How I use noninvasive prenatal testing for red blood cell and platelet antigens Free

Alloimmunization during pregnancy is a condition in which a pregnant woman produces alloantibodies against antigens on fetal blood cells, which may lead to fetal blood cell destruction necessitating perinatal treatment. Alloimmunization can happen during pregnancy, as well as in cases of incompatible blood transfusions or other scenarios involving direct exposure to allogeneic blood cells, such as organ transplants or sharing needles during drug use.¹ 

In hemolytic disease of the fetus and newborn (HDFN) alloantibodies cause clinically significant destruction of red blood cells (RBCs); whereas in fetal and neonatal alloimmune thrombocytopenia (FNAIT), the alloantibodies lead to destruction of platelets, and probably also affect the endothelial cells. In the White population, severe antenatal disease may occur if the alloantibodies target the Rh (Rhesus) antigens: RhD; RhC; Rhc RhE; and, very infrequently, Rhe or the K (Kell) antigen. However, the distribution of RBC antigens varies among different ethnic populations. For example, in non-White populations, alloimmunization against RhD or K is much less frequent, whereas in Asian populations, antibodies against blood group antigens expressed by glycophorin A, such as anti-GPMur or anti-M, and anti-Di(a) cause clinically relevant problems.^1-4 In FNAIT, the alloantibodies target human platelet antigens (HPAs) expressed by fetal platelets but which may also be present on leukocytes, endothelial cells, and even placental tissue.⁵ In FNAIT, the implicated antigens are most often HPA-1a (80% of cases) or HPA-5b in the White population, but again, occurrence of disease and implicated antigens differ among populations with different ethnic backgrounds. For example, anti–HPA-4b in the Asian population and anti-CD36 in Asian and African populations can be involved.^5-7 

If the fetus carries the implicated antigen, this can lead to severe complications. In HDFN, these include severe fetal anemia, leading to immune hydrops fetalis and even fetal death. After birth, the newborn may need treatment for both anemia and hyperbilirubinemia to prevent occurrence of kernicterus.^1,3,4,8 In FNAIT, because of very severe thrombocytopenia and possible damage to endothelial cells, antenatal bleeding can occur with intracranial hemorrhage (ICH) being the most severe outcome, potentially resulting in irreversible brain damage or death.⁵ Postnatally, FNAIT can cause bruising, petechiae, and brain or organ bleeds, although most severe bleeding occurs before birth.⁵ 

Depending on the father’s zygosity, the paternally derived antigens on fetal blood cells may be present or absent in a pregnancy. If the implicated antigen is absent, further laboratory testing and clinical follow-up can be omitted and future parents can be reassured.¹ If the father is heterozygous for a certain blood group antigen, there is a 50% chance the fetus will be antigen positive. Before the turn of the century, clinicians relied on invasive techniques such as chorionic villous sampling or amniocentesis to collect fetal genetic material to perform genotyping. These invasive procedures carried a miscarriage risk of ∼1% and the risk of additional alloimmunization and boostering.^9,10 In 1997, Lo et al discovered that during pregnancy, from the end of first trimester, there is sufficient cell-free fetal DNA (cff-DNA) present in maternal plasma for genotyping.¹¹ The amount of cff-DNA increases during pregnancy,¹² but the fraction of cff-DNA compared with maternal circulating DNA remains low, usually a few percent up to 10% to 20%.¹³ The source of cff-DNA is trophoblast cells undergoing apoptosis, releasing in general fragments of a small size of <150 base pairs, although recently it was found that larger fragments of cff-DNA also circulate.^14-16 For a comprehensive overview of cff-DNA characteristics, including its potential applications in various testing methods and test design considerations using cff-DNA, we refer to the works of Chiu et al,¹⁷ Kjeldsen-Kragh and Hellberg,¹⁸ and Hyland et al.¹⁹ 

The first example of noninvasive prenatal testing (NIPT) for blood group antigen genotyping was the determination of the fetal RhD blood group.^11,20 Following this example, numerous other tests and assays were developed, allowing identification of fetal blood group genotype for most common antigens involved in HDFN.^21-33 Similar advancements have occurred for fetal HPA typing.^{27,29-31,34-37} However, NIPT for fetal RBC or HPA antigens has not yet been adopted by all clinical centers worldwide providing care to alloimmunized women. As suggested by Clausen and van der Schoot, this may be related to the advanced equipment necessary to perform the tests.³⁸ At the start of this century, NIPT for alloimmunized women was primarily developed within reference laboratories of academic or blood institutes^18,19 in Europe and Australia, which offered the tests internationally. It has not yet been widely included in guidelines for the management of alloimmunized pregnant women.^39-42 In Asia, fetal genotyping for clinically relevant populations is expanding.^42,43 The latest technology now allows laboratories specializing in NIPT testing for chromosomal abnormalities to offer this type of testing³³ more widely, including in the United States.

