F that many congenital cardiac anomalies can be detected

Transcription

1 Fetal Cardiac Bypass: Improved Placental Function With Moderately High Flow Rates John A. Hawkins, MD, Steven M. Clark, MD, Robert E. Shaddy, MD, and William A. Gay, Jr, MD Divisions of Cardiothoracic Surgery and Cardiology, The University of Utah School of Medicine and Primary Children s Medical Center, Salt Lake City, Utah, and Division of Pediatric Cardiothoracic Surgery, University of Cincinnati School of Medicine and Children s Hospital Medical Center, Cincinnati, Ohio Prenatal correction of certain cardiac lesions with a poor prognosis may have advantages over postnatal repair. For this to be done, safe and effective support of the fetal circulation must be devised. Studies involving fetal cardiac bypass have demonstrated progressive fetal hypoxemia, hypercapnia, and acidosis, indicating placental dysfunction. We performed fetal cardiac bypass in 18 fetal lambs (126 to 140 days gestation) to assess the effect of flow rate on fetal oxygenation and metabolism and function of the placenta as an in vivo oxygenator. Fetal cardiac bypass was done for a 30-minute study period at normothermia in all fetuses. During the study period the fetal aorta was cross-clamped and cold cardioplegia was administered to the heart so there was no fetal cardiac contribution to systemic output. Nine fetuses underwent studies at low flow rates (109 f 20 ml * kg- * min- ) and 9 at higher flow rates ( ml * kg- min- 1. At the lower flow rate, mean aortic pressure, arterial ph, and oxygen tension decreased whereas carbon dioxide tension and lactate levels increased when compared with prebypass levels. At the higher flow rate mean aortic pressure, ph, oxygen tension, carbon dioxide tension, and lactate levels remained similar to prebypass levels during the 30-minute study period. When the animals were weaned from the bypass circuit after studies at high flow rates, arterial oxygen tension and ph decreased whereas carbon dioxide tension increased to levels similar to those in the low-flow group. We conclude that low fetal cardiac bypass flow rates (100 to 125 ml * kg- * min- ) are inadequate to maintain hemodynamics, oxygenation, CO, removal, and normal lactate levels when the placenta is used as an in vivo oxygenator. Higher flow rates (300 to 400 ml * kg-. min- ) may limit these changes by improving placental perfusion and function during bypass. Despite high flow rates, placental dysfunction and fetal blood gas abnormalities still occur after fetal cardiac bypass. (Ann Thorac Surg 1994;57:293-7) eta1 echocardiography has now developed to the point F that many congenital cardiac anomalies can be detected relatively early in gestation [l]. Some of these cardiac abnormalities recognized during fetal development by echocardiography are complex anomalies and have a higher rate of intrauterine fetal demise or early neonatal death [l, 21. These observations have led some to consider intrauterine surgical intervention for certain cardiac conditions [>5]. Before prenatal cardiac surgical For editorial comment, see page 279. intervention can be investigated, methods for intrauterine extracorporeal circulation and myocardial preservation must be perfected. Previous studies in sheep have shown a consistent dysfunction of the placenta and increase in placental vascular resistance both during and after fetal bypass [6, 71. This dysfunction is characterized by acidosis of the Presented at the Twenty-Ninth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 2527, Address reprint requests to Dr Hawkins, Department of Surgery, The University of Utah Medical Center, 50 North Medical Dr, Salt Lake City, UT fetus and impairment of transplacental gas exchange. Previous studies from this laboratory have shown that this detrimental effect is worsened by the use of hypothermia and bypass flow rates that have been traditionally used in the human neonate [8]. This study was designed to examine the role that higher fetal bypass flow rates may have on the improvement in fetal gas exchange and placental function. Material and Methods Animals We studied a total of 18 fetuses from pregnant ewes of mixed western breed. Breeding dates were known, and gestational ages ranged from 126 to 140 days gestation (term = 145 days). Weights of the fetuses ranged from 2.3 to 5.3 kg. All animals were treated humanely in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication 85-23, revised 1985). The protocol for the experiments was reviewed and approved by the respective committees for animal care at Children s Hospital Medical Center in Cincinnati, OH, and the University of Utah, Salt Lake City, UT by The Society of Thoracic Surgeons /94/$7.00

2 294 HAWKINS ET AL Ann Thorac Surg Pressure (mm Hg) 35 15, Fig 1. Fetal mean arterial blood pressure for the both the low- and high-flow groups. Mean arterial pressure in the low-flow group was significantly less than that in the high-flow group during bypass. Mean arterial pressure was similar in the two groups both before and after bypass. ("p < 0.05, high-flow versus low-flow group.) Surgical Preparation We performed all studies under acute experimental conditions. Each maternal ewe was fasted for 24 hours before the operative procedure. The maternal ewes underwent endotracheal general anesthesia with isoflurane and 100% oxygen with controlled ventilation. Adequacy of maternal ventilation and oxygenation was determined by blood gas analysis, and the ventilator was adjusted appropriately. The ewe then had 200 ml of blood drawn from the jugular vein for use in priming the pump circuit. The fetus was exposed through a midline maternal laparotomy and a vertical hysterotomy directly overlying the fetal sternum. The fetus was not removed from the uterus to minimize fetal manipulation and temperature loss. An arterial line was placed into a fetal axillary artery and was used for monitoring mean and phasic aortic pressure. The fetal heart was exposed through a median sternotomy, and cannulation of the pulmonary trunk was used for arterial inflow and the right atrium for venous uptake [l]. Typically a 10F cannula (Electro-Catheter Corp, Rahway, NJ) was used for the arterial cannulation and a single 16F metal-tipped cannula was used for venous uptake (DLP Corp, Grand Rapids, MI). The bypass circuit consisted of a venous reservoir, a low-prime heat exchanger (Avecor Cardiovascular Inc, Plymouth, MN), and a calibrated roller pump. The placenta functioned as an in vivo oxygenator through umbilical artery perfusion from the descending aorta. The exact details of this perfusion circuit and schematic representation have been published previously [8]. Approximately 200 ml of the heparinized maternal blood was used to prime the pump circuit. For the 30-minute study period the ascending aorta was cross-clamped. The heart was arrested using 50 ml of 4 C crystalloid potassium cardioplegia to have no cardiac contribution to total body perfusion during the studies. Experimental Studies The 18 fetuses were divided into a control group of 9 fetuses that underwent perfusion at a low flow rate (mean k standard error of the mean, ml - kg-' * min-') and a study group of 9 fetuses that underwent perfusion at a high flow rate (324 * 31 ml * kg-' * min-i). This higher flow rate was essentially at the upper limits of flow for the size of cannulas that were used in these experiments. All studies were carried out at normothermia for the sheep fetus (37" to 39 C). The perfusion period lasted for 30 minutes of cross-clamp time, and study measurements were taken before fetal sternotomy and at 10 minutes and 30 minutes after institution of fetal bypass. These measurements included measurement of mean arterial pressure, arterial blood gases, and serum lactate level. At the end of the 30-minute study period, the aorta was unclamped and perfusion of the coronary arteries and heart was reinstituted. The fetus was then weaned from bypass and additional hemodynamic, blood gas, and lactate measurements were taken 15 minutes later. At the conclusion of the acute experimental studies, the ewe and fetus were sacrificed using pentobarbital and potassium chloride injection. The fetus was removed from the uterus and weighed to obtain exact weights for normalization of the high and low bypass perfusion flow rates. Statistical Analysis All values are expressed as mean * standard error of the mean. Significant differences between measurements in the high-flow group and the low-flow group were tested with the use of Students's t test with Bonferroni correction; differences were considered to be significant if the value of p was less than Results All data are shown in Figures 1 through 5. Baseline values for blood pressure, blood gases, and lactate levels are similar in the two study groups and were similar to values found in other studies using similar experimental conditions [8-11]. Mean arterial pressure decreased markedly in the animals undergoing bypass with the lower flow rates, but returned to normal once bypass was discontinued (see Fig 1). The arterial ph in the high-flow group was maintained at baseline levels throughout the study and was significantly higher than in the low-flow group :I - ph HghFbw C LowFbw ' I Fig 2. Fetal arterial ph for both the low- and high-flow groups. The ph in the low-flow group was significantly less than that in the highflow group after 30 minutes of bypass and also after bypass was discontinued. ("p < 0.05, high-flow versus low-flow group.)

3 Ann Thorac Surg HAWKINS ET AL pc02 (tom) T High Flow Low Flow 40 1 Fig 3. Fetal carbon dioxide tension (pc0,) for both the low- and high-flow groups. The fetal pc0, in the high-flow group is significantly less than that in the low-flow group after 30 minutes of bypass. After bypass is discontinued the pc0, is still less in the high-flow group, but not significantly SO. (*p < 0.05, high-flow versus h-flow group.) Lactate 400- (pmwl) High Flow Low Flow Fig 5. Fetal serum lactate levels for both high- and low-flow groups. The Serum lactate level is significantly elevated in the l0w-f ~ group beginning 10 minutes after the institution of bypass and remaining very elevated after bypass is discontinued. (*p < 0.05, high-flow versus low-flow group.) after 30 minutes of bypass and after bypass was discontinued (see Fig 2). Arterial carbon dioxide tension (pc0,) in the low-flow group was markedly increased over that in the high flow group after 30 minutes, yet seemed to improve slightly after bypass was discontinued (see Fig 3). Similarly the arterial oxygen tension (PO,) in the low-flow group was significantly lower than PO, in the high-flow group 30 minutes after instituting bypass, but did improve again after bypass was discontinued (see Fig 4). The most striking differences were seen in the serum lactate levels (see Fig 5), with serum lactate levels in the low-flow group significantly elevated over those in the high-flow group as early as 10 minutes after the start of bypass (see Fig 5). Five fetuses in the high-flow study group were successfully weaned from the bypass circuit and survived at least 15 minutes afterward. Four fetuses from the low-flow group could be successfully weaned from bypass and survived 15 minutes. The inability to be weaned from bypass was uniformly due to insufficient cardiac activity and function. Comment Developments in fetal echocardiography and the advancements in treatment of certain congenital defects by 30 7 pre-bypass 10 rnin 30 min post-bypass Fig 4. Fetal oxygen tension (PO,) for both the low- and high-flow groups. The fetal PO, in the low-flow group is significantly less than the PO, in the high-flow group after 30 minutes of bypass. After bypass is discontinued, the PO, values in the high- and low-flow groups are similar. ( p < 0.05, high-flow versus low-flow group.) in-utero operation has led to early attempts to perform in-utero therapy for cardiac abnormalities in humans including congenital heart block and critical aortic stenosis [4, 51. In addition to this human work, several investigators have begun work on basic research into cardiac operations in fetal lambs using closed-heart techniques to relieve both pulmonary artery stenosis and aortic stenosis [3]. Despite the importance of these early attempts, the greatest impact of surgical intervention for the treatment of prenatal cardiac conditions will likely require bypass techniques to be developed so more complex and lethal conditions can be treated. Several investigators have begun the experimental work neckssary to develop effective fetal bypass techniques [&lo]. Early work has demonstrated a consistent problem with placental gas transfer and acidosis when the fetal circulation is supported extracorporeally and the placenta functions as an oxygenator [6, 8, 91. This placental dysfunction seems to occur at normothermic temperatures and even more so at hypothermic temperatures [8]. This phenomenon has been demonstrated to be due to an increase in placental vascular resistance, even when perfusion occurs at normothermic temperatures [7]. The earliest attempts to counteract this increase in placental vascular resistance used an extracorporeal oxygenator, nitroprusside, and flow rates that approximated those in human neonatal cardiac operations and averaged about 150 ml - kg-i - min- [9]. These experiments were still characterized by elevated pc0, levels and progressively worsening acidosis of the fetus after bypass was discontinued [9]. For this reason and the fact that normal biventricular output in the fetus is 400 to 450 ml - kg- * min- [12], we hypothesized that higher flow rates may improve placental function by more closely approximating normal systemic and placental blood flow. These experiments did demonstrate maintenance or slight improvement in pco,, ph, PO,, and lactate levels during the time the fetus was supported at higher flow rates by the bypass circuit, as compared with baseline values (see Figs 2-5). Elevation in pc0, and lactate levels with decrease in ph and PO, was seen at the lower flow rates as compared with baseline levels. It can be hypothesized that relative hypoperfusion of the placenta and

4 296 HAWKINS ET AL Ann Thorac Surg fetus is responsible for these changes. However, at the higher flow rates, all of these values worsened when the fetus was weaned from bypass. Although we did not measure postbypass cardiac output, these findings can be partially explained by presumed myocardial dysfunction once bypass was discontinued because of the 30-minute cross-clamp period. However this decrease in ph and PO, with an increase in pc0, is still seen in the fetal lamb model if the aorta is not clamped and the heart is allowed to beat during the bypass period [9]. This indicates that the postbypass acidemia and hypercapnia may have as much to do with placental dysfunction as myocardial dysfunction [ 131. This placental dysfunction after fetal cardiac bypass has been extensively studied by Hanley and his associates [7, 13, 141. They have demonstrated a consistent increase in placental vascular resistance and corresponding decrease in placental blood flow, which becomes even more marked in the postbypass period [7]. They have seen results very similar to those in the current study in that PO,, pco,, and ph are maintained at near-baseline levels during bypass at flow rates of 400 ml - kg-' * min-', but still deteriorate after bypass [13]. They have hypothesized that this is due to fetal stress and the production of prostaglandins [13]. The administration of indomethacin or corticosteroids seems to blunt this response and return placental vascular resistance and blood flow to more normal levels [13, 141. The current experiments were characterized by a relatively low rate of successful weaning from the bypass circuit, as compared with other studies [9, 131. Our experimental design differed from other experiments in that the aorta was cross-clamped during the 30-minute study period and cardioplegia was used. We have noticed that when the aorta is left unclamped at normothermic temperatures, there continues to be cardiac activity with a small amount of pulsatile flow, particularly at lower flow rates. We wanted the study to more closely approximate conditions that might be in use if an open fetal cardiac operation were to be done. Also, we did not want an additional, unknown amount of cardiac output to be delivered to the placenta, making true differences between flow rates more difficult to interpret. It is likely that clamping the aorta and administering cardioplegia significantly reduced cardiac function in the postbypass phase and contributed to the low success rate of weaning from the bypass circuit. It is also likely that normothermia and unavoidable ventricular distention during fetal cardiac bypass contribute to postbypass myocardial dysfunction. In summary, these experiments have demonstrated that fetal bypass flow rates that more normally approximate fetal biventricular output improve fetal blood gases, lactate levels, and acid-base balance during bypass as compared with the lower flow rates that have been used in previous experiments. Despite near-normalization of fetal blood flow and blood gases during bypass, placental dysfunction still occurs after the bypass period. A clearer understanding of the mechanism responsible for placental dysfunction during fetal cardiac bypass will be needed to further refine this technique for future experimental applications. Supported by a grant from the Ohio Affiliate of the American Heart Association. References 1. Allan LD, Grawfore DC, Anderson RH, Tynan M. Spectrum of congenital heart disease detected echocardiographically in prenatal life. Br Heart J 1985;54: Huhta JC. Uses and abuses of fetal echocardiography. A pediatric cardiologist's view. J Am Coll Cardiol 1986;8: Turley K, Vlahakes GJ, Harrision MR, et al. Intrauterine cardiothoracic surgery. The fetal lamb model. Ann Thorac Surg 1982; Carpenter RJ, Strasburger JF, Garson A, Smith RT, Deter RL, Engelhardt HT. Fetal ventricular pacing for hydrops secondary to complete atrioventricular block. J Am Coll Cardiol 1986;8: Maxwell D, Allan L, Tynan MJ. Balloon dilation of the aortic valve in the fetus. A report of 2 cases. Br Heart J 1991;65: Richter RC, Slate RK, Rudolph AM, Turley K. Fetal blood flow during hypothermic cardiopulmonary bypass in-utero. J Cardiovasc Surg 1985;26: Lee FY, Assad RS, OHare RE, Hanley FL. Cardiopulmonary bypass in the isolated in-situ lamb placenta. Hemodynamic characteristics. Circulation 1990;82(Suppl 3): Hawkins JA, Paape KL, Adkins TP, Shaddy RE, Gay WA. Extracorporeal circulation in the fetal lamb. Effects of hypothermia and perfusion rate. J Cardiovasc Surg 1991;32: Bradley SM, Verrier ED, Duncan BW, et al. Cardiopulmonary bypass in the fetal lamb. Effect of sodium nitroprusside. Circulation 1989;8O(Suppl 2): Slate RK, Richter RC, Rudolph AM, Turley K. Cardiopulmonary bypass in fetal lambs. A technique for intrauterine cardiac surgery. Circulation 1984;7O(Suppl2): Fisher DJ, Heymann MA, Rudolph AM. Fetal myocardial oxygen consumption and carbohydrate consumption during acutely induced hypoxemia. Am J Physiol 1982;242(Heart Circ Physiol 11):H Rudolph AM. Distribution and regulation of blood flow in the fetal and neonatal lamb. Circ Res 1985; Sabik JF, Assad RS, Hanley FL. Prostaglandin synthesis inhibition prevents placental dysfunction after fetal cardiac bypass. J Thorac Cardiovasc Surg 1992;103: Sabik JF, Hanley FL, Heinemann MK, Assad RS. High dose steroids prevent placental dysfunction after fetal cardiac bypass. J Thorac Cardiovasc Surg (in press). DISCUSSION DR MARKUS K. HEINEMANN (Hannover, Germany): I was quite surprised by your choice and definition of the flow rates. I think if you choose 100 and 200 ml kg-'. min-', it is pretty obvious (if you take into account that about 40% of the combined ventricular output of a fetus goes into the placenta) that 100 ml 3 kg-'. min-' will not work, offhand. I think what your

5 Ann Thorac Surg HAWKINS ET AL 297 figures show is that you inflict placental injury by inducing hypoperfusion. If you talk about high flow, why didn't you compare 200 ml * kg-'. min-' with 400 ml * kg-' min-'? Four hundred or 500 ml * kg-'. min-' would be real high flow versus 200 ml. kg-'. min-', which is just about normal, or even less than that. DR HAWKINS: Admittedly some of the experimental design in this study is somewhat dated. Initially I performed the first set of experiments a number of years ago and chose at that time, as did several other investigators, a flow rate that approximated what we use in a normal human neonate for bypass. In the early experiments we did not understand that when using the placenta as an oxygenator, it becomes a circuit that is in parallel rather than one in series, which is typically used on our bypass machines for human use. The reason this particular "high' flow was chosen was that it was essentially the upper limit of flow that could be attained with the cannulas that were used. Admittedly this was somewhat arbitrary. DR HEINEMANN: My second question is about cardioplegia, which is a thing that we in Boston never used. Did you take the chance to investigate the effects of cardioplegia on a totally immature heart, like measuring cardiac output or histologic studies? I have my doubts that it is good for this kind of tissue. DR HAWKINS: No, I did not. I chose to give cardioplegia because I observed at normothermia that there was still some pulsatile flow, at least at the flow rates that I used in the circuit. I was not sure what beneficial effect pulsatile flow may have, which would make it more difficult to sort out what effect the bypass had. I have not looked at cardiac function after cardioplegic arrest. I can only anecdotally say that the fetal heart does not tolerate cardioplegia well. The mortality rate in this series was about 50% in terms of weaning from bypass, and uniformly it was because the heart just did not work. And so there is a long way to go on this. Doctor Hanley and Dr Heinemann have done some very elegant experiments looking at prostaglandins and their role. I think a bigger problem is going to be myocardial preservation in the fetus. DR JEFFREY M. DUNN (Philadelphia, PA): I have a quick comment. First of all, it is interesting that we are increasingly getting an armamentarium of ways to maintain placental blood flow. As an aside, our experiments have used flow rates of 150 to 200 ml * kg-'. min-' and they mirror your low-flow group completely. In your high-flow group your last measurement was 30 minutes after bypass, and I noticed that the CO, tension was a little higher than control there. In our experience, the C02 tension has been the first sign of placental dysfunction, and in studies that we did both in Philadelphia and in Reading with Dr Hanson's laboratory, we found that even 2 or 3 hours after bypass, after the uterus was closed and the animal was back to baseline, the CO, tension could sometimes start to increase showing placental dysfunction and the same result. Do you have any late studies showing what happens after that 30-minute period? DR HAWKINS: I am sorry, I do not have any results with that. I would have to agree with you that the changes that we see early on become even more marked and more pronounced as time goes on. There is a lot we do not understand, in fact, much more we do not understand than we do understand about placental function and fetal cardiac bypass. DR FRANK L. HANLEY (San Francisco, CA): Our laboratory has an interest in this area. It appears that there are two major types of placental dysfunction after bypass. The first is addressed in this paper, ie, the hemodynamic component. Work from our laboratory with an isolated placental model has shown that if the mean umbilical perfusion pressure to the placenta decreases to less than 40 mm Hg, a detrimental positive feedback loop occurs. As the pressure decreases to less than 40 mm Hg, the umbilicalplacental vascular resistance increases, instead of decreasing as you should expect with an ordinary autoregulatory mechanism. The best explanation for this behavior is the following. The umbilical-placental vasculature is maximally vasodilated in its normal state; therefore, there is little potential for further vasodilation. Furthermore, the placental blood flow from the fetus is competing in the same placental tissue space with the maternal uterine vascular pressure, which keeps the tissue pressure relatively high. If umbilical flow to the placenta decreases to less than 150 ml. kg-' * min-', the pressure typically approaches 40 mm Hg. Any further reduction in flow will be accompanied by a lower umbilical perfusion pressure. The surrounding tissue pressure will then compress the vessels, causing increased resistance. This hemodynamic problem can be avoided by having high flow rates on bypass. An analogous situation is a single ventricle patient with a systemic to pulmonary artery shunt. If bypass is instituted in this patient and the shunt is left open, flow must reach 300 to 400 ml kg-' * min-' to provide an adequate cardiac output to the body because at least one or two cardiac outputs will go through the shunt to the lungs. In the fetus, the umbilical vessels in the placenta are analogous to the shunt to the lungs in the single ventricle patient. Therefore bypass flow must reach 300 to 400 ml * kg-'. min-' at a minimum in the fetus to provide adequate flow both to the body and to the placenta. If this is achieved, fetal gas exchange and acid-base status are very stable. The second type of placental dysfunction is hormonal. Our laboratory has done some work with indomethacin, steroids, nitroprusside, and various other prostaglandin blockers, which suggests that prostaglandins and stress hormones are very important in the secondary responses of the placenta to cardiopulmonary bypass. We are slowly beginning to understand these very complex mechanisms which really do not have an analogous situation in the independent postnatal individual undergoing bypass.