Fetomaternal blood incompatibility can be responsible for severe fetal anemia. Its optimal management today is based on in utero fetal blood transfusion by a vascular access to the umbilical cord.1,2 In some severe cases, anemia can occur very early, before 20 weeks of gestation, and even before the end of the first trimester. In this situation, performing an intravascular transfusion can present particular technical difficulties related to the thinness of the umbilical cord, which may make the procedure impossible or life-threatening for the fetus.1,3,4 Other techniques, thought to be less effective, such as intraperitoneal transfusion can be considered in this situation.5,6
The aim of this study was to estimate the outcome of the cases of very early fetal anemia that required an in utero transfusion before 20 completed weeks of gestation as a function of the techniques used in our referral department.
MATERIALS AND METHODS
This retrospective study covered all pregnancies with fetomaternal alloimmunization followed in or referred to the French National Reference Center for Perinatal Hemobiology from January 1, 1990, through August 15, 2011, in which at least one in utero transfusion was performed or attempted before 20 completed weeks of gestation. This is a continuous series of unselected patients in a single center by the same medical team. For each case, we collected the woman's obstetric history and characteristics, the type of alloimmunization, the methods by which fetal anemia was diagnosed, gestational age at the in utero transfusion, in utero transfusion technique, pregnancy outcome, and neonatal condition at birth. The study was approved by an ethics committee for research in obstetrics and gynecology.
The complications were analyzed 1) by pregnancy to assess the survival rate in these high-risk pregnancies requiring multiple transfusions; and 2) by transfusion performed before 20 weeks to assess the complication rate associated with each of these ultraearly procedures.
Student t test was used for comparison of the continuous variables. P<.05 was selected to define statistical significance.
During the study period, 25 transfusions took place (or were attempted) before 20 weeks for 18 pregnancies of 16 patients. During the same period, a total of 393 transfusions for red-cell alloimmunization were performed during 149 pregnancies of 141 women. Mean gestational age (±standard deviation) at the first in utero transfusion was 26 4/7 weeks (±4 3/7 weeks) for all of these pregnancies.
The very early in utero transfusions, before 20 weeks, therefore, accounted for only 6.3% of the transfusions performed in our series but involved 11.3% of the women and 12.1% of the pregnancies.
Among these 18 pregnancies with early diagnoses of anemia, the immunization was anti-RhD in 14 cases and anti-K1 in four. Table 1 summarizes the women's obstetric characteristics. Ten cases had a history of severe fetal anemia resulting from alloimmunization observed very early in a previous pregnancy: four spontaneous fetal deaths resulting from undiagnosed fetal anemia, three deaths from fetal anemia treated by in utero transfusion, and three live births after one or more in utero transfusions. The other eight pregnancies had shown no signs of fetal anemia in the previous pregnancy (Table 1).
Before 2002, peak systolic velocity of the middle cerebral artery was not used and fetal anemia was diagnosed based on ultrasonographic signs (ascites, pericardial effusion, subcutaneous edema), even minimal, in patients with severe alloimmunization (high antibody titer), especially if there was a history of fetal anemia, death, or both. Thus, even the slightest pericardial effusion or the thinnest strip of ascites were often subjectively considered pathological during this period (Table 2). After 2002, the diagnosis was based on an elevated peak systolic velocity of the middle cerebral artery of fetuses of immunized women. When peak systolic velocity of the middle cerebral artery increased above 1.5 MoM, a control was organized within 2 days and a transfusion was decided when a permanent elevation of peak systolic velocity above 1.6 MoM was confirmed. In the majority of these cases, no sign of hydrops could be observed (Table 2). Surveillance began especially early when there was a history of in utero death, early in utero transfusion or anemia in preceding pregnancies, or when there was a particularly high antibody titer. Before 2002, the mean fetal hemoglobin at the first in utero transfusion was 7.4±1.1 g/dL (pregnancies 1–8). After 2002, the use of peak systolic velocity of the middle cerebral artery made it possible to decide to perform this invasive procedure on the basis of more specific information: the mean hemoglobin was 4.3±2.0 (for pregnancies 9–17; P<.001).
The technical particularities of the 22 in utero transfusions are reported in Table 2. All the procedures were performed by the same team of physicians. In the 24 cases in which it was judged technically possible, direct access to the umbilical vein was attempted, either by a transplacental route for anterior placentas or by a transamniotic route, near the cord insertion. We most often used a 22-gauge to 25-gauge needle depending on gestational age, the caliber of the umbilical vein, and the conditions of vascular access reassessed immediately before the procedure. In case of failure to access the umbilical vein, no more than two attempts were made. The volume of blood to be transfused was calculated using charts established in our reference center (unpublished data) taking into account the estimated fetal weight, the initial fetal hemoglobin level, the target fetal hemoglobin level (less than fourfold the initial hematocrit level), and the hematocrit of the packed red cells. We usually checked fetal hemoglobin level using a hemoglobinometer when half of the calculated volume had been transfused to avoid blood volume overload.7 The total volume to be transfused could thus be adapted according to the efficiency of the initial transfusion.
Twenty-two intravascular in utero transfusions thus were performed successfully at or after 17 3/7 weeks of gestation. Median gestational age at the first in utero transfusion was 19 2/7 weeks (mean±standard deviation: 18 6/7 weeks±7.4 days).
In two cases, the first fetal transfusion was performed by the intraperitoneal route. In one case, fetal bradycardia occurred during the attempted vascular approach to the cord, at 17 3/7 weeks, and led to withdrawal of the needle before any blood could be sampled. Myocardial hypokinesia followed the bradycardia; its persistence led us to decide against attempting another vascular approach. An intraperitoneal transfusion was therefore performed. The in utero death observed at 18 weeks seems directly attributable to the first vascular attempt. In the second case, an in utero transfusion was planned because of the sharp rise of the peak systolic velocity of the middle cerebral artery at 17 2/7 weeks. The ultrasonographic observations indicated a vascular approach was likely to be impossible, and we began with an intraperitoneal transfusion. The peak systolic velocity of the middle cerebral artery did not decrease after the procedure, despite its apparent technical success, demonstrated by the ultrasonographic visualization of a fetal peritoneal effusion during the procedure. The first hemoglobin measurement from fetal blood at 18 2/7 weeks, 7 days after the intraperitoneal transfusion, before the intravascular transfusion, was 2.1 g/dL. This very low hemoglobin level, combined with the failure of the peak systolic velocity of the middle cerebral artery to fall after the intraperitoneal transfusion, suggested that this peritoneal approach was ineffective. Nonetheless, the Kleihauer-Betke test performed on the fetal blood sampled immediately before the in utero transfusion showed that 79.8% of the red blood cells were adult cells, therefore, from the intraperitoneal transfusion 7 days earlier.
The pregnancy outcomes are reported in Table 1. There were three fetal deaths during the study period, all in the same patient, in three successive pregnancies. The first death took place at 29 weeks, after four in utero transfusions. The first had been performed at 19 3/7 weeks, during severe fetal hydrops, and showed a fetal hemoglobin level of 1.8 g/dL. Despite correction of the anemia and three subsequent in utero transfusions at 19 6/7, 23, and 26 weeks, the fetal hydrops did not improve, and the fetus remained practically immobile. A medically indicated termination of pregnancy was considered, but the parents decided against it. Death occurred 3 weeks after the last in utero transfusion, although the peak systolic velocity of the middle cerebral artery no longer indicated any sign of fetal anemia. The other two in utero deaths, in this woman's next two pregnancies, seemed directly attributable to the technical difficulties of attempting a vascular approach for very early in utero transfusions at 17 3/7 and 17 2/7 weeks. One death, described previously in this article, occurred after an attempted intravascular transfusion, which was accompanied by fetal bradycardia and followed by an intraperitoneal transfusion. In the other case, the death took place immediately after an in utero transfusion was attempted by a transplacental vascular approach at 17 2/7 weeks, during which severe bradycardia occurred. The vascular approach failed and no transfusion could be performed. The in utero death, attributable to the attempted in utero transfusion, was noted the next day. These two cases are the only ones for which no fetal blood sample could be obtained.
The other 16 pregnancies that survived beyond 20 weeks resulted in 14 live births at a median gestational age of 36 0/7 weeks (mean±standard deviation 35 4/7 weeks±16 days), one ongoing pregnancy at 24 weeks of gestation, and one fetal death at 29 weeks (described previously, not attributed to any transfusion, attempted or completed).
During the study period, seven other pregnancies required a first transfusion between 20 and 22 weeks. These in utero transfusions did not entail any particular technical difficulty and no fetal loss occurred in this group. The survival rate in fetuses requiring a first in utero transfusion before 20 weeks was therefore 83.3% (95% confidence interval [CI] 58.5–96.4) and that of fetuses whose first in utero transfusion took place before 22 weeks was 88.0% (95% CI 68.8–97.5). The fetal loss rate attributable to an attempted in utero transfusion before 20 weeks was 8.0% in our series, all before 18 weeks. Before 22 weeks, the overall fetal loss rate per procedure was 6.3%, but nonetheless, no fetal loss attributable to transfusion by the vascular approach occurred between 18 and 22 weeks. Table 3 summarizes the results for survival and fetal loss associated with procedures before 22 weeks in our series and those reported in the literature.
In utero transfusion by the vascular approach has transformed the prognosis for red-cell alloimmunization and resulted in an overall survival rate on the order of 90%1,2 at the gestational ages when fetal anemia most generally occurs. In our center, the mean gestational age at which the first in utero transfusion is performed is 26 4/7 weeks (95% CI 21 3/7–30 4/7 weeks). Nonetheless, when anemia develops very early, especially before 20 weeks, the procedure can turn out to be technically difficult or even impossible by the vascular route. In our center, 11% of the patients with alloimmunization and anemia needed an in utero transfusion before 20 weeks, but they account for only approximately 6% of all the in utero transfusions performed.
Two principal publications have looked at early fetal anemia. Poissonnier et al and Yinon et al report 32 and 30 cases, respectively, of transfusions before 22 weeks.3,4 These two series report survival rates of 78% and 80%. In our experience, the principal technical difficulties are encountered before 20 weeks. In this situation, we had a survival rate of 83.3%, similar to that in other series, despite the earlier gestational age at the in utero transfusions. When we also included the cases that occurred from 20 to 22 weeks, the survival rate reached 88.0%, hardly different from that of in utero transfusions performed at later gestational ages.1,2
The risk of fetal loss associated with the procedure is higher before 20 weeks than after. In our series, the two fetal deaths associated with the procedure occurred during attempted vascular transfusions at a very early gestational age, less than 18 weeks. These two fetal deaths were associated with difficulties in accessing the umbilical cord rather than with blood volume overload as previously reported7 because no blood could be transfused in any of these two cases. The fetal loss rate was two of five (40%) for the attempted vascular approaches at terms of 18 weeks or younger, whereas there was no fetal loss associated with a procedure after 18 weeks. The lethality rate associated with the procedure before 20 weeks was 8.0% in our series. It fell to 6.3% before 22 weeks, almost identical to the rate reported by Yinon et al.4 In a series of a large number of in utero transfusions, Van Kamp et al1 reported a lethality rate per transfusion of 5.6% before 20 weeks but did not mention either the number of procedures performed at this term or the techniques used.
The vascular approach therefore seems to us reasonably secure above 18 weeks. Although cases of intravascular transfusion during fetal parvovirus B19 infections have been reported at 13 and 13 1/7 weeks and led to live births,8 the procedure remains dangerous. It seems likely that only the rare successes of ultraearly in utero transfusions have been published, whereas the failures are not publicized.
Of the other available routes, Yinon et al used the intrahepatic umbilical vein 19 times in 30 in utero transfusions, at gestational ages less than 22 weeks, and with a high success rate.4 The teams that regularly use an intrahepatic approach at more advanced gestational ages report success and complications rates similar to or slightly lower than those observed with a direct approach to the umbilical vein.1,9 We have no experience with this approach, but the similarity of the procedure-related complication rates in our series and in that of Yinon et al4 leads us to think that there is no evidence beyond the operator's experience for preferring one or the other technique.
Westgren et al10 proposed an intracardiac puncture when other techniques failed; this procedure had a high complication rate: 20%. In more recent series,3,4 this technique was used essentially in rare situations of very early severe anemia—from 15 to 16 weeks. The five intracardiac transfusions described in these two studies all resulted in fetal loss, probably associated as much with the severity of the situation as with the technique itself.3,4 Nonetheless, this vascular approach of last recourse appears to have a very high risk, and we think it should be avoided.
We used intraperitoneal transfusion when a direct vascular approach failed or appeared too dangerous from the outset. Other authors have used this route for very early transfusions, before 20 weeks. Howe et al6 performed intraperitoneal transfusions weekly in cases with a history of very early anemia in five fetuses from 15 weeks until a vascular approach became possible, at approximately 20 weeks. Outcome was favorable in all cases. Gallot et al11 reported one case with a favorable outcome after intraperitoneal transfusion at 14 weeks. Fox et al5 described intraperitoneal transfusions in seven fetuses between 15 and 18 weeks followed by vascular in utero transfusions from 20 to 22 weeks with an 87% survival rate. The principal criticism of these series involves the impossibility of measuring fetal hemoglobin before the procedure. In the studies that did not use peak systolic velocity of the middle cerebral artery measurements, it is impossible either to prove formally the reality of severe anemia or to demonstrate the efficacy of the peritoneal transfusion. In our series, an intraperitoneal transfusion performed after the vascular approach failed at 17 3/7 weeks was followed by an in utero death attributable to the failure to puncture the cord. In another case, the intraperitoneal transfusion was performed successfully as a first-line treatment of a sharply elevated peak systolic velocity of the middle cerebral artery at 17 2/7 weeks. The absence of any subsequent improvement in the peak systolic velocity of the middle cerebral artery initially cast doubt on its efficacy. The peak systolic velocity of the middle cerebral artery continued to rise, and an intravascular transfusion was performed 7 days afterward. The blood sample then confirmed very severe anemia with hemoglobin at 2.1 g/dL. Despite this very low hemoglobin level, the Kleihauer-Betke test performed on the sample before the in utero transfusion showed that 79.8% of the red blood cells present were adult cells and therefore came from the intraperitoneal transfusion. This result confirmed that the intraperitoneal transfusion had certainly helped to stabilize the fetal anemia and provide time until a vascular transfusion could reasonably be performed. Inversely, Howe et al showed an improvement in the peak systolic velocity of the middle cerebral artery after intraperitoneal transfusions, repeated in some fetuses, at early terms; nonetheless, no Kleihauer-Betke test was performed to confirm the efficacy of the procedure during the first vascular procedure.6
Classically, a diagnosis of very early anemia, before 20 weeks, must be suspected, particularly in cases with a history of in utero death, of in utero transfusion in the preceding pregnancy, or of exchange–transfusion at birth.12 Early management, if possible before the onset of fetal hydrops, is associated with improved survival.1 The ultrasonographic signs of fetal hydrops are rarely observed before 17 weeks, probably because of the immaturity of the endothelial system.3,13 On the other hand, peak systolic velocity of the middle cerebral artery measurements are very well correlated with the fetal hemoglobin level from 18 weeks onward14 and precede the onset of fetal hydrops.6 At earlier gestational ages, reference values for peak systolic velocity of the middle cerebral artery have been established in normal fetuses15 and for those with anemia due to fetal thalassemia.16 In our series, the performance of peak systolic velocity of the middle cerebral artery in predicting fetal anemia before 18 weeks seems excellent. All patients with an elevated peak systolic velocity of the middle cerebral artery above 1.6 MoM (from 44 to 51 cm/sec) for whom an intravascular puncture was possible did indeed have severe fetal anemia, identified before the appearance of any sign of fetal hydrops. Our observation that fetal hemoglobin levels were significantly lower since the introduction of peak systolic velocity of the middle cerebral artery when compared with fetuses previously diagnosed anemic on ultrasonographic signs of hydrops was an unexpected result. However, before the use of peak systolic velocity of the middle cerebral artery, the diagnosis of fetal anemia was suspected on the smallest signs of fetal hydrops such as a very limited pericardial effusion or strip of ascites. Most probably, such minor ultrasonographic signs were subjectively interpreted as pathologic in a context of severe alloimmunization, especially with a history of early fetal anemia, death, or both, but they were poorly specific.
In conclusion, fetal anemia before 20 weeks remains a situation at high risk of lethal complications compared with later gestational ages. The risk of early fetal anemia can be reliably monitored with peak systolic velocity of the middle cerebral artery, even before 20 weeks. Ideally management is based on intravascular transfusion when it is technically possible. In the case of foreseeable difficulties in a vascular access, especially before 18 weeks, intraperitoneal transfusion could provide the extra several days necessary to perform a vascular procedure more safely.
