Long-term survivors of fetal cardiac bypass in lambs

Long-term survivors of fetal cardiac bypass in lambs The initial experience with cardiac bypass in fetal lambs resulted in early fetal death from placental insufficiency. Subsequent work in our laboratory indicated that vasoactive cyclooxygenase products were released as mediators of this response. The placental dysfunction could be blocked by the administration of indomethacin, allowing longer fetal survival. This unmasked a more subacute (but fatal) problem: fetal surgical stress resulted in diminished fetal cardiac output and progressive metabolic acidosis and contributed to the placental vasoconstriction. In acute studies, when indomethacin was given and the stress response was inhibited by the use of total spinal anesthesia, the fetus maintained normal blood gas levels, cardiac output, placental blood flow, and acid-base status for several hours after bypass. We hypothesized that beyond this point, no further fetal or placental compromise would occur and that this management technique would thus aUow long-term fetal survival. With the use of total spinal anesthesia and sterile technique for long-term study, 12 fetal lambs at 120 days (80 %) gestation underwent exposure, line placement, and cannulation for fetal cardiac bypass. Indomethacin was given intravenously on obtaining venous access. After 20 minutes of normothermic cardiac bypass at flow rates of 250 to 300 m1jkgjmin, the fetus was weaned from bypass, the cannulas and lines were removed, the uterus and abdomen were closed, and the ewe and fetus were allowed to recover. There was one maternal death (pneumonia) and one early abortion (of twins); the remaining 10 ewes progressed to term. At term, five healthy lambs that had undergone fetal cardiac bypass were delivered (including one twin), four ewes delivered a mummified study fetus and one or two healthy siblings, and one delivered a dead term fetus. With the use of techniques that inhibit fetal stress and block placental vasoconstriction, cardiac bypass can be performed in single-gestation fetal lambs with a high degree of recovery and survival (80% in this study). The cause of the elevated abortion rate associated with twin gestation is unclear. (J THORAC CARDIOVASC SURG 1994;107:14237)

Kathleen N. Fenton, MD, Hannah E. Zinn, BA, MA, Markus K. Heinemann, MD, John R. Liddicoat, MD, and Frank L. Hanley, MD, Boston, Mass., and San Francisco, Calif

Accurate echocardiographic imaging of the fetal heart and great vessels can now be done as early as 10 to 12 From the Department of Surgery, Children's Hospital, Boston, Mass., and Division of Cardiothoracic Surgery, University of California, San Francisco, Calif. Presented in part at the 1993 Meeting of the International Fetal Medicine and Surgery Society, Whistler, B.C., Canada, April 28-May 1,1993. Supported by a grant from the National Institutes of Health (NIH ROI HL 43357-01) and by a grant from the American Heart Association (13-456-901). Received for publication July 9, 1993. Accepted for publication Nov. 8, 1993. Address for reprints: Frank L. Hanley, MD, Division of Cardiothoracic Surgery, M593, 505 Parnassus Ave., University of California, San Francisco, San Francisco, CA 94143-0118. Copyright © 1994 by Mosby-Year Book, Inc. 0022-5223/94 $3.00 + 0

12/1/52%3

weeks of gestation by transvaginal echocardiography, I which opens the possibility of intrauterine repair for selected congenital cardiac defects that might benefit from this approach. Such repairs will require some form of fetal circulatory support. The early experience with cardiac bypass in fetal lambs demonstrated adequate preservation of fetal cardiac function, but placental insufficiency(as evidenced by hypercapnia and hypoxia) developed during and after bypass and ultimately caused fetal death within 30 to 90 minutes.v ' This placental dysfunction was found to be the result of markedly elevated placental vascular resistance, which led to extremely diminished placental blood flow. The vasoconstrictive response could be attenuated (but not obliterated) by the administration of high doses of sodium nitroprusside to the fetus during bypass," but this nonspecific smooth-muscle relaxation had the disadvantage of causing greatly elevated fetal pulmonary blood flow (which 1423

I 424

The Journal of Thoracic and Cardiovascular Surgery June 1994

Fenton et at.

ultimately came at the expense of fetal systemic and placental perfusion). Subsequent studies demonstrated that vasoactive cyclooxygenase products are released as mediators of the placental vasoconstrictive response; their release can be blocked by the administration of indomethacin.' When indomethacin is given to the fetus before bypass, fetal arterial oxygen and carbon dioxide tensions are maintained near normal during and for several hours after bypass, but a severe metabolic acidosis gradually develops in the fetus," which is ultimately fatal. A massive catecholamine-mediated response to surgical stress is known to develop in the fetus," 8 which raises afterload. Because the immature fetal myocardium is unable to compensate for the increased afterload, there is a substantial drop in cardiac output, which is associated with a redistribution of blood flow. Fetal stress also leads to the release of prostaglandins.U'' which are known to affect placental vascular resistance. We have previously demonstrated that the hemodynamic effects of the fetal stress response to operation can be avoided by the use of a total spinal anesthetic!" and that the use of a total spinal anesthetic for fetal cardiac bypass markedly improves fetal cardiac output, placental blood flow and gas exchange after bypass.P When this technique is used in a short-term model, the fetus is able to maintain normal cardiac output and gas exchange for up to several hours. We hypothesized that beyond this point no further placental injury would occur and that use of a total spinal anesthetic in conjunction with indomethacin administration would thus allow us to achieve long-term survival after fetal cardiac bypass.

Methods Anesthesia and monitoring. Twelve mixed-breed pregnant ewes (120 to 126 days gestation) were fasted for 24 hours. On the morning of operation, the ewe was placed in a sling, and intravenous access was obtained by cannulating the lateral tarsal vein with a large-bore catheter. After the administration of 1000 ml intravenous fluids, the ewe was sedated with sodium pentobarbital (100 to 200 mg intravenously), and tetracaine 20 mg was given as a spinal anesthetic. The ewe was then placed supine on the operating table, gently restrained, and allowed to breathe oxygenated (5 L/min by face mask) room air. Maternal sedation was maintained with sodium pentobarbital 100mg intravenously as needed. A Foley catheter was placed in the bladder to monitor urine output, and the abdomen was shaved and scrubbed. Operation. A midline laparotomy was done; the uterus was exposed and the number and orientation of the fetuses were determined. Fetal weight was estimated by palpation. A small (3 to 4 ern) hysterotomy was made over the fetal neck for the introduction of tetracaine (2 mg/kg) into the cisterna magna. A forelimb was gently extracted from the uterus, and the carotid artery and jugular vein were exposed.The carotid artery was ligated and catheterized distally with a 5F umbilical vessel

catheter. This catheter was used to monitor fetal heart rate and arterial bloodgases.The jugular veinwascannulated toward the heart with a 12F or IOF Bio-Medicus(Medtronic Bio-Medicus, Eden Prarie, Minn.) venous cannula, and indomethacin (0.5 rug/kg) was given through the venous cannula. The carotid artery was similarly cannulated with a 12F, !OF, or 8F Bio-Medicus arterial cannula. Cardiac bypass. A Bio-Medicus centrifugal pump was primed with heparinized venousblood(from adult donor sheep). The circuit included a pump and reservoir, the placenta was perfused and was used as the only oxygenator (to minimize prirningvolume, whichwas about 300 mI). Pump flows were 200 to 300 ml/kg per minute. After a 20-minute normothermic bypass period, the fetus was weaned from bypass, with volume (donor blood) given as required. The cannulas and arterial catheter were removed, and the uterus and abdomen were closed. Data collection. Fetal heart rate was monitored continuously during operation with a Statham transducer and a HewlettPackard 78342A monitor (Hewlett-Packard, Inc., Andover, Mass.). Fetal arterial bloodgases and hematocrit were sampled before and after neck cannulation, at 10 and 20 minutes during bypass, and after cessation of bypass and were analyzed on a Ciba-Corning 280 bloodgas system (Ciba Corning Diagnostics Corp., Medfield, Mass.). All samples for analysis of fetal blood gases were drawn from the carotid artery. Postoperative course. The ewe was observed upright in her sling and given intravenous fluids for several hours or until she was alert and eating, at which time she was returned to her pen. Lambs were allowed to progress to term, at which time they were delivered naturally. After examination by the investigators, the lambs were returned to the farm with their mothers to feed and grow. Euthanasia. After an observation period of I week to 3 months, the lambs were put to death by ketamine sedation followed by an overdose of KCI. Autopsies were done to confirm previousligation of the neck vessels (that is, to verify prior operation) and to screen for any pathologic conditions. All animals received humane 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 Academy of Sciences and published by the National Institutes of Health. The experimental protocol was reviewed and approved by the Children's Hospital of Boston Committee on the Care and Use of Laboratory Animals. Statistical analysis. Blood gas data were analyzed with a statistical program (Statview SE, Abacus Concepts, Berkeley, Calif.), using one-way analysis of variance to look for factors predictive of survival and repeated-measures analysis of variance to look at changes in pH, carbon dioxide tension (Pco-), and oxygen tension (P02) over time.

Results Fetal heart rate and arterial blood gas levels were normal before bypass in all lambs (Table I). There was a trend toward a small rise in mean Pco- over time; mean values for pH and POz and oxygen saturation remained normal (Table I). Long-term results are summarized in Table II. One

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 6

Fenton et al.

I 425

Table I. Fetal arterial blood gas data (intraoperative blood gas data from all lambs undergoing bypass) Time

pH

Before cannulation After cannulation 10 minutes of bypass 20 minutes of bypass After bypass

7.348 7.327 7.277 7.303 7.279

± ± ± ± ±

Pe02

0.047 0.035 0.063 0.055 0.092

(mmHg)

42.7 41.5 45.7 44.5 47.6

± ± ± ± ±

8.0 6.5 8.4 8.3 11

P02

(mmHg)

23.2 24.4 24.9 24.8 22.4

± ± ± ± ±

5.2 6.8 7.4 8.4 6.7

Sa 0 2 (%) 17.9 10.9 19.5 20.7 19.1

± ± ± ± ±

18.5 14.6 20.3 21.3 20.5

Data are given as mean plus or minus the standard deviation. Sa02, Arterial oxygen saturation.

Table II. Long-term outcome for each ewe and her fetuses No. offetuses

Outcome of FB lamb

1

1

2 3 4

2

Premature birth; ultimately did well Early abortion Maternal death Born alive at term Aborted at term Born alive at term Aborted at term (7 kg) Born alive at term Aborted at term Born alive at term Aborted at term Aborted at term

Ewe

5 6 7 8 9 10 11 12

2

I 3

1 1 2

2 I 2 2

Outcome of sibling(s) Early abortion Maternal death Born alive at term

Born alive at term Born alive at term Born alive at term Born alive at term

FB lamb, Lamb undergoing fetal cardiac bypass.

ewe (carrying twins) died on the first postoperative day of hypoxemia and metabolic acidosis, which on autopsy was found to be related to a severe pneumonia. Both of the fetuses were dead at the time of autopsy. There was also one early (postoperative day 3) abortion of twins. One ewe delivered a live lamb 7 days before term; the lamb was small and weak and required gavage feeding for 2 days, but ultimately grew normally. The remaining 9 ewes progressed to term. At term, of five single-gestation lambs, four were born alive. The remaining lamb was dead at delivery, but weighed approximately three times its weight at operation and appeared to have recently died. Autopsy revealed a massive left-ventricular infarct and a closedductus arteriosus. Offivemultiple-gestation lambs, one study lamb was born alive; the remaining lambs had living siblings, but appeared to have died shortly after operation. All of the live-born lambs appeared healthy and behaved and grew normally. The slight rise in Pco- during operation was not related to fetal death. Instead, a drop in fetal P02 was strongly predictive offetal death (p < 0.005): mean fetal P02 after bypass was 26.4 ± 3.3 mm Hg for survivors but only 13.4 ± 5.8 mm Hg for nonsurvivors (Fig. I). Twin gestation was also statistically related to fetal death (p < 0.005).

Discussion Preliminary investigation of the feasibility of performing fetal cardiac bypass in sheep demonstrated a good hemodynamic recovery of the fetus after bypass.l? but placental dysfunction developed during and after bypass and ultimately caused fetal death within hours. The placental insufficiencyis the result of elevated placental vascular resistance, which leads to diminished placental blood flow and poor gas exchange. Sodium nitroprusside can be used to block placental vasoconstriction, but the applicability of this technique is limited by a significant rise in pulmonary blood flow. The stimulus for the vasoconstrictive response of the placenta to bypass has not been completely elucidated, but it is clear that it is partly related to fetal stress'S 15 and partially related to exposure of blood to the extracorporeal circuit.l" We have previously demonstrated that administration of indomethacin to the fetus before bypass at least partially blocks the placental vasoconstriction, implying a cyclooxygenase-mediated mechanism. In short-term studies, by inhibiting prostaglandin synthesis with indomethacin and blocking the fetal stress response with a total spinal anesthetic, we have been able to maintain normal fetal gas exchange for up to 6 hours after bypass. The goal of the current study was to apply this technique to a long-term fetal lamb

I 426

The Journal of Thoracic and Cardiovascular Surgery June 1994

Fenton et al.

40 -e--dead 0;

:x:

----alive

30

E

E-

N

0

20

Q.

<;;

s

1ii

10

0

pre bypass

bypass

post bypass

Fig. 1. Intraoperative fetal arterial Po-, stratified by longterm outcome.Error bars representstandard deviation. Drop in fetal Poz was strongly predictiveof fetal death (p < 0.005) by analysis of variance. model, to see whether it would permit long-term fetal survival. Of 12 fetal lambs undergoing bypass, 4 of 5 single lambs and 1 of 7 twins survived. The overall survival of 42% represents the first reported long-term success of fetal cardiac bypass, and the 80% survival in single-gestation fetal lambs is promising. The one lamb that appeared to have died close to term clearly grew normally after operation, and apparently died of a left ventricular infarct. This may represent a complication of (incorrect) cannula placement at the time of operation. We have not previously seen this complication in our short-term studies. The elevated abortion rate in twins is of greater concern. With useofthe technique described in this study, there is typically a drop in placental blood flow of 35% to 50% in most lambs when blood flows are examined in acute preparations. Although normal gas exchange is maintained in the short term, it is possible that twin fetuses have less placental reserve and that over time this amount of reduced placental flow does have notable consequences. A drop in placental blood flow, however, usually causes hypercapnia and respiratory acidosis long before hypoxemia develops. Hypoxemia alone is more commonly associated with poor uteroplacental perfusion or inadequate oxygen delivery to the maternal side of the placenta. The association of death with early hypoxemia implies insufficient maternal blood flow to the placenta; this could easily be mechanistically related to multiple gestation. If this were combined with a moderate drop in fetal placental blood flow after bypass, and with replacement of some fetal hemoglobin by adult hemoglobin (because of the pump prime), the increased incidence of fetal death found in the twin fetuses could be explained.

Cardiac bypass can be done successfully in single-gestation fetal lambs by blocking the fetal stress response to operation (with the use of total spinal anesthesia) and by inhibiting prostaglandin synthesis (with indomethacin). A number of other problems remain to be solved before the application of these techniques to human beings. Perhaps foremost is the problem of the pump prime. We have minimized our circuit to limit priming volume, but there is still some replacement of fetal hemoglobin by adult hemoglobin. The addition of other lines (for example, suction lines) and oxygenators would need to be considered carefully in light of the increased priming volume required. This will become more of a problem when repairs are extended to younger (smaller) fetuses. Early fetal operations will probably involve repair of relatively simple flow-altering defects, which can be done with continuous normothermic bypass. If more complex intracardiac repairs are eventually attempted that require hypothermic bypass, cardioplegia, and, potentially, circulatory arrest, the specific effects ofthese on the fetus (and on the mother) will need to be studied carefully. Further research is needed to preserve normal placental blood flow, to optimize uterine perfusion, and to apply these techniques to a primate model.

REFERENCES 1. Dolkart LA, Reimers FT. Transvaginalfetal echocardiography in early pregnancy: normative data. Am J Obstet GynecoI1991;165:688-91. 2. Richter RC, Slate RK, RudolphAM, Turley K. Fetalblood flow during hypothermiccardiopulmonarybypassin utero. [Abstract]. J CardiovascSurg I985;26(Suppl):86. 3. Hawkins JA, Paape K, Adkins TP, Shaddy RE, Gay WA. Extracorporeal circulation in the fetal lamb: effects of hypothermia and perfusion rate. J CardiovascSurg 1991; 32:295-300. 4. Bradley SM, Hanley FL, Jennings RW, Duncan BW, Jester JA, Verrier ED. Regional blood flows during cardiopulmonarybypassin fetal lambs:the effectof nitroprusside [Abstract]. Circulation 1990;82(Suppl):III413. 5. Sabik JF, Assad RS, Hanley FL. Prostaglandin synthesis inhibition prevents placental dysfunction after fetal cardiac bypass.J THORAC CARDIOVASC SURG 1992;103:733-42. 6. Fenton KN, Heinemann MK, Hanley FL. Exclusion of the placenta during fetal cardiac bypass augments systemic perfusion and provides important information about the mechanism of placental injury. J THORAC CARDIOVASC SURG 1993;105:502-12. 7. Heymann MA, Rudolph AM. Effect of exteriorization of the sheep fetus on its cardiovascular function. Circ Res 1967;21:741-5. 8. Cohn HE, Sacks EJ, Heymann MA, Rudolph AM. Car-

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 6

9.

10.

II.

12.

diovascular responses to hypoxemia and acidemia in fetal lambs. Am J Obstet GynecoI1974;120:817-24, Hooper SB, Coulter CL, Deayton JM, Harding R, Thorburn GD. Fetal endocrine responses to prolonged hypoxemia in sheep. Am J Physiol 1990;259:R703-8. Challis JR, Hart I, Louis TM, et al. Prostaglandins in the sheep fetus: implications for fetal function. Adv Prostaglandin Thromboxane Res 1978;4:115-32, Andrianakis P, Walker DW, Ralph MM, Thorburn GD. Effects of hyperthermia on fetal and maternal prostaglandins and uterine activity in sheep. Prostaglandins 1989; 38:541-55. O'Brien WF, Cefalo RC, Davis SE, Ramwell PW. Maternal and fetal prostaglandins during acute fetal acidosis. Eur J Obstet Gynecol Reprad Bioi 1983;14:317-21.

Fenton et al.

I 427

13. Clyman RI, Mauray F, Roman C, Rudolph AM, Heymann MA. Circulating prostaglandin E2 concentrations and patent ductus arteriosus in fetal and neonatal Iambs. J Pediatr 1980;97:455-61. 14. Fenton KN, Heinemann MK, Hickey PR, Hanley FL. The stress response during fetal surgery is blocked by total spinal anesthesia. Surg Forum 1992;43:631-4. 15. Fenton KN, Heinemann MK, Hickey PR, Klautz RJM, Liddicoat JR, Hanley FL. Inhibition of the fetal stress response improves cardiac output and gas exchange after fetal cardiac bypass. J THORAC CARDIOVASC SURG 1994; 107:1416-22. 16. Fenton KN, Heinemann MK, Hanley FL. Fetal cardiac bypass using cross-circulation. Cor Europeum [In press].