Intrauterine blood transfusion
P. N. Adama van Scheltema, Department of Prenatal Diagnosis and Therapy, B3-87, PO BOX 9600, 2300 RC Leiden, the Netherlands
Fifty years ago, haemolytic disease of the newborn was one of the most common causes of perinatal mortality. Early delivery and neonatal exchange transfusions were the only treatment options, until in the 1960s, intrauterine transfusion became possible. Freda and colleagues performed open foetal surgery, transfusing the foetus using a vein in the exteriorized leg [1,2]. A major breakthrough was the development of percutaneous intraperitoneal transfusion under X-ray guidance by Sir William Liley . However, hydropic foetuses did not take up the transfused blood from the peritoneal cavity very well. In addition, the technique was not feasible before 27 weeks’ gestation . In the 1980s, Sir Charles Rodeck introduced the technique of intravascular transfusion by needling the umbilical artery under direct fetoscopic guidance . Shortly afterwards, Jens Bang in Denmark and Ferdnand Daffos in France both pioneered foetal blood sampling under ultrasound guidance. In the last two decades, their approach is still used worldwide as the technique for intrauterine transfusions [6,7].
The main indication for intrauterine blood transfusion still is foetal anaemia because of red cell alloimmunization. However, in any foetal disease with severe anaemia, intrauterine blood transfusion can be considered. Intravenous access to the foetus also allows treatment of other conditions, including alloimmune thrombocytopenia by platelet transfusions, and cardiac arrhythmias by injecting antiarrhythmic agents. The procedure of intrauterine transfusion is identical in all these diseases, independent of the indication. We will therefore focus on intrauterine transfusion for red cell alloimmunization.
When to transfuse: monitoring a pregnancy at risk
As in all surgical procedures, the technical details are only part of the clinicians’ skills. The correct indication and timing are at least as important. Especially in invasive procedures in obstetrics, in which any complication is potentially lethal for the foetus, a careful individual risk-benefit analysis must be made each time. In red cell alloimmunization, maternal serum antibody levels and obstetric history help to identify foetuses at risk for anaemia. However, not all foetuses in alloimmunized pregnancies develop severe anaemia. Serial assessment of signs of anaemia is necessary to enable optimal timing of intervention. The goal and the challenge is to transfuse only in case of severe anaemia, but before the foetus develops hydrops. In all series, survival of intrauterine transfusion in hydropic foetuses is significantly lower compared to non-hydropic foetuses . Several studies have shown that hydrops only develops when the foetal haemoglobin drops to levels 6–7 SD below the mean for the gestational age. Ideally, a transfusion is given when the foetal haemoglobin level is between 4 and 6 SD below the mean. In such cases, one can transfuse a relatively large amount of O-negative donor blood, which is not haemolysed by the red cell antibodies. The next transfusion can then be given several weeks later. Transfusion intervals will be discussed in more detail later.
As timing of transfusion is such an important issue, we will briefly describe the currently used diagnostic tests for evaluation of red cell–alloimmunized pregnancies.
Weekly ultrasonographic evaluation is important for early detection of foetal hydrops. Hydrops is a sign of severe foetal anaemia and, as mentioned before, associated with adverse outcome. The first signs of hydrops because of anaemia are ascites and pericardial effusion, practically always in conjunction with cardiomegaly. Skin oedema develops later, and pleural effusions are only apparent in advanced stages. The latter is also true for increased placental thickness and polyhydramnios. Some severely hydropic anaemic foetuses even have oligohydramnios. Foetal movements are sometimes remarkably normal despite severe anaemia.
There are other, less specific signs that indicate anaemia: enlargement of the heart, spleen and liver because of cardiovascular adaptation and increased extramedullary erythropoiesis, respectively.
Progressive anaemia leads to an elevated cardiac output and decreased blood viscosity. This results in an increased maximum or systolic flow velocities in various parts of the foetal venous and arterial circulation [9–13]. For clinical practice, the most reliable and reproducible of these Doppler parameters is the middle cerebral artery peak velocity [14,15].
The end product of foetal red cell destruction is bilirubin. Most bilirubin is removed from the foetus via the placenta but small amounts enter the amniotic fluid in the second and third trimester and increases in bilirubin level in amniotic fluid correlate with the foetal red cell destruction . The amniotic bilirubin concentration can be quantified spectrophotometrically by assessing the change in optical density at a specific wavelength . In 1961, Liley devised a system for managing pregnancies complicated by rhesus alloimmunization based on the correlation of bilirubin levels in amniotic fluid and neonatal outcome . This Liley chart, or an extended version, is still in use for predicting foetal anaemia. An alternative chart was produced by Queenan et al. , based on a reference range of bilirubin levels in normal pregnancies.
In most centres, the procedure is performed by a team consisting of at least three or four members. One operator to actually put the needle in the foetal vessel of choice, either using a free hand technique or with an ultrasonographer to guide the procedure, to monitor the transfusion site and the foetal condition, one nurse with sterile gloves on to assist with the various medications and syringes, and one assistant to perform the on-site blood tests and calculations [19,20]. In general, foetal blood transfusions including all preparations take 40–60 min and can be carried out as outpatient procedure or with an overnight stay. In our unit, we plan these procedures in the afternoon, and we discharge the patient either the same evening or the next morning.
Premedication and foetal paralysis
Some sedation may be useful for anxious patients, as it is important that they lie quietly on the operating table for about 30 min. Maternal pain relief by local anaesthetic injection at the puncture site is usually sufficient. Some operators administer fentanyl to the foetus in case of transfusion through the intrahepatic vein, to avoid foetal stress and pain [21,22]. Foetal movements during the procedure can cause needle displacement, haematomas or even tearing of the vessel wall with life-threatening bleeding as a result. The routine administration of muscle relaxants such as pancuronium or atracurium to achieve foetal paralysis, either intravenously or intramuscularly, was suggested to be the preferred policy in a recent study by Van Kamp et al. . Others only administer these substances when the foetus is very active, and the extremities are in close proximity of the transfusion site.
Site for transfusion
Depending on both foetal position and placenta localization, the operator chooses the site for transfusion, either the intrahepatic umbilical vein or the umbilical cord at placental insertion. Some operators prefer the placental cord insertion, because transfusion in the intrahepatic vein has been reported to be associated with an increase in foetal stress hormones [24–26]. Others prefer the intrahepatic vein because this technique minimizes blood loss from the cord puncture site, as the blood is absorbed from the peritoneal cavity . If the placenta lies anterior, we consider the umbilical cord at placenta insertion to be the safest choice, unless the foetus is blocking the view and/or insertion. If the placenta lies posterior, we consider the intrahepatic vein to be the safest choice. Puncturing a free loop of the umbilical cord has the risk of tearing during foetal movement, and ‘jet’ with an unknown amount of blood loss in the amniotic cavity after removal of the needle post procedure. Accessing or transfusing in the umbilical artery has a high complication risk, because the artery is smaller in diameter than the vein and is more likely to go into spasm during the procedure and cause a foetal bradycardia [23,28]. Cardiac punction is rarely used because of the potential hazards such as cardiac tamponade, hemopericardium and arrhythmia including asystole .
Intravascular versus intraperitoneal approach
Intraperitoneal transfusion relies on placing the donor cells into the peritoneal cavity so that they become absorbed into the foetal circulation via the subdiaphragmatic lymphatics and thoracic duct. This absorption is somewhat unpredictable and may be impaired in case of severe hydrops. Apart from the fact that intravascular transfusion delivers the donor cells immediately where they should be, the other major advantage is that the foetal haemoglobin levels can be obtained before and after the transfusion. The amount of blood (V) that needs to be transfused can be calculated using the following formula:
Commonly, the donor blood has an hematocrit of around 80%. Fetoplacental volume has been estimated by several authors, either based on gestational age  or estimated foetal weight . The aim for post-transfusion hematocrit, in case of complete intravascular transfusion, is generally set at 40–50%. This slightly supranormal value allows for a longer transfusion interval, minimizing procedure-related risks.
The volume given when intraperitoneal transfusion alone is performed has traditionally been calculated as follows:
IPT V = (gestational age in weeks −20) × 10
Intravascular transfusion is considerably more successful at reversing hydrops than intraperitoneal transfusion, possibly because hydrops reduces the absorption of red cells from the lymphatic system .
A few studies show considerable advantages of a combined technique of intravascular and intraperitoneal transfusion [33,34]. This should result in considerably longer transfusion intervals, although in one study the mean gain was only 3 days [19,33].
One remaining indication for intraperitoneal transfusion could be a high likelihood of severe anaemia in the very early (e.g. before 18–20 weeks) gestational age foetus, although experienced operators are usually able to access the cord or intrahepatic vein from 17 weeks gestation onwards. Sometimes, such cases, especially if associated with maternal obesity, can be a real challenge. We recommend watchful waiting, monitoring twice weekly if needed, and only attempt to transfuse at the first signs of hydrops. In very severely affected pregnancies, e.g. with previous foetal loss or hydrops before 20 weeks’ gestation, this policy can be combined with maternal IvIG administration, plasmapheresis or both. For details, we refer to an excellent review of these options by Moise .
Volume and rate of transfusion
On obtaining access to the foetal circulation, a 1 ml sample of foetal blood is obtained to determine the foetal hematocrit and haemoglobin level. Most centres have a Coulter Counter in the procedure room so that the result is available in a few seconds. Verification of correct placement of the needle is done by injecting a small amount of saline, followed by a foetal weight adjusted dose of muscle relaxant. The needle is connected to a sac with donor blood. The blood is guided through a heater, which warms the blood to body temperature. After transfusing the estimated amount of blood, the needle is flushed with some saline and left in place for 1–2 min to allow mixing of blood in the fetoplacental unit. Then, a 1- ml sample is taken for immediate assessment, and if the desired hematocrit is not yet reached, more blood is given. At the end of the procedure, a final sample is taken again after 1–2 min mixing time.