1. Van Kamp IL, Klumper FJ, Oepkes D, Meerman RH, Scherjon SA, Vandenbussche FP, et al.. Complications of intrauterine intravascular transfusion for fetal anemia due to maternal red-cell alloimmunization. Am J Obstet Gynecol 2005;192:171–7.
2. Carbonne B, Castaigne-Meary V, Cynober E, Gougeul-Tesnière V, Cortey A, Soulié JC, et al.. Use of peak systolic velocity of the middle cerebral artery in the management of fetal anemia due to fetomaternal erythrocyte alloimmunization [in French]. J Gynecol Obstet Biol Reprod 2008;37:163–9.
3. Poissonnier MH, Picone O, Brossard Y, Lepercq J. Intravenous fetal exchange transfusion before 22 weeks of gestation in early and severe red-cell fetomaternal alloimmunization. Fetal Diagn Ther 2003;18:467–71.
4. Yinon Y, Visser J, Kelly EN, Windrim R, Amsalem H, Seaward PG, et al.. Early intrauterine transfusions in severe red blood cell alloimmunization. Ultrasound Obstet Gynecol 2010;36:601–6.
5. Fox C, Martin W, Somerset DA, Thompson PJ, Kilby MD. Early intraperitoneal transfusion and adjuvant maternal immunoglobulin therapy in the treatment of severe red cell alloimmunization prior to fetal intravascular transfusion. Fetal Diagn Ther 2008;23:159–63.
6. Howe DT, Michailidis GD. Intraperitoneal transfusion in severe, early-onset Rh isoimmunization. Obstet Gynecol 2007;110:880–4.
7. Radunovic N, Lockwood CJ, Alvarez M, Plecas D, Chitkara U, Berkowitz RL. The severely anemic and hydropic isoimmune fetus: changes in fetal hematocrit associated with intrauterine death. Obstet Gynecol 1992;79:390–3.
8. Kempe A, Rösing B, Berg C, Kamil D, Heep A, Gembruch U, Geipel A. First-trimester treatment of fetal anemia secondary to parvovirus B19 infection. Ultrasound Obstet Gynecol 2007;29:226–8.
9. Nicolini U, Nicolaidis P, Fisk NM, Tannirandorn Y, Rodeck CH. Fetal blood sampling from the intrahepatic vein: analysis of safety and clinical experience with 214 procedures. Obstet Gynecol 1990;76:47–53.
10. Westgren M, Selbing A, Stangenberg M. Fetal intracardiac transfusions in patients with severe rhesus isoimmunisation. Br Med J (Clin Res Ed) 1988;296:885–6.
11. Gallot D, Boiret N, Vanlieferinghen P, Laurichesse H, Micorek JC, Berger M, et al.. The peritoneal route as a safe pathway for early in utero therapies: illustration by a 12-year follow-up after conservative management of severe Rhesus allo-immunization. Fetal Diagn Ther 2004;19:170–3.
12. Lobato G, Soncini CS. Relationship between obstetric history and Rh(D) alloimmunization severity. Arch Gynecol Obstet 2008;277:245–8.
13. Ryan G, Morrow RJ. Fetal blood transfusion. Clin Perinatol 1994;21:573–89.
14. Mari G, Deter RL, Carpenter RL, Rahman F, Zimmerman R, Moise KJ Jr, et al.. Noninvasive diagnosis by Doppler ultrasonography of fetal anemia due to maternal red-cell alloimmunization. Collaborative Group for Doppler Assessment of the Blood Velocity in Anemic Fetuses. N Engl J Med 2000;342:9–14.
15. Tongsong T, Wanapirak C, Sirichotiyakul S, Tongprasert F, Srisupundit K. Middle cerebral artery peak systolic velocity of healthy fetuses in the first half of pregnancy. J Ultrasound Med 2007;26:1013–7.
16. Lam YH, Tang MH. Middle cerebral artery Doppler study in fetuses with homozygous alpha-thalassaemia-1 at 12–13 weeks of gestation. Prenat Diagn 2002;22:56–8.