoroethylene (PTFE) intraatrial baffie,* which can be closed with a catheter-mounted occlusion device.P (The adjustable ASD may be placed either in the native atrial septum or in a PTFE patch, as used in the

*Gore-Tex intraarterial baffle and patch, registered trademark of W.L. Gore & Associates, Inc., Elkton, Md.

Volume 104 Number 6 December 1992

Quantification offlow through interatrial communication

Open to air Adjustable column

I 703

Adjustable reservoir

Pressure monitors

"\

Roller pump

l

Reservoir

Fig. 1. ASD flow chamber constructed of transparent acrylic. A PTFE patch with one or more holes is placed in the center of the chamber. Bloodis suppliedat various pressuresto the chamber, and flow is collectedin a graduated cylinder.

lateral tunnel procedure. Our clinical report further describes the several techniques for adjustable ASD construction.") The size of the ASD may be critical in the clinical application of the partial Fontan principle. This is particularly true if the defect cannot be adjusted postoperatively except by its complete closure by means of a device. ASD size must be customized for each patient according to the patient's size and right-to-left atrial pressure gradient. We have conducted a series of experiments with the use of porcine blood and a model to simulating blood flow across interatrial defects of varying sizes at several pressure gradients. Absolute flow values are reported and compared with those predicted from the formula of Gorlin and Gorlin." Recommendations for sizing of the residual interatrial communication in patients undergoing a partial Fontan procedure are given according to the size of the patient and the anticipated transpulmonary gradient.

Table I. ASD flow versus gradient-single holes Gradient

Flow (mlimin)

(mmHg)

5 8

[0 12 Diameter (mm)

156 203 225 256 2

385

546

460

774

503 558

879 952

3

4

780 1165 1334 [420 5

1536 1978 2154 2354 6

Flows obtained in ml/rnin through single holes at pressure gradients of 5, 8, 10, and 12 mm Hg. The relationship is linear.

Methods and materials An experimental chamber was constructed with transparent acrylic (Fig. I). Standard bypass tubing connected one half of the chamber to a heated glass bowl whose height could be adjusted to providevarious inflow pressures,thereby simulating right atrial pressure. Blood was pumped into this bowl from a large bloodreservoir, and the levelwas maintained at a constant

17 0 4

The Journal of Thoracic and Cardiovascular Surgery

Pearl et al.

2400 2200 2000

.5 E

"0

~

:=o u:

Hole Size II

1800

•

1600 1400 1200 1000 800 600 400

•

2mm 3mm 4mm 5mm 6mm

•

...

II II 200 O-l----.------.,....--.....----.---"T--.------.------r-----. 12 8 10 5 Pressure Gradient (mmHg)

Fig. 2. Flow through single ASDs at various pressure gradients. There is a linear relationship between flow and pressure gradient.

1500 1300

:5

.§

1100

• •

v

.:::;:

900

0

li:::

0

5-1mm holes 10- Imm holes 15- Imm holes

700 500 300 5

8

10

12

Pressure Gradient (mmflg) Fig. 3. Flow through multiple I mm holes increases in a linear fashion with increasing pressure gradient as well as with increasing hole number.

height by an overflow. The outflow portion of the ASD chamber was connected to a column of bypass tubing with a Y connection. Blood reaching the top of this tubing would then overflow into a collecting vessel or back into the reservoir. By adjustment of the height of the outflow tubing, the outflow pressure or simulates left atrial pressure could be adjusted. Neither the inflow nor the outflow tubing contained valves; thus they simulated the human superior vena cava and inferior vena cava. In addition, the outflow conduit was of sufficient size to allow an unrestricted flow. Camino 110-4 pressure catheters (Camino 420 amplifier, Camino Laboratories, San Diego,

Calif.) were placed in both chambers to measure inflow and outflow pressures. Pieces ofPTFE measuring 3 X 5 cm were cut from an 18 mm PTFE tube graft, with an approximate thickness of 0.8 mm, and placed securely in a Styrofoam template. The edges were glued together to prevent leakage around the patch. A vascular punch or other circular cutting instrument of known size was used to make a hole in the middle of the PTFE patch. The template was then secured between the inflow and outflow portions of the experimental chamber, and the chamber was sealed. Blood was supplied to the circuit by exsanguination of a

Volume 104 Number 6 December 1992

Quantification offlow through interatrial communication

1 70 5

3000

'2 .... S

2000

'U

.::!~

0 .~

1000

O+------r---,------r---,--------,--------, 2

4

3

6

5

ASD Size (mm) Fig. 4. Linear relationship between ASD size and flowat a constant (10 mm Hg) pressure gradient. Small changes in ASD size results in significant changes in flow.

Table II. Comparison of rated hole size to actual holes obtained Diameter of rated hole size (mm)

Diameter of actual hole size (mm)

1

1.60 2.39 3.14 3.86 5.04 6.23

2

3 4

5 6

Table III. Flow predicted by Gorlin formula versus actual flow measured Rated hole size (mm)

Actual hole size (mm)

2

2.39

3

3.13

Some of the holes were actually oval.

Duroc support pig weighing 70 to 80 kg. The animal was fully heparinized with 40,000 units before exsanguination, and an additional 40,000 units of heparin was added to the circuit. Full heparinization was confirmed by an activated clotting time (ACT) of infinity (Hemochron, International Technidyne Corp., Edison, N.J.). The hematocrit value was adjusted to either 25% or 35% by hemodilution with crystalloid (normal saline solution) or by hemoconcentration (Hemocor Plus P650, Minntech Corp., Minneapolis, Minn.). Blood temperature was maintained at 38° C to ensure stable blood viscosity and simulate the clinical setting. Porcine blood has properties very similar to those of human blood, including similar viscosity, and served as a good model. 7 After a stable pressure and flow were obtained, flow was measured by collections in a graduated cylinder for I minute. Timed collections were repeated three times, and the average was recorded. Results were plotted asflow versus gradient, and flow versus hole size. Testing was performed with hole sizes 2, 3, 4, 5, and 6 mm in diameter at pressure gradients of 5,8, 10, and 12 mm Hg. In addition, patches with 5, 10, or 15 mm holes were tested. Flows were measured at hematocrit values of both 25% and 35% for each patch. At the conclusion of the experiment, the exact

4

3.86

5

5.04

6

6.23

Gradient (mmHg)

Percent of predicted flow

5 8 10 12 5 8 10 12 5 8 10 12 5 8 10 12 5 8 10 12

58 60 59 62 84 79 77

78 78 88 89 88 65 77

79 77

84 86 84 84

Actual flows obtained were consistently less than those predicted by the Gorlin formula when the actual hole sizes were used. This inaccuracy was particularly evident with smaller hole sizes.

diameter of each hole was determined with electronic calipers (W.L. Gore & Associates, Flagstaff, Ariz.). All animals received care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Institutes of Health (NIH Publication No. 85-23, revised 1985). Baseline lactate dehydrogenase (LDH), hematocrit, platelet

The Journal of Thoracic and Cardiovascular Surgery

1 7 0 6 Pearl et al.

1600 1400

--c E ( .) (.)

Hole Size

1200 II

1000

~ 0

5mm 5- 1mm holes

800

u:::

600 400

5

8 10 Gradient (mmHg)

12

Fig. S. A single 5 mm hole gives significantly more flow than do five I mm holes. Hole size is not additive.

Table IV. Body surface area versus one third of cardiac output Flow (mljmin)

BSA

0.3

208

0.4

277

0.5 0.6 0.7 0.8 0.9 1.0 1.2 1.4

CI

347 417

486 556 625 694 833 972 2.0

250 333 417

500 583 667

750 833 1000 1167

2.5

300 400 500 600 700 800 900 1000 1200 1400

3.0

Flow required for right-to-left shunt of one third of cardiac output at different cardiac indices. BSA, Body surface area; C/, cardiac index.

count, and free plasma hemoglobin values were determined after the circuit was primed to monitor for hemolysis, which could affect the blood viscosity. Repeat values were obtained during and at the end of the experimental period. Results Hemodynamic data. The flow rates are presented in Table I. There was a linear increase in flow with increasing pressure gradient across the ASD for all hole sizes tested (Table I, Figs. 2 and 3). In addition, there was a linear relationship between ASD size and flow at each pressure gradient (Fig. 4). It is evident from these data that relatively small changes in the size of the ASD result in significant changes in flow, Alterations of the pressure gradient also result in marked linear changes in flow, although to a lesser degree than with alterations of hole SIze. Flow through a single 5 mm hole proved to be roughly

double the flow through five individual 1 mm holes (Fig. 5). This finding documents the nonadditive nature of hole size on flow,which becomes clinically relevant for patients undergoing a fenestrated Fontan operation with multiple small fenestrations. There was no significant difference in measured flow between hematocrit values of 25% and 35% for any of the hole sizes tested (Fig. 6). The hematocrit values chosen were from measurements encountered clinically after cardiopulmonary bypass. In these experiments, a hematocrit value of 35% to 40% is similar to those observed in our patients in the early postoperative period. Perhaps at much higher hematocrit values (greater than 50%), blood viscosity would be altered, resulting in a decreased flow. The actual size of the holes placed in the PTFE patches varied slightly from the rated hole size of the punches used, and some ofthe holes had an oval shape (Table 11). The measured hole sizes were used for calculation of estimated flow according to the Godin formula, and estimated flow was compared with the experimentally measured flows (Table III). Comparison of the estimated flows with those measured directly demonstrates that the Godin formula consistently overestimates flow by 11% to 40%. Hemolysis data. After an initial increase in LDH and plasma hemoglobin above baseline, the values stabilized (267 U /L and 51.3 mg/dl, respectively). The initial increase was thought to be related to a small amount of hemolysis which in turn was related to exsanguination and extracorporeal circulation. Platelet count also remained stable throughout the experiment (322,000, 310,000, and 305,000). Microscopic examination of blood at the conclusion of the experiment revealed no sig-

Volume 104 Number 6 December 1992

Quantification offlow through interatrial communication

1 707

1000

800 ~

•

a

600

0

•

400

~

8

200

hole 2 25% hole 3 25% hole 4 25% hole 2 35% hole 3 35% hole 4 35%

==e

iii

......- -..---..,

O+-----,----r-----,---~---,---~-

5

8

10

12

GRADIENT (mmllg) Fig. 6. There is no significant difference in flow between hematocrit values of 25% and 35% at any of the hole sizes tested.

nificant hemolysis. Hence, after an initial increase in plasma hemoglobin and LDH values, continued flow through the ASD chamber and PTFE patches did not appear to result in further hemolysis over a period of 8 hours. Discussion The partial Fontan procedure has become an accepted alternative for the high-risk candidate.l? Creation of a small right-to-left shunt will lower the systemic venous pressure and improve systemic cardiac output while maintaining an acceptable systemic arterial saturation. However, a residual interatrial communication that is too large will result in excessive right-to-left shunting, leading to significant systemic arterial oxygen desaturation. Conversely, an interatrial communication that is too small will not allow sufficient shunting to achieve the desired result: reduction of systemic venous pressures and increased cardiac output. In this experiment, we have developed guidelines for the proper sizing of an interatrial communication in patients undergoing a partial Fontan procedure. The values presented are for creation of a right-to-left shunt equal to one third of the systemic venous return and take into account a range of body surface areas and various transpulmonary gradients. The difficulties associated with improper sizing of the residual interatrial communication in patients undergoing a partial Fontan procedure are demonstrated by the followingexample. If an ASD 3 mm in diameter is placed in a patient whose right atrial-left atrial pressure gradient is 5 mm Hg, a right-to-left shunt of 460 ml/rnin will be present (Table III). This would equal approximately

one third of the systemic venous return in a patient with a body surface area (BSA) of 0.4 to 0.5 m 2 and a cardiac index of 3 (cardiac output = 1200 to 1500 ml/rnin) (Table IV). If, however, a 4-mm ASD had been placed in this patient, an excessive right-to-left shunt of 774 mljmin (roughly one half of the systemic venous return) would be present and would result in significant systemic arterial desaturation to 75%. This example demonstrates the delicate balance between the size of the interatrial communication and the degree of right-to-left shunting; small size differences result in significant differences in the amount of shunting. The ability to adjust the size of the ASD with a snare-controlled device, as we have described, 1,4 allows for postoperative control of right-toleft shunting. This should improve both the safety and the effectiveness ofthe partial Fontan procedure by allowing for precise control of right-to-left shunting. As seen in Table I and Fig. 2, there is a linear relationship between flow and pressure gradient through a single fixed hole. In the clinical setting, small changes in the right atrial-left atrial pressure gradient will result in significant alteration in the degree of right-to-left shunting. Transpulmonary gradient usually decreases in the postoperative period. In a patient with a residual ASD, a decreased right atrial-left atrial pressure gradient would result in decreased right-to-left shunting. This may be appropriate if it corresponds with acceptable systemic venous pressure and good cardiac output. Conversely, a defect that provides the appropriate right-to-left shunt and saturation in the operating room may soon result in excessive right-to-left shunting in the intensive care unit; marked arterial oxygen desaturation

The Journal of Thoracic and Cardiovascular Surgery

1 7 0 8 Pearl et al.

would also occur as the patient awakens from anesthesia and as the transpulmonary gradient increases. Without the benefit of an adjustable interatrial communication, the patient would require urgent, complete closure of the interatrial communication, either surgically or by the transcatheter technique, to maintain an acceptable systemic arterial oxygen saturation. Complete closure of the interatrial communication in such a patient with elevated pulmonary vascular resistance would result in markedly increased venous pressures and low cardiac output. In contrast, if the interatrial communication were partially closed with an adjustable snare, right-to-left shunting could be decreased to obtain an acceptable systemic arterial saturation while the right atrial decompression was still maintained. Bridges and colleagues have recently presented their experience with 20 patients undergoing a partial Fontan procedure in whom a nonadjustable ASD was created. They successfully closed 11 ofthese communications with a transcatheter method in the early postoperative period before discharge of the patients and four additional communications after discharge. The interatrial communication varied from 2 to 9 mm in diameter. Some of their patients had insufficient shunting, at least initially, as a result of an interatrial communication that was too small. Other patients had marked cyanosis either from excessive right-to-left shunting or from poor ventricular function, resulting in the need for early closure. Their experience demonstrates the difficulty of predicting the appropriate size of interatrial communication necessary in patients undergoing partial Fontan procedures and the difficulties encountered when an incorrectly sized interatrial communication is used. Although the Gorlin formula may be used to predict the amount of flow through an interatrial communication of known size at a known pressure gradient, the results presented in Table II show the Gorlin formula to be inaccurate by as much as 40%. This inaccuracy is particularly evident with smaller hole sizes. Even if the appropriate interatrial communication size for the desired flowcan be predicted accurately, as from our data, variations in the postoperative pressure gradient may result in undesirable levels of shunting, requiring either adjustment or closure of the interatrial communication. The data presented are meant to serve as a guideline for the sizing of a residual interatrial communication in patients undergoing a partial Fontan procedure. Despite the limitations imposed by our model, including both the nondistensible characteristics of our chamber and the lack of pulsatile flow, we believe that the conclusions reached regarding flowdynamics and variability are validand closely represent the clinical setting. Although

these data may help determine the approximate size of ASD needed for a particular patient, differences in the postoperative transatrial pressure gradient, transpulmonary gradient, and ventricular function make it difficult to predict precisely the degree of right-to-left shunting. Our clinical experience with the snare-controlled adjustable ASD has demonstrated the benefits of a controllable right-to-left shunt in patients undergoing the Fontan procedure. The ability to control the excessive right-to-left shunt without completely eliminating it is a potential advantage of the adjustable ASD over other methods used for the partial Fontan procedure." Disclaimer The material used for the interatrial patches in this experiment was taken from an 18 mm PTFE tube graft, which is used clinically for lateral tunnel Fontan operations at our institution. Although the tube graft is the material we prefer for the lateral tunnel Fontan procedure because it is thicker and has a natural curve, W.L. Gore & Associates, Inc., does not recommend use of the tube graft in this fashion because of potential weakening of the material when it is cut through the radial fibers. Additional concerns have been raised over the possibilityof thrombus formation on the outer surface of the tube graft, which does not have the same thromboresistant properties as the inner surface. In our experience, we have not found either of these factors to be a problem. W.L. Gore & Associates, Inc., is currently developing a thicker (0.8 mm) cardiovascular patch. REFERENCES I. Laks H, Pearl J, Wu A, Haas G, George B. Experience with the fontan procedure including use of an adjustable intraatrial communication, vol 2. In: Crupi G, Parenzan L, Anderson R, et aI, eds. Perspectives in pediatric cardiology: pediatric cardiac Surgery (pt 2). Mount Kisko, New York: Futura, 1989. 2. Haas G, Laks H, Pearl J. Modified Fontan procedure. In: Advances in cardiac surgery, Volume I, Chicago: Year Book Medical Publishers, Inc. 1990. 3. Laks H. The partial Fontan procedure. Circulation, 1990;82:

1866-7.

4. Pearl JM, Laks H, Haas GS, et al. Partial Fontan: advantages of an adjustable interatrial communication. Ann Thorae Surg 1991;52:1084-95. 5. Bridges ND, Lock lE, Casteneda AR. Baffle fenestration with subsequent transcatheter closure. Circulation, 1990;82:

1681-9.

6. Gorlin R, Gorlin SG. Hydraulic formula for calculation of the area of the stenotic mitral valve, other cardiac valves, and central circulatory shunts. Am Heart J 1951;41:1-

29.

7. Grabowski EF, Didisheim P, Lewis lC, Franta ]T, Shopp lQ. Platelet adhesion to foreign surfaces under controlled conditions of whole blood flow: human vs. rabbit, dog, calf, sheep, pig, macaque and baboon. Trans Am Soc Artif Intern Organs 1977;23:141.