Direct Fetal Transfusion

Present-day guidelines for the administration of Rhesus (Rh) immune globulin (RhIG) to every Rh_o(D)-negative women at 28 weeks' gestation and postpartum have greatly reduced the incidence of red blood cell alloimmunization.¹ Before the use of RhIG, nearly 1% of all Rh_o(D)-negative pregnant women were sensitized.² Nonetheless, in 1986 the Centers for Disease Control reported an incidence of Rh hemolytic disease of 10.6 per 10,000 total births in the United States.³ For every 100 affected patients, nine fetuses require intrauterine transfusion (IUT).^{4, 5, 6}

IUT was the first effort at in utero treatment of the human fetus. During the last 30 years, it has become the treatment of choice for patients alloimmunized to red blood cell antigens in whom severe fetal anemia develops remote from term.^{4, 5, 6} In these cases, IUT generally is indicated if the fetal hemoglobin concentration is less than the 5th percentile for gestational age⁷ or if the fetal hematocrit is less than 30%.^{4, 5, 6}

We refer the reader elsewhere for a description of the complete diagnostic evaluation of the pregnancy affected by Rh hemolytic disease.^{2, 4, 5, 6, 7, 8, 9} (See also GLOWM Chapter Alloimmune hemolytic disease of the fetus and newborn (erythroblastosis fetalis): diagnosis, management, and prevention) Instead, this chapter stresses the technical aspects of IUT, including the use of ultrasound, the physiologic effects of IUT on the fetus, complications, neonatal “top-up” transfusions, and pregnancy outcome. The use of IUT in the treatment of fetomaternal hemorrhage, fetal parvovirus infection, and platelet alloimmunization is described and the future of IUT briefly evaluated.

In 1963, Liley^{10, 11, 12, 13} initiated the in utero therapy of Rh hemolytic disease by developing the technique of fetal intraperitoneal transfusion (IPT). During an amniocentesis, Liley inadvertently placed his needle into the fetal abdomen. After returning from Africa, one of Liley's research associates described how IPT was commonly used there to transfuse children with sickle cell anemia. Liley concluded that this could be a feasible route for IUT if the fetal peritoneal cavity could be entered intentionally.^{14, 15} After instilling radiopaque dye into the amniotic cavity by amniocentesis, Liley made use of the fetal ability to concentrate this contrast material in its lower gastrointestinal tract, thereby providing a radiographic target for needle placement. He then employed paper clips fastened with adhesive tape to mark the maternal skin surface. During the next step of the procedure, Liley inserted a 16-gauge Tuohy needle into the fetal peritoneal cavity and advanced an epidural catheter through it to aspirate 200 mL of fetal ascitic fluid and replace it with 100 mL of packed red blood cells. After three fetal deaths, Liley reported his first success in September 1963.¹² Virtually simultaneous, though more aggressive attempts to access the fetal circulation directly included hysterotomy-facilitated catheter placement into chorionic plate vessels or into the fetal femoral artery, saphenous vein, or internal jugular vein.^{16, 17, 18, 19}

Before real-time ultrasound was available, IPT involved blindly injecting radiopaque dye into the amniotic cavity. The fetus would then swallow this contrast material, allowing for identification of its peritoneal cavity at fluoroscopy. Often a “chicken-wire” grid molded to the contour of the maternal abdomen was used to ensure accurate needle placement under fluoroscopic visualization.^{20, 21} This technique was later modified to include the direct injection of radiopaque dye into the fetal abdomen.²² The use of static gray-scale ultrasound imaging for placement of the transfusion needle was first reported in 1975, significantly reducing fetal radiation exposure by eliminating the need for pretransfusion amniography.^{23, 24} Real-time ultrasonography to guide the transfusion needle during IPT was first described in 1977.^{25, 26, 27} Only a short time later, fluoroscopy was no longer used. Saline, which produces small bubbles on real-time ultrasound when briskly shaken and injected under pressure, was used instead to observe the transfusion needle successfully entering the fetal abdominal cavity.^{25, 26} Bowman and associates²⁸ stressed that the diameter of the epidural catheter is important to the success of the IPT. To avoid side holes with small diameters causing hemolysis during red blood cell infusion, they recommended that the end of the epidural catheter be removed before being used for IPT. They estimated the appropriate transfusion volume for IPT using the following calculation:

IPT remained the only technique for in utero transfusion therapy until 1981, when Rodeck and co-workers²⁹ performed the first direct fetal intravascular transfusion (IVT) by placing a needle into chorionic plate vessels under fetoscopic visualization. One year later, Bang and colleagues³⁰ successfully transfused a fetus by inserting a needle into the umbilical vein under ultrasound guidance. Seven different techniques for IUT have been reported since then (Table 1). The use of fetoscopy in guiding the IUT needle is described in several studies. This method failed to achieve great popularity.^{31, 32, 33, 34} Both the direct fetal intravascular approach and fetal intravascular exchange transfusion were evaluated by investigators in the United States during the following 4 years.^{35, 36, 37, 38} Proponents of exchange transfusion postulated that with this technique the fetus was less likely to become volume-overloaded while more fetal cells were removed from circulation. Advocates of direct IVT believed that the procedure time was shorter in comparison with fetal-exchange IVT. They also asserted that the placental vascular bed would absorb the increased circulatory volume without negatively affecting the fetus. The direct intravascular technique became more widely adopted as experience with both techniques grew. Ultrasound imaging resolution continues to improve, and simple direct IVT has become the procedure of choice at most centers in the United States; IPT alone is performed rarely unless direct access to the umbilical cord is not technically feasible. Transfusion of red blood cells directly into the fetal circulation seems especially important in the treatment of the hydropic fetus whose peritoneal absorption of red blood cells is not inhibited, but it is thought to require several weeks.^{39, 40, 41} Comparing neonatal outcome after IVT to that after IPT using historic controls, Harman and co-workers⁴² clearly demonstrated that IVT significantly improves the chance for survival in the hydropic fetus.

Table 1. Techniques for intrauterine transfusion

Intraperitoneal
Intravascular by cordocentesis
     Exchange transfusion
     Simple direct transfusion
Intravascular by intrahepatic venous puncture
     Simple direct transfusion
Intracardiac by fetal cardiocentesis
     Simple direct transfusion
Combined approaches
     Exchange intravascular transfusion followed by intraperitoneal transfusion
     Simple direct intravascular transfusion followed by intraperitoneal transfusion

Because complications unique to IVT had been reported, the need for IVT in the severely anemic but nonhydropic fetus remained controversial.^{43, 44, 45} Moreover, with the use of direct IVT alone, fetal hematocrits between procedures varied widely. The combination of IVT and IPT therefore was tested at Baylor College of Medicine.⁴⁶ First, packed red blood cells with a hematocrit of 80% or higher were infused by IVT to increase the fetal hematocrit to approximately 40%. Next, a standard IPT was performed to create a red blood cell reservoir in the fetus for later absorption between IUTs. This combined approach resulted in a more stable fetal hematocrit and longer intervals between procedures (Fig. 1).^{36, 46} As a result of the combined IVT/IPT approach, the average daily decline of the fetal hematocrit was only 0.01% compared with 1.14% when IVT alone was used. Other centers have validated these findings and also use the combined IUT technique.^{47, 48}

As an alternate site for IUT, the intrahepatic umbilical vein has been used when access is impossible at the placental cord insertion.^{49, 50, 51} Nicolini and co-workers^{50, 51} transfused 72 severely anemic fetuses intrahepatically with a 90% success rate. There were few complications, which consisted mainly of fetal bradycardia and intraperitoneal bleeding. Fetal bradycardia during IUT often is related to inadvertent umbilical arterial puncture and vasospasm.^{7, 52} During intrahepatic transfusion, the incidence of fetal bradycardia was low because of the absence of the umbilical artery at the needle insertion site. Only occasional fetal intraperitoneal bleeding occurred in this series. The extravasated blood was absorbed from the peritoneal cavity in all cases, and there were no adverse fetal effects.^{50, 51}

Westgren and co-workers⁵³ described direct fetal intracardiac transfusion (ICT) in patients with severe Rh hemolytic disease. Six patients with evidence of severe erythroblastosis fetalis underwent the procedure at 19–31 weeks' gestation. A total of 25 ICTs were completed, with a complication rate of 20%. According to these authors, ICT offers an alternative if direct IVT into the umbilical cord is impossible.⁵³ Using ICT, Harman⁷ described the resuscitation of five fetuses from exsanguinating hemorrhage after umbilical cord puncture. After the initial ICT was successful, serial IVTs were performed subsequently with good outcome.

Target fetal hematocrits at the conclusion of IUT vary among treatment centers.⁵⁴ A final fetal hematocrit of 50–60% is usually the goal when IVT alone is performed. A rise in fetal whole-blood viscosity during transfusion can be minimized by restricting the posttransfusion fetal hematocrit to approximately 50–55%.⁵⁵ The final target hematocrit in the combined IVT/IPT approach is usually approximately 40%. This is more physiologic because the fetal hematocrit normally ranges from 33% at 17 weeks' to 47% at 40 weeks' gestation.⁵⁶ To estimate the volume of red blood cells needed for an IVT, the fetoplacental volume is determined first.⁵⁷ The fetoplacental volume relative to fetal weight estimated by ultrasound is rather constant throughout the course of pregnancy.⁵⁸ This value in milliliters is equal to 1.046 plus the ultrasound-estimated fetal weight in grams multiplied by 0.14. The volume to be transfused during IVT is then computed:

This calculation is based on a simple dilution formula and can be computerized. Repeated hematocrit determinations during the IVT can be avoided in this manner.⁵⁷ Alternatively, a transfusion coefficient can be used to determine the amount of blood to be transfused (Table 2).⁵⁹

Table2. Volume of blood (hematocrit 78%*) to be transfused to obtain desired increase in fetal hematocrit

Most fetuses undergoing IVT with a normal umbilical venous pressure have a pressure of less than 10 mmHg at the end of the transfusion.⁶⁰ Monitoring fetal umbilical venous pressure is still useful for assessing the fetal response to IUT⁶¹ because pressure increases of more than 10 mmHg have predicted fetal death within 24 hours after IVT with a sensitivity of 80%.⁶² Severely anemic and hydropic fetuses experience particularly high loss rates. Radunovic and co-workers⁶² therefore recommended that the posttransfusion fetal hematocrit not exceed 25% or a fourfold increase from the pretransfusion value. Selbing and associates⁶³ noted that if the fetus is given more than 20 mL of fluid for each kilogram of its ultrasound-estimated weight, its chance of survival is reduced significantly. These authors strongly suggested an upper transfusion limit of 20 mL/kg, which corresponds to approximately one fifth of the fetoplacental blood volume.

Timing of subsequent fetal transfusions varies among institutions. At Baylor College of Medicine, anemic fetuses are transfused every 2 weeks for the first two transfusions; then the interval between procedures is lengthened to 3–4 weeks. Severely anemic fetuses diagnosed early in the second trimester are an exception to this guideline. Such fetuses are transfused at first to achieve a final hematocrit of approximately 25%. In a subsequent IUT approximately 48 hours later, a posttransfusion hematocrit of 35% is the goal. Equations to predict fetal hematocrit at the beginning of subsequent IUTs have been developed, and timing of repeat transfusions can thus be estimated optimally.^{7, 64, 65}

Initial studies of the effects of IUT focused on the rate of absorption of transfused blood from the peritoneal cavity by the fetus.^39,66 At first, there was great concern that hydropic fetuses were unable to absorb red blood cells after IPT.³⁹ One case report showed that 2 hours after a 60-mL IPT, only 18% of circulating red blood cells in the fetus were adult-type cells.⁶⁷ Using chromium-labeled packed red blood cells during IPT, it was later demonstrated that, though over several weeks' time, even the hydropic fetus can absorb them from its peritoneal cavity.^40,41

The survival of transfused red blood cells in the fetal circulation also has been investigated. Maternal donor lymphocytes were shown to persist for more than 2 years in the peripheral blood of four infants transfused in utero.⁶⁸ A case report by Jones and co-workers⁶⁹ suggested that the half-life of the donor adult erythrocyte is only 30 days in a second-trimester fetus, compared with approximately 60 days if transfused into an adult. Pattison and Roberts,⁷⁰ however, studied 27 anemic fetuses and found that adult erythrocyte survival in the fetus was similar to that in the adult circulation. These authors also noted that red blood cell survival was independent of the route of transfusion, gestational age, or presence of hydrops. The fraction of transfused red blood cells surviving at 29 days after IVT was approximately 65%, compared with 8% in the above-cited case report.⁶⁹

There has been considerable change in the source of donor red blood cells for IUT. Virtually all centers use fresh, O-negative cells for IUT; several centers use maternal blood as an alternate source. This source of fresh red blood cells is readily available, the risk of maternal alloimmunization to new red blood cell antigens due to any fetomaternal hemorrhage is theoretically reduced, and transmission of viral infections to the fetus is likely to be decreased. However, Vietor and colleagues⁷¹ studied 91 women treated with IUT for severe red blood cell alloimmunization and tested them for the development of additional alloantibodies. These investigators detected new alloantibodies against red blood cell antigens in 24 women (26%). These were usually directed against fetal rather than donor antigens. The use of maternal red blood cells during IUT in these cases would not have prevented further alloantibody formation. Finally, the use of maternal red blood cells for IUT has a psychologic benefit in that any guilt over fetal rejection would be alleviated by providing mothers with a practical way to help their fetuses.

Up to 6 U of maternal blood can be safely harvested for fetal transfusion during a gestation.⁷² Regular intake of prenatal vitamins, folate, and ferrous sulfate prevents significant maternal donor anemia. All mothers are screened routinely for syphilis, West Nile virus, HIV1 and 2, human T-cell lymphotropic virus type 1 (HTLV-1), and all known forms of hepatitis before donating for their fetuses. Donated red blood cells are washed to remove the offending antibodies and packed to achieve a hematocrit of approximately 80%. Units then are processed through a leukocyte-poor filter and irradiated with approximately 2500 rad (25 Gy) to prevent any graft-versus-host reaction. Great care must be taken to ensure timely processing of fresh blood to avoid transfusion of blood with high plasma potassium levels, which could cause fetal arrhythmias after IUT.^{73, 74} Mothers with antibodies to cytomegalovirus may still donate because this virus resides in white blood cells removed during the filtration process. If initial cordocentesis reveals an ABO incompatibility between mother and fetus, maternal red blood cells should be used only with great caution because of the risk of fetal sensitization. At Baylor College of Medicine, maternal blood was utilized in two such cases with no adverse effects observed in the fetus or neonate.

Ultrasound has played an important role in improving the outcome in pregnancies affected by Rh hemolytic disease. It is used to establish the correct gestational age on which parameters such as normal fetal hematocrit and amniotic bilirubin levels are based. Ultrasound provides invaluable needle guidance during amniocentesis, cordocentesis, and IUT. Unfortunately, its use in the diagnosis of fetal anemia is limited until overt hydrops fetalis is present. Although some investigators have advocated serial ultrasound examinations to detect signs of impending hydrops fetalis, Nicolaides and co-workers⁷⁵ were unable to correlate fetal hematocrit with increased placental thickness or increased umbilical vein diameter. Because the liver and spleen are sources of extramedullary hematopoiesis in response to fetal anemia, these organs have been studied separately to predict fetal anemia. Two studies^{76, 77} have proposed that an increase in fetal liver dimensions is a good predictor of anemia. Oepkes and co-workers⁷⁸ measured the fetal splenic perimeter with ultrasound and noted that splenomegaly was present in all nonhydropic, anemic fetuses. Splenomegaly predicted a hemoglobin deficit in excess of five standard deviations from the normal range for a given gestational age, with a sensitivity of 93%. Some hydropic fetuses still had no splenomegaly.

The cardiac output of the anemic fetus has been demonstrated to be increased over controls.⁷⁹ This would lead to enhanced velocity of blood in various fetal vessels. For this reason, Doppler ultrasound has been studied to see whether it could serve as a valuable adjunct in the evaluation of the status of the fetus requiring IUT.⁸⁰ Pulsed Doppler wave forms from the fetal descending aorta have been used to predict hematocrit before IVT.^{81, 82} Copel and co-workers⁸³ have proposed the following formula to predict fetal hematocrit:

When they applied this formula to 16 fetuses, they achieved a sensitivity of 90% and a specificity of 69%. Copel and co-workers^{83, 84} further concluded that their Doppler technique was useful in predicting fetal hematocrit and the need for an initial IUT, but the timing of subsequent IUTs was not improved. Pulsed Doppler flow-velocity indices in the umbilical artery do not correlate with fetal blood gas values in red blood cell-alloimmunized pregnancies, nor do they predict fetal hematocrit except in very severely affected fetuses clearly requiring IUT.^{85, 86} Study of fetal cerebral vessels, such as the common carotid and middle cerebral arteries, has not proved Doppler findings in these vessels to be any more useful in the prediction of fetal anemia.^{87, 88}

Investigations to evaluate the clinical usefulness of pulsed Doppler flow-velocity indices in the management of severe red blood cell alloimmunization confirmed the hyperdynamic circulation produced by the fetal anemia in both arterial and venous vessels.⁸⁹ An association with the degree of fetal anemia was again found only for the flow-velocity wave forms in the fetal aorta. Steiner and associates⁹⁰ showed that the clinical usefulness of the peak aortic velocity is hampered by its low predictive value. Bahado-Singh and co-workers⁹¹ demonstrated that the splenic artery resistance index increases in fetuses with severe anemia. These authors postulated that this occurs as a reflection of vascular congestion from red blood cells trapped in the splenic circulation.

Doppler flow velocity assessment is now the standard of care for the diagnosis of fetal anemia.^{92, 9, 93}Chapter Alloimmune hemolytic disease of the fetus and newborn (erythroblastosis fetalis): diagnosis, management, and prevention

The operator has been plagued by fetal movement during IUT since the first description of the procedure by Liley.¹² Fetal immobilization was first described by Liggins,⁹⁴ who performed IPT by an impaling technique using multiple needle punctures of the fetus. Since then, IUT has been made easier by the widespread use of fetal paralytic agents. Fetal paralysis was first introduced in 1985 in Australia.⁴⁹ In the United States, D-tubocurarine initially was injected into the fetal thigh under ultrasound guidance.⁹⁵ The intravascular use of pancuronium bromide through cordocentesis was reported subsequently.^{96, 97} The absorption of red blood cells from the fetal peritoneal cavity was found to be markedly reduced after injection of pancuronium bromide in a sheep model.⁹⁸ Therefore, short-acting agents such as atracurium besylate and vecuronium bromide are currently in use.^{99, 100, 101} These two paralytic agents also do not cause the fetal tachycardia and loss of short-term fetal heart rate variability frequently demonstrated with the use of pancuronium bromide.¹⁰² A vecuronium or pancuronium bromide dose of 0.1 mg/kg of ultrasound-estimated fetal weight results in almost immediate cessation of fetal movement when injected intravascularly at the beginning of an IUT. This fetal paralytic effect lasts for up to 2 hours, and no untoward effects in neonates have been reported.

The use of color flow Doppler ultrasound also has improved the technique of IUT.¹⁰³ Color flow Doppler has an advantage over gray-scale ultrasound in that it allows for accurate localization of fetal vessels in difficult cases, such as early gestational age, oligohydramnios, or poor fetal positioning over the umbilical cord insertion site. When targeting the fetal hepatic vein for IUT, color flow Doppler helps to identify the vessel. During the actual injection of blood, intravascular turbulence at the tip of the transfusion needle is visible on color flow Doppler imaging.¹⁰³ In addition, the fetal heart rate can be monitored by observation of the umbilical artery pulsations. Thus, the operator's view of the transfusion needle is not interrupted by the need to observe the fetal heart rate directly on the ultrasound monitor. Finally, during IPT the color flow Doppler can be used to monitor the free flow of fluid into the peritoneal space, thereby confirming proper placement of the transfusion needle.

At Baylor College of Medicine, the following IUT procedure is used.¹⁰³ The IUT is performed in close proximity to the labor and delivery suite. With the patient under mild sedation and the abdomen prepped and draped in the usual sterile fashion, a 20-gauge needle, 6 inches in length, is directed under local anesthesia and real-time ultrasound guidance into the placental umbilical cord insertion. An initial fetal blood sample is sent for spun hematocrit, complete blood and reticulocyte counts, Kleihauer-Betke stain, and total and direct bilirubin. Next, the fetus is paralyzed with a pancuronium dose of 0.1 mg/kg of ultrasound-estimated fetal weight; fetal analgesia is provided by the use of fentanyl (10 μg/kg of estimated fetal weight).  These medications can be mixed in the same syringe and injected directly into the umbilical vessel. Red blood cells then are transfused at a rate of approximately 5–10 mL/min. A final sample of fetal blood is obtained for a hematocrit and a Kleihauer-Betke stain. After the IVT is completed, a standard IPT is undertaken in the combined IVT plus IPT approach performed at our center; we accomplish this simply by redirecting the IVT needle. After the procedure, the fetal heart rate is monitored continuously for at least 2 hours, after which the patient is discharged home. A follow-up ultrasound is performed the next morning.

The timing of delivery in fetuses undergoing serial IUTs has undergone considerable change. During the era of IPT, fetuses affected by hemolytic disease were delivered routinely at 32 weeks' gestation. Hyaline membrane disease and hyperbilirubinemia necessitating neonatal exchange transfusion were frequent complications. The widespread use of IVT during the past decade has led many centers to perform the final procedure at approximately 35 weeks' gestation with delivery planned approximately 3 weeks later. This change in management has virtually eliminated hyaline membrane disease and the need for neonatal exchange transfusions for elevated bilirubin.  The addition of at least 10 days of maternal phenobarbital (30 mg three times daily) at our center has resulted in a 75% reduction in the need for neonatal exchange tranfusions.¹⁰⁴ Once fetal viability is reached, IUTs should be undertaken close to the labor and delivery suite so that an immediate cesarean section can be performed in cases of fetal distress.

IVT leads to extensive fetal cardiovascular changes. The procedure greatly stresses the developing fetus with acute changes in intravascular volume and viscosity. Umbilical venous pressure increases between 1.7 and 4.6 mmHg.^60,105,106 Fetal cardiac output decreases by as much as 25%,¹⁰⁵ begins to return to pretransfusion levels within 2 hours,¹⁰⁷ and returns to normal by approximately 24 hours after the procedure.⁷⁹ To explain fully the fetal hemodynamic response to IVT, it has been suggested that the procedure leads to increased fetal cardiac afterload secondary to an increase in fetal blood viscosity. The fetus responds to this rise in afterload by a decrease in stroke volume, leading to a decrease in cardiac output and fetal heart rate and an increase in right atrial and umbilical venous pressures.^7,105 Pulsed Doppler assessment of umbilical and fetal femoral, renal, and cerebral vessels has demonstrated generalized fetal vasodilatation in response to this increase in afterload.^108,109 Results of investigations using pulsatility index as an indicator of fetal anemia and its correction by IVT also are indicative of a decrease in fetal vascular impedance immediately after IVT.^{105,108,109,110}

An increase in fetal vasodilator prostaglandins, which allows the human fetus to tolerate large increases in intravascular volume, has been reported after IVT.^111,112 An increase in atrial natriuretic peptide, a vasodilator, also contributes to the apparent acute decrease in vascular tone observed in the human fetus immediately after IVT.^113,114,115 Interestingly, a paradoxic increase in fetal plasma arginine vasopressin has been observed in association with IVT.¹¹⁶ This finding remains unexplained.

IVT also leads to significant alterations in fetal metabolism. Nicolini and co-workers¹¹⁷ found that the umbilical venous pH and base excess decrease while the PCO₂ acutely increases after IVT. In this study, the transfused blood had a mean pH of 6.76. These investigators believed that the fetoplacental circulation was effective in correcting this exogenous source of acidosis. The oxygen dissociation curve of fetal hemoglobin favors the uptake of oxygen in the placenta. Therefore, replacing fetal red blood cells with adult cells could seriously affect oxygen delivery to fetal tissues. Soothill and colleagues¹¹⁸ compared blood gas values of fetuses transfused with adult hemoglobin-containing red blood cells to those of controls. Transfused fetuses had a lower umbilical arterial pH and greater base deficit, while umbilical venous PO₂ was higher. These authors postulated that an increase in uteroplacental blood flow is the compensatory mechanism for fetal tissue hypoxia caused by transfused red blood cells carrying only adult hemoglobin.

An important compensatory response to anemia in the adult is an increase in the 2,3-diphosphoglycerate level, which reduces the oxygen affinity of adult hemoglobin and facilitates oxygen delivery to the tissues. Soothill and co-workers¹¹⁹ demonstrated that when adult red blood cells were transfused into the fetal circulation of 34 anemic fetuses, the 2,3-diphosphoglycerate concentration increased in direct correlation with the degree of anemia, allowing for improved fetal tissue oxygenation. Socol and associates¹²⁰ reported that IVT decreases hemolysis and reduces circulating levels of fetal plasma glutathione. Glutathione, liberated from hemolyzed fetal red blood cells, causes the inhibition of insulin activity in anemic fetuses, leading to a compensatory fetal hyperinsulinemic response.¹²¹ IVT therefore may prevent the hyperinsulinemia frequently associated with red blood cell alloimmunization.¹²⁰

Nasrat and associates¹²² demonstrated that there is a potential risk for iron overload in Rh-alloimmunized fetuses undergoing IUT. These authors described plasma ferritin levels indicative of iron overload in several fetuses. Severe fetal hemolytic anemia leads to elevated iron stores, which further increase in direct correlation with the volume of red blood cells transfused. Neonatal cholestasis and hepatitis due to severe intrahepatic iron deposition after IUT have been reported recently.¹²³ Therefore iron supplementation should be withheld in newborns with hemolytic disease until serum ferritin levels return to the normal range.

Before paralytic agents were in use to immobilize the fetus during IUT, fetal movement had the potential to result in visceral injury or umbilical cord trauma during the procedure. One case report even describes an IPT with subsequent fetal omental herniation through the fetal abdominal needle puncture site.¹²⁴ Anterior placental location increases the risk of damage to a major fetoplacental vessel with subsequent fetal exsanguination.¹²⁵ Perinatal infectious complications, such as viral hepatitis, also have been reported in the early experience with IPT.^{126, 127}

Several reports have described sinusoidal fetal heart rate patterns immediately after IPT.^{128, 129} A transient sinusoidal fetal heart rate after IPT is not necessarily an ominous sign; however, if persistent, it may herald fetal cardiac decompensation. This likely is caused by an increase in intra-abdominal pressure adversely affecting fetal heart rate control as a result of increased vagal stimulation.¹³⁰ Animal experiments have revealed that IPT causes increased fetal intraperitoneal pressure, which can lead to fetal hypoxia from umbilical vein compression.¹³⁰ The need for intraperitoneal pressure monitoring during IPT has since been supported by human data as well.^{106, 131}

IVT has led to complications not previously described with IPT, even though IVT generally is safe with a procedure-related pregnancy loss rate of approximately 1–3%.^{8, 52, 132} Entering umbilical vessels during IVT can cause an umbilical cord hematoma with resulting fetal compromise.⁴⁴ Some authorities have proposed that transient fetal bradycardia results from a perivascular blood collection leading to umbilical arterial vasospasm, with the extravasated fetal blood acting as an irritative focus. If umbilical vein thrombosis occurs in addition to the hematoma, fetal death may be the consequence.⁴⁵ Thus, visible echogenic turbulence is a sine qua non feature of the successful IVT. Disappearance of this marker signifies loss of proper needle location.¹³³ Furthermore, Moise and colleagues¹³⁴ noted that increased bilirubin levels due to hemolysis in the fetus of an alloimmunized pregnancy resulted in icteric serum in spun capillary hematocrit tubes. This finding is useful in confirming fetal vascular access in cases of anterior placentation, where visualization of the needle tip with ultrasound may be problematic.

Fetal bradycardia secondary to umbilical arterial vasospasm transiently occurs in approximately 4% of IVTs, especially as the target transfusion volume is being approached.^{7, 52} When the umbilical artery is entered, the incidence of fetal bradycardia approaches 30%.¹³² Therefore, the umbilical vein should be targeted routinely for IVT. On occasion, however, the umbilical artery is inadvertently punctured; in these cases, infused red blood cells will stream back toward the placenta during ultrasound visualization. Although immediate removal of the transfusion needle frequently is recommended, at our center we proceed with caution and frequently monitor the fetal heart rate using color flow Doppler. Slowing or stopping the transfusion and maternal oxygen administration usually leads to complete resolution of the fetal bradycardia. If these maneuvers do not reverse the bradycardia, the needle should be removed. Maternal repositioning in the left lateral position is undertaken. If the fetus is viable, we observe the bradycardia for up to 10 minutes before proceeding with an emergency cesarean section. Normal function of the atrioventricular valve leaflets in the fetal heart is a good indication that cardiac output is being maintained despite the slow heart rate. Cessation of movement of these valve leaflets on ultrasound examination warrants delivery.

Congenital viral infections can occur when donor blood is used for IVT. An unfortunate case of fatal congenital cytomegalovirus infection acquired through IUT was described previously.¹³⁵ Rare, but especially concerning, are cases of severe fetal brain injury after IVT.^{136, 137} Dildy and co-workers⁴³ documented an unexplained fetal porencephalic cyst after IVT. An increase in blood viscosity and fetal bradycardia during the IVT may have contributed to this complication by leading to fetal cerebral hypoperfusion. Fortunately, the infant later was discharged home with a normal neurologic status.

Fetomaternal hemorrhage can accelerate fetal hemolytic disease in the pregnant patient previously sensitized to red blood cell antigens. Moise and co-workers¹³⁸ described a case of poor fetal outcome after first-trimester transcervical chorionic villi sampling in a previously alloimmunized patient. The patient's antibody titers became elevated, resulting in the demise of a hydropic fetus early in the second trimester. These authors proposed that red blood cell alloimmunization is an absolute contraindication for chorionic villi sampling. Similarly, the placenta should be avoided during placement of the IUT needle. Nicolini and associates¹³⁹ described fetomaternal hemorrhages with an average volume of 2.4 mL in 21 of 32 patients undergoing IVT with an anterior placenta. Maternal alloantibody titers became elevated when the estimated volume of fetomaternal hemorrhage exceeded 1 mL. Hence, any significant fetomaternal hemorrhage may lead to more severe fetal hemolytic disease during a patient's subsequent pregnancies.

Survival of the fetus after IUT varies with the operator's experience, the institution, and the technique used. Results of studies on the outcome after IPT alone are presented in Table 3.^{135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164} Of 2483 reported fetuses who underwent IPT between 1963 and 1988, 39% survived. Only 17% of hydropic fetuses survived after IPT alone. Table 4 summarizes the reported experience with both IVT and IVT/IPT. Of 411 fetuses who underwent the procedure, 84% did well^{165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178}; 94% of nonhydropic fetuses and 74% of hydropic fetuses survived. An overall fetal survival rate of approximately 80% was reported by Baylor College of Medicine IUT using the combined IVT/IPT approach.^{4, 5, 6}

Table 3. Summary of studies using fetal intraperitoneal transfusion (IPT)

NA, not available

Table 4. Summary of studies using direct fetal intravascular transfusion

IPT, intraperitoneal transfusion; IVT, intravascular transfusion; NA, not available.

Initially it seemed that infants severely affected by red blood cell alloimmunization and transfused in utero had lower birth weights than matched controls,^{184, 185} but fetal catch-up growth after IVT has been observed in more recent studies.^{186, 187} It has been shown clearly that birth weights of infants transfused in utero are comparable to those of matched controls.

Few long-term outcome studies exist assessing infants transfused in utero.^{188, 189, 190, 191} Follow-up evaluations of fetuses treated with IPT have identified increased numbers of inguinal hernias in males and umbilical hernias in females as compared with matched siblings.¹⁹² In assessing overall physical, intellectual, and social maturity in IPT survivors for several years, Turner¹⁹³ considered only 50% to be completely normal. Other investigators have evaluated survivors of IPT for up to 11 years and found no significant developmental or physical abnormalities.^{192, 194}

Follow-up of IVT survivors is limited and must be viewed cautiously considering that the lives of more moribund fetuses have been saved.^{188, 191} Doyle and co-workers¹⁸⁸ studied 38 infants who had undergone IVT and found 35 to be developing normally. The three abnormal outcomes were attributed to prematurity.

Lozinska¹⁹⁵ demonstrated that IUT suppresses hematopoiesis. Therefore, many infants treated with IUT before birth require “top-up” transfusions during the early months of life. Because their red blood cell population consists mainly of transfused adult erythrocytes, reticulocytes are virtually absent. Neonatal exchange transfusions, however, rarely are necessary. Typically, infants with a history of IVT present with symptomatic anemia at 1 month of age and require a simple red blood cell transfusion. A review of 40 cases at Baylor College of Medicine showed that a top-up transfusion is necessary in approximately 50% of infants at a mean age of 38 days of life.¹⁹⁶ Most infants received only one such transfusion although some required as many as three. A recent retrospective series from the Netherlands found that 76% of the 52 infants in their series that had received intrauterine transfusions required top-up transfusions.¹⁹⁷ Infants requiring late neonatal transfusion are characterized by a lower reticulocyte count at their final IUT, higher umbilical cord hemoglobin at delivery, and a greater number of adult red blood cells seen on a cord Kleihauer-Betke smear. The affected infants have exhibited bone marrow erythroid hypoplasia, low serum erythropoietin, and a decreased number of circulating reticulocytes.^{198, 199} The mechanism causing the decreased erythropoietin is poorly understood; however, because passively acquired maternal antibodies remain elevated for at least 6 weeks in these neonates, the reticulocytopenia could be caused by high bone marrow levels of anti-red blood cell antibodies.

Therefore, weekly hematocrit and reticulocyte determinations should be performed for the first 2 months of life in infants with a history of IUT.²⁰⁰ Infants with a hemoglobin of less than 5–6 g/dL clearly should be transfused, even if they are symptom-free. Infants showing signs related to severe anemia, such as failure to thrive and lethargy, also are candidates for transfusion. Scaradavou and colleagues²⁰¹ reported the use of erythropoietin to treat late anemia caused by suppression of erythropoiesis due to IUT for Rh alloimmunization. Subcutaneous injection of 200 U erythropoietin per kilogram of body weight three times per week, combined with ferrous sulfate and folic acid supplementation, was effective in markedly decreasing the need for postnatal transfusion.

Massive fetomaternal hemorrhage is the cause of fetal death in 1 per 1000 births and results in significant fetal morbidity in at least 1 per 800 deliveries.²⁰² Seven cases of IUT in this situation have been described. Decreased fetal movement was the chief complaint in four of these cases, and subsequent fetal monitoring revealed a sinusoidal fetal heart rate pattern.^203,204,205 In the remaining three cases, routine ultrasound showed a hydropic fetus.^206,207,208 The maternal Kleihauer-Betke smear was positive in every case, and IUT was performed immediately. A single IUT procedure was needed in three cases, two procedures in another three cases, and five procedures in the remaining case. With this intervention, pregnancy was prolonged in three cases. In the first case, an IPT at 21 weeks' gestation was successful in reversing hydrops fetalis with subsequent delivery of a normal infant at 38 weeks' gestation.²⁰⁶ Thorp and co-workers²⁰⁸ performed two IVTs at 26 and 27 weeks' gestation. Hydrops fetalis was noted to resolve, and a normal infant was born at 39 weeks' gestation. At Baylor College of Medicine, Montgomery and co-workers²⁰⁷ transfused an affected fetus five times beginning at 27 weeks' gestation using the combined IVT/IPT approach. Abdominal delivery of a 1740-g infant at 30 weeks' gestation was required as a result of chorioamnionitis. The infant's initial hematocrit was 57%, and its neonatal course was uneventful. In all other reported cases, fetal bradycardia or the return of decreased fetal movement necessitated delivery by cesarean section. Cordocentesis in three of these fetuses revealed a decreasing hematocrit due to continued fetomaternal hemorrhage. The use of IUT in the treatment of fetomaternal hemorrhage therefore may prove beneficial only in selected cases.

Human parvovirus B19 causes erythema infectiosum (fifth disease) in children. During outbreaks, household contacts as well as school teachers frequently are exposed. Approximately 50% of persons lack immunity, and 20% of these will become infected after exposure.^{209, 210} One third of infected pregnant women will remain asymptomatic and will not manifest the exanthem typical of this disease.²¹¹ Reported rates of fetal infection range from 2.5% to 38%.^{212, 213} Parvovirus suppresses the fetal bone marrow and inhibits fetal erythropoiesis with the subsequent development of aplastic anemia, nonimmune hydrops, and fetal death.

Maternal parvovirus infection is often confirmed by detection of a specific IgM antibody that appears 3–4 days after the onset of clinical disease and persists for 3–4 months.²¹⁴ Fetal infection usually occurs 4–6 weeks after maternal infection, although hydrops fetalis has been reported as late as 12 weeks after the maternal illness. If maternal infection has been confirmed, the fetus should undergo weekly ultrasound examinations for the presence of hydrops fetalis for a total of 12 weeks. If hydrops is noted, donor red blood cells and platelets should be prepared, and cordocentesis performed. Studies of fetal blood usually show a negative direct Coombs' test, severe anemia, occasional thrombocytopenia, an inappropriately low reticulocyte count, normal serum bilirubin, elevated total IgM, and elevated liver enzymes.²¹⁵ Under the electron microscope, viral particles may be observed in fetal ascitic fluid or blood.²¹⁶ Using probes specific for human parvovirus B19, viral DNA also can be identified in some fetal body fluids.²¹⁴ Spontaneous resolution of hydrops fetalis secondary to fetal parvovirus infection has been reported by several authors.^{217, 218} Thus, if the fetal reticulocyte count is elevated before the first IUT, a second procedure is not necessary because recovery of the fetal bone marrow is ongoing. Nonimmune hydrops due to in utero parvovirus infection may require several weeks to resolve after the fetal hematocrit has returned to normal. 

In one series of 539 cases of parvovirus-induced hydrops, spontaneous resolution of the hydrops occurred in one third of cases while an additional 30% of fetuses died without treatment. IVT resulted in an 84% overall survival.²¹⁹ Therefore, IVT should be strongly considered as a therapeutic option in cases of fetal parvovirus infection with resultant hydrops fetalis. Although initial neonatal outcome was thought to be normal in those cases treated with IVT, more recent data call this into question. Nagel et al.²²⁰ followed 16 survivors that had presented with hydrops in utero. Thirty-two percent demonstrated psychomotor delay – mild in three cases and severe in two cases.

Neonatal alloimmune thrombocytopenia, a disease process analogous to red blood cell alloimmunization, complicates 1 per 1000–2000 live births.²²¹ In this situation, transplacental passage of maternal antibodies to fetal platelet antigens of paternal origin results in severe fetal and neonatal thrombocytopenia. In utero intracranial hemorrhage occurs in 10% of cases and has been documented as early as the second trimester.^{222, 223} In 75% of affected fetuses, the platelet antigen involved is HPA-1.²²⁴ Unlike Rh alloimmunization, the first fetus may already be severely affected; maternal antibody titers are not predictive of the degree of fetal thrombocytopenia.^{222, 225} The clinician often becomes aware of the disease only after an affected infant is born to the patient or her sister. Subsequent pregnancies are commonly complicated by progressively worsening fetal disease.²²⁶

The initial evaluation of an affected neonate should include platelet antigen testing of the newborn, mother, and father. Determining the paternal zygosity with respect to the HPA-1 antigen will assist in the prediction of the fetal antigen status in the vast majority of instances because only 5% of fathers are heterozygous.²²⁷ The likelihood of recurrent HPA-1 fetomaternal incompatibility is therefore approximately 94%.²²⁷ Khouzami and associates²²⁸ described the use of amniocentesis to determine the fetal platelet antigen status employing allele-specific oligonucleotide probes in uncultured amniocytes. Currently, fetal platelet count and antigen status of any subsequent pregnancy are determined most commonly by cordocentesis at 20 weeks' gestation. Unfortunately, this approach is not without risk. Several cases of fetal death caused by hemorrhage have occurred after cordocentesis in the investigation of platelet alloimmunization.²²⁹ Therefore, maternal platelets for immediate transfusion should be available at the time of cordocentesis should excessive streaming of blood occur from the umbilical vessel puncture site. Alternatively, if the fetal platelet count can be assessed rapidly while the cordocentesis needle is still in place, platelet transfusion should be performed for counts less than 50,000/mm³. This threshold is based on neonatal courses complicated by bleeding in infants born to women with autoimmune thrombocytopenia.²³⁰ Because approximately 98% of the population is HPA-1 antigen positive, maternal platelets should be used for fetal transfusion.²³¹ The platelets are obtained by apheresis 24 hours before their anticipated use, washed to remove any trace of offending antibodies and resuspended in ABO-compatible plasma.

The volume of platelet concentrate to be used can be calculated with the following formula:

where V represents the volume of platelets to be transfused, VFP is the estimated fetoplacental blood volume for the respective gestational age,⁵⁸ C₁ is the fetal platelet concentration before transfusion, C₂ is the concentration of the donor platelets, and C₃ is the fetal platelet concentration desired after transfusion.²³² The optimal posttransfusion fetal platelet count to prevent bleeding complications has yet to be determined. For this reason, most clinicians will transfuse a quantity of platelets empirically, based on the formula presented here. Although a final fetal platelet count should be obtained at the conclusion of the procedure, it is not necessary to wait for the result before the transfusion needle is removed.

Initially, fetal platelet transfusions in cases of platelet alloimmunization involved serial infusions of maternal platelets by cordocentesis. Four cases have been described.^{232, 233, 234} In two of these cases, the median daily decline in fetal platelet count was 23,600/mm³. Because of the short platelet half-life of 4–7 days, weekly transfusions are required to maintain an adequate fetal platelet count (Fig. 2).²³³ Bussel and colleagues²³⁵ have argued that this approach is associated with undue risk to the fetus from multiple umbilical cord punctures. Using an alternative approach, Lynch and co-workers²³⁶ successfully combined the maternal administration of steroids and weekly intravenous gamma globulin injections to increase the fetal platelet count at delivery to greater than 50,000/mm³ in 11 of 18 cases. No case of intracranial hemorrhage occurred in the treatment group, compared with a 33% incidence in antecedent siblings.²³⁶ Therefore, serial in utero platelet transfusions are not a reasonable primary approach to the management of affected fetuses with alloimmune thrombocytopenia.

Daffos and co-workers²³² proposed cordocentesis at term in fetuses suspected of having thrombocytopenia due to maternal platelet alloimmunization. If fetal thrombocytopenia is detected, transfusion with maternal platelets can then be performed. Subsequent induction of labor with vaginal delivery would be a safe option. Ten cases of fetal transfusion with platelets just before delivery have been reported.^{225, 232, 234, 237} Initially, investigators performed the fetal platelet transfusion and then proceeded with elective cesarean section, with the reasoning that cesarean delivery alone is not completely protective against intracranial hemorrhage in fetuses affected by alloimmune thrombocytopenia.²³⁸ More recently, vaginal delivery was allowed after fetal platelet transfusion in four of the 10 cases in this series. There were no adverse neonatal sequelae. Sia and colleagues,²³⁸ however, recommended vaginal delivery only with a posttransfusion fetal platelet count greater than 100,000/mm³.

Severe fetal thrombocytopenia has been associated with hydrops fetalis caused by red blood cell alloimmunization and parvovirus infection.^{215, 239, 240} In these cases, random-donor platelets also should be available at the time of IUT with red blood cells. If the fetal platelet count is noted to be less than 50,000/mm³, transfusion with platelets may prevent fetal death due to prolonged bleeding from the umbilical vessel puncture site.

Unfortunately, IUT only briefly relieves the symptoms of severe fetal anemia or thrombocytopenia caused by red blood cell or platelet alloimmunization and is intended only to prolong the intrauterine period available to the developing fetus to achieve maturity. Even though gamma globulin and steroids are promising alternatives to serial fetal platelet transfusion in the treatment of alloimmune thrombocytopenia,²³⁶ efforts to treat Rh hemolytic disease directly by administration of intravenous gamma globulin, plasmapheresis, oral Rh antigen ingestion, or promethazine have failed to date.^{241, 242, 243, 244} IVT is well established at many centers throughout the United States, and improvement of fetal survival beyond the 84% presented here is unlikely. To achieve better treatment results, a direct approach to the disease is needed.

The human Rh gene has been cloned, and the exact blood type of the developing embryo can be determined.^{245, 246, 247, 248} Using in vitro fertilization techniques, preimplantation diagnosis would allow for the exclusive return of Rh_o(D)-negative embryos to the uterus.²⁴⁹ Such an approach would lead to an unaffected fetus in all successful implantations.

Using a rabbit model of Rh hemolytic disease developed by Moise and associates,²⁵⁰ immunization to paternal leukocytes has been demonstrated to offer promise as a method of maternal immunotherapy for the treatment of red blood cell alloimmunization.²⁵¹

Widespread use of RhIG has reduced the incidence of Rh hemolytic disease in the United States to only 1 per 1000 live births.³ Only 9% of affected patients require IUT yielding less than 500 cases in the United States annually.^4,5,6 To optimize outcome with IUT as state-of-the-art therapy for severely anemic fetuses remote from term, regionalization of treatment should be considered at centers where a minimum of 10 fetal transfusions per year are performed. After diagnostic testing using ultrasound and amniocentesis performed by local obstetricians has shown a severely affected fetus requiring cordocentesis and possible IUT, the patient would enter the referral network. Concentrating patients with severe hemolytic disease at such centers would also provide the opportunity for continued basic and clinical research.

IUT is a safe procedure^8,52,132 and has saved the lives of many fetuses, including several infected with human parvovirus.^{211,215,252,253} It is the best available therapy until red blood cell alloimmunization can be prevented completely. IUT in the treatment of fetomaternal hemorrhage may prove beneficial only in selected cases.^{203,204,205,206,207,208} Fetal platelet transfusions in cases of severe platelet alloimmunization are of limited therapeutic value because of the short half-life of this blood component.^{232,233,234,235}