In 2022, a group of experts representing the cell-free DNA subgroup from the International Society of Blood Transfusion working party on red cell immunogenetics and blood group terminology published comprehensive recommendations for validation and quality assurance of NIPT for blood fetal groups, including for fetal RHD typing in alloimmunized women.⁴⁴ For further information on requirements for validation assays for routine fetal RHD typing in the screening setting (beyond the scope of this review) the reader is referred to a number of recent publications.^38,39,45 

In this article, we describe 3 representative case stories, 2 related to HDFN and 1 to FNAIT, to highlight important factors to consider when using NIPT in alloimmunized pregnancies. In addition, some key learning points on technical issues are addressed separately. The clinical monitoring and treatment of HDFN are discussed by Savoia et al⁸ in this How I Treat series.

A gravida 2, para 1 RhD-negative woman presented with a positive RBC antibody screen in the first trimester. This routine antibody screening was performed as part of the free-of-charge, nation-wide Dutch population screening program. Her first pregnancy was uncomplicated and no antibodies were detected in the first trimester or at 27 weeks of gestation. She had also received anti-D prophylaxis (Rh-immunoglobulin, 1000 IU; 200 μg) at 30-weeks gestation and immediately after birth, based on the routine screening RHD genotyping results obtained with the sample taken at gestational week 27.⁴⁶ There were no complications during or after birth. In the current pregnancy, antibody specification showed anti-D antibodies with a titer of 32, and a 30% antibody-dependent cell-mediated cytotoxicity test result. The antibody-dependent cell-mediated cytotoxicity test is used in the Netherlands to establish the hemolytic activity of the antibodies.⁴⁷ Both values were above the critical cutoff value used to identify a risk of severe HDFN, making further testing imperative.^1,3,8 Although there were no recognized high-risk events in the previous pregnancy, such as cesarean delivery or manual removal of the placenta, sometimes alloantibodies can still develop.^48,49 

An EDTA-anticoagulated blood sample was drawn from the mother and sent in for NIPT using cff-DNA to test for fetal RHD. The RHD NIPT test results showed no amplification of RHD sequences, whereas amplification of the fetal control marker hypermethylated RASSF1a (mRASSF1a) was above the test acceptance criteria to confirm that sufficient cff-DNA was present.⁵⁰ These 2 results combined, predicted an RhD-negative blood group of the fetus with no further need for laboratory or clinical monitoring. As an example, in Figure 1 the interpretation for the current diagnostic NIPT used in The Netherlands is visualized. This approach is in line with the recommendations of Clausen et al who strongly advocate the use of fetal DNA controls to confirm the presence of fetal DNA in cases of negative blood group genotyping result. The authors also mention the possibility to use 2 independent negative fetal typing results before concluding the fetal blood group status.⁴⁴ In next-generation sequencing (NGS)–based platforms, controls for so-called individual identification single-nucleotide polymorphisms can be combined with the fetal RBC or HPA typing in 1 single test, as reviewed by McGowan et al.³¹ The pregnant woman had an uncomplicated course for the remainder of her pregnancy and delivered a healthy newborn. A cord blood sample was used to confirm RhD negativity of the newborn.

In Western settings, anti-D alloimmunization mostly occurs because of a pregnancy with an RhD-positive fetus and not because of RhD-incompatible transfusions or organ transplantation. Before NIPT for fetal RHD typing was possible, serological testing of the biological father’s Rh phenotype was used to predict fetal RhD status when the mother had anti-D antibodies. The father’s RhDCcEe phenotype could suggest whether he was likely homozygous for RhD expression, allowing clinicians to assess fetal risk without invasive testing. However, variations in CE-D haplotype distribution across populations made this approach unreliable in non-White or mixed-ethnicity groups.^51,52 Currently, genomic methods to test for the presence of 1 or 2 copies of RHD are possible and could be used. However, one might now rely more on NIPT for fetal RHD typing, which may be the only possibility in cases involving assisted reproductive technology with a donor or if the biological father is not known, certain, or available.⁵³ 

As illustrated by Figures 1 and 2, NIPT with cff-DNA isolated from the mother’s plasma can be used in cases of RhD alloimmunization and is currently already used in many centers^18,38,41 However, not all centers using NIPT for fetal RHD testing in alloimmunized women base their antenatal management decisions on the fetal typing result. For example, in the Netherlands, if the test predicts that the fetus is RhD negative, no further laboratory testing or clinical monitoring is conducted.²⁴ When NIPT shows a positive fetal blood group, repeated anti-D titers are performed, and after the critical cutoff is reached, monitoring occurs via Doppler ultrasound at a specialized referral center.⁵⁴ 