The blood is transfused at a rate of 5–10 ml per minute. During the procedure, the blood flow is continuously visualized on the ultrasound screen to confirm the needle position. Periodically, the foetal heart rate is checked for arrhythmias, especially bradycardia. If a bradycardia occurs, we wait with further transfusing until the heart rate has normalized. The procedure is performed using a 20 or 22 gauge needle, depending on gestational age.
The primary source of blood for intrauterine transfusion is O-negative unrelated donor blood. The blood has been collected within 24 h and cross-matched with the maternal blood. It should have been screened for hepatitis B and C, cytomegalovirus and HIV as well as irradiated to remove the white-blood-cells to avoid ‘graft-versus-host’-like complications in the foetus. The blood is packed to a hematocrit of 75–85% to minimize the volume of blood that needs to be transfused.
Maternal blood is also a good source of blood for intrauterine transfusions. It has the theoretical advantage of decreasing the risk for sensitization to new red cell antigens. In addition, a longer circulating half-life can be expected because of the fresh source of cells. Parents seem to have some preference for the use of mothers’ own blood to donate to the foetus. However, the blood must undergo rigorous testing for antibodies against the infectious diseases mentioned earlier. For logistic reasons, we have therefore chosen to use unrelated donor blood. A study comparing the foetal effects of maternal red cells vs. unrelated donor cells showed no significant differences in decline rate of red blood cells until 33 weeks gestation .
Top-up versus exchange transfusion
Concern has been raised that directly transfusing blood into the foetus without removing any blood (the top-up procedure) may lead to volume overload and cardiac compromise. Some operators aspirate small amounts of blood from the foetus at regular intervals during the transfusion with the intention of preventing hypervolemia (the exchange procedure) [36,37]. Others suggest that the umbilical venous pressure should be routinely monitored and if the change in pressure exceeds 10 mmHg then blood should be removed and replaced with an equal volume of saline . In practice, with many centres performing top-up transfusions for 20 years now, the foetus appears to tolerate the top-up transfusion without any adverse effects.
Timing of subsequent transfusions
The second transfusion usually has to be performed 2–3 weeks after the first transfusion, or earlier if the desired post-transfusion hematocrit has not been reached. The mean fall in hematocrit is around 1% per day, but there is a wide variation (SD 0·44) . This donor cell fall is due mainly to growth of the foetus with increasing fetoplacental volume and only partly because of actual destruction of red cells. This rate of fall is quite unpredictable between the first and second transfusion, because at this stage, the percentage of foetal erythrocytes in the foetal circulation and the suppression of erythropoiesis are variable. After the second transfusion, the mean fall in hematocrit is generally better to predict, because at this stage, almost all of the foetal erythrocytes are replaced with donor erythrocytes and the erythropoiesis is almost completely suppressed. After the second transfusion, the interval can safely be 4 or even 5 weeks if the hematocrit after transfusion is at least 45%. In general, transfusions are performed until the gestational age of 35 weeks is reached . This means that a pregnancy in which transfusions are performed can be continued until approximately 37 weeks, thereby reducing the risks of respiratory and other prematurity problems.
Several groups have published their experience showing that repeated intravascular transfusions throughout pregnancy are associated with survival rates ranging from 76% to 96% [30,36,40–42].
Intrauterine transfusion is a safe procedure with a relatively low procedure-related complication and perinatal loss rate. However, complications do sometimes occur. Foetal complications during or after an invasive procedure may either result from the procedure or from the underlying pathologic condition necessitating treatment. This means that not all complications are because of the procedure but some are a consequence of the compromised foetal condition.
Transient foetal bradycardia during transfusion is the most common complication, occurring in 8% of procedures . Foetal distress during or after transfusion is the most feared complication and may result in foetal death or emergency delivery with the risk of neonatal asphyxia and death. Foetal distress can occur after cord accidents (rupture, spasm, tamponade from a haematoma), haemorrhage from the puncture site, volume overload, chorioamnionitis, preterm rupture of membranes or preterm labour. Fortunately, all these complications are rare. Between 1988 and 2001, our centre has systematically scored all procedure-related complications (32). The total procedure-related complication rate was 2·9% in the absence of hydrops and 3·9% in the presence of hydrops. The risk of procedure-related perinatal death was 1·6% per procedure (32).
After the transfusion, the woman is advised to rest on her left side for half an hour, to optimize blood flow to and from the uterus, during which time the foetal heart rate is monitored by cardiotocography. She then remains in bed until all effects of the sedatives have gone, usually 2–4 h. She is discharged home the same or the next day. A follow-up appointment is typically scheduled for 10–12 days afterwards.
Outcome of treatment
Survival after intrauterine transfusion varies with centre, experience and the presence of foetal hydrops. Foetal loss rate per intravascular transfusion ranges from 0·6% to 4% [32,33].
Several groups have studied the long-term outcome for children treated with intrauterine transfusions with follow-up ranging from 6 months to 6 years [43,44]. The neurodevelopmental outcome of these children is normal, even for the children who initially presented with foetal hydrops. Hearing deficits in the neonate have been reported in association with high serum bilirubin levels [45,46]. However, a study by Janssens et al.  did not show any permanent hearing problems in children treated with intrauterine transfusions.
The authors declare that there are no potential conflicts of interest.