A gravida 2, para 1 woman with an uncomplicated first pregnancy presented in her second pregnancy in gestational week 10 with anti-K (Kell) antibodies at first screen. The anti-K titer was 4, which was above the critical cutoff for detection of high-risk anti-K complicated pregnancies.^1,3,8 Priority was given to provide an additional blood sample from the mother for NIPT fetal K genotyping with cff-DNA. The test result indicated fetal K-positivity, and the immunohematology consultant of the laboratory immediately contacted the obstetric care provider to explain the implications of these results. When anti-K antibodies are found and the fetus carries the K antigen, there is a 50% chance that severe HDFN will occur.⁵⁵ The pregnant woman was thus examined as soon as possible at a highly specialized tertiary facility and arrangements for weekly close monitoring were made with the referring hospital (Figure 2). It is important that care is provided in centers in which pregnancies complicated by alloimmunization are seen regularly to ensure adequate and timely treatment of severe anemia of the fetus, as this yields the best outcome.^8,56 The woman was seen weekly for monitoring, and at 28 weeks of gestation, she required her first intrauterine transfusion (IUT). Thanks to a prompt referral, she was well-prepared for the procedure. After the initial IUT, 2 additional IUTs were necessary. A healthy newborn was delivered at 37 weeks of gestation. During the first 3 months of life, the newborn required 1 additional top-up transfusion because of the prolonged suppression of erythropoiesis associated with K-mediated HDFN.¹ Fortunately, children who experience HDFN typically have excellent long-term outcomes when treatment is initiated in a timely manner.⁵⁷ 

For blood group incompatibilities other than RhD, the decision to use cff-DNA testing upon identification of an RBC alloantibody depends on specificity of the alloantibody and whether the occurrence of RBC alloantibodies in women of childbearing age is prevented by matching RBC units. This approach aims not only to prevent the development of anti-D but also to prevent the formation of other types of RBC alloantibodies. In the Netherlands, RBC units have been matched for RhD for many years, for K since 2004, and for cE since 2011 for women of childbearing potential.⁵⁸ In other countries, it is common to match for D and, in some cases, for c and K.⁵⁹ Especially in cases of K immunization, it is crucial to determine whether the mother has ever received a nonmatched K blood transfusion. K-negative women who receive K+ blood have a relatively high risk of developing anti-K antibodies.^60,61 In the White population, the prevalence of K negativity is >90% and it is even much higher in other ethnic populations.^61,62 Hence, if a K-negative woman develops anti-K upon a K-positive blood transfusion, there is a very high chance she is pregnant by a K-negative partner. In low resource settings, paternal serological K typing might be a first approach to select only high-risk pregnancies. In our pregnant population the policy of K-matched transfusions for women aged <45 years has dramatically reduced the prevalence of anti-K and subsequently also halved the prevalence of K-mediated HDFN.⁶¹ Because in K-alloimmunization, fetal disease can occur early in pregnancy, we currently advise to perform NIPT for fetal K typing as early as possible without the need to first type the father (Figure 1). In serologically typed as K+ k− individuals (apparently homozygous K+), still ∼7% have a k allele encoding a low or absent expression of k also making identification of the relative rare homozygous K+ phenotype of less importance in our setting.⁶³ 

The K/k polymorphism is caused by only 1 single nucleotide variation.⁵¹ Early genotyping assays using real-time quantitative polymerase chain reaction (RQ-PCR) struggled with specificity of the test because of background amplification of the mother’s k allele, necessitating assay modifications that reduced the sensitivity and making repeated testing at later gestational age necessary.^21,24,64 These limitations have been solved using other technical platforms for genotyping.^{16,19,21-28,32} Both droplet digital PCR (ddPCR) platforms and techniques such as massive parallel sequencing (NGS) can be used early in pregnancy with very reliable results.^19,28,37,65 This approach further reduces the likelihood of false-negative results caused by rare instances of mispriming.⁶⁶ 

NIPT can also be used when antibodies against other Rh antigens are present: RhC, Rhc, and RhE. In our setting, for RhC and Rhc, fetal genotyping is performed without typing of the father when alloantibodies are found. However, the approach differs for anti-E antibodies. It is known that anti-E can occur “naturally” during pregnancy, without prior incompatible blood transfusions or pregnancy.⁶⁷ In every population, this concerns a high number of pregnant women, because >60% of women will be E negative, irrespective of their ethnic background.⁵² This is why, if the routine RBC screen early in pregnancy shows anti-E, it is reasonable first to type the partner and only if he has an E+ e+ phenotype, the fetus will be genotyped. If no samples from the biological father are available, NIPT fetal RhE genotyping is also performed.

Recent literature, including several meta-analyses and overviews, concludes that NIPT for blood groups can be reliably performed as early as gestational week 11.^18,38,45 In-house developed assays have been designed and validated at several reference centers, but for RHD typing commercial kits are also available.¹⁸ These commercial kits were initially developed for high-throughput fetal RHD typing to target anti-D prophylaxis in RhD-negative, nonalloimmunized women.¹⁸ However, both in-house developed and some commercially available tests can be used for fetal RHD typing in alloimmunized women.¹⁸ Most of these tests still use RQ-PCR, whereas more recently, assays based on ddPCR, NGS, or other technical platforms have been published.^25-33,68 Because samples from alloimmunized women are relatively scarce, it is a challenge to validate these assays for all blood group antigens with a range of samples, including those from early gestational age. In general, for diagnostic NIPT, experts recommend using controls to ensure the accuracy of the test and to confirm the presence of sufficient amount of amplifiable cff-DNA in cases of negative blood group genotyping results.⁴⁴ For male fetuses, a relatively simple and robust control is testing for Y chromosome sequences (sex determining region of the Y chromosome).^24,69 Alternatively, we, and others, have used sets of genetic markers that differ between the mother and fetus to confirm a sufficient cff-DNA fraction for conclusive results.^24,33,69 Although challenging, one can also use the universal fetal marker mRASSF1a.^31,50 In our experience, when robust tests for fetal blood group typing and fetal Y chromosome markers are combined with the universal mRASSF1a marker, only ∼4% of cases require an additional blood sample taken later in pregnancy because of insufficient fetal DNA concentration.⁷⁰ 

At the start of using NIPT for blood group antigens in alloimmunized pregnancies as part of clinical decision-making, the primary concern was the sensitivity of the test to ascertain a 100% negative predictive value. A false-negative NIPT result would imply risk of development of unnoticed severe disease or even fetal demise. Early studies showed that high levels of maternal DNA could affect the specificity of the test result, and the use of blood collection tube with a preservative to prevent decay of leukocytes is recommended.⁷¹ In some women, the cff-DNA concentration is extremely low, making testing for cff-DNA markers in diagnostic testing a prerequisite to prevent false-negative typing results.

In time, the impact of false-positive NIPT results has evolved. In the early years of blood group typing using NIPT, the risk of false-positive results was accepted. Although these would lead to unnecessary diagnostic testing and clinical follow-up during pregnancy and an induced delivery, this did not significantly harm the mother or the fetus because, the peak systolic velocity of the middle cerebral artery would generally remain below 1.55 multiple of the mean, and no cordocentesis would be conducted.⁷² However, a false-positive result could affect the timing and setting of the delivery, leading to iatrogenic prematurity. Moreover, with the development of new immunomodulatory agents such as FcRn blockers, there is an additional reason to prevent false-positive blood group antigen typing results, because they could lead to unnecessary initiation of antenatal treatment.^8,73,74 FcRn-blocking therapy should be started early in pregnancy, making it necessary to have reliable assays available at a gestational age of ∼11 weeks. The reviews of test performances are reassuring, showing that false negatives are extremely rare. Moreover, if the testing laboratory accounts for variations in RHD genes, these instances will become even rarer.^18,38,41 

False positives in RHD genotyping often result from the presence of RHD variant alleles. While most White individuals with an RhD-negative phenotype have a complete deletion of the RHD gene, RHD variants exist that result in a complete absence of RhD expression on RBCs while retaining large parts of the RHD gene.^51,52 Therefore, analyzing multiple RHD exons is essential. Some of these variants are more prevalent in different ethnic groups, leading to differences in the genetic background of the RHD genotype across the world.⁵² For instance, the RHD pseudogene (RHD∗08N.01), common in the Black African population, retains most of the gene but because of a 37–base pair insertion and a missense mutation in exon 6, it does not express the RhD protein on RBCs, making the mother susceptible to immunization.^51,52 To accurately determine the fetal RHD type, amplification of genetic sequences absent in the RHD pseudogene but present in the wild-type gene, particularly in exons 4 and 5, is necessary.⁵¹ Additionally, distinguishing between a fetus with an RHD pseudogene (who is RhD negative and will not suffer from HDFN because of maternal anti-D) and a DVI variant fetus (eg, RHD∗06.01, who will express fewer RhD molecules on their RBCs, but may theoretically suffer from HDFN) is critical to avoid false positives and unnecessary treatment. Similarly, it is important to detect fetal variants that lead to extremely low levels of RhD expression, such as variants such as RHD∗01EL.01 present in people from Asian descent, when designing test algorithms to personalize the start of IV immunoglobulin or FcRn-targeting therapies. Until comprehensive NGS-based typing becomes available, in the most genetically complex cases, we currently still rely on genotyping the biological father for RHD variants to collect additional information for correct prediction of the fetal RhD type.^51,75 

The variation in RHD-null alleles makes it very challenging to mitigate all risks of false-positive results.⁷⁶ Furthermore, in women carrying RHD-null alleles of which the molecular background is still not yet determined, it remains difficult to have assays available for routine typing other than NGS-based assays.

Overall, evidence is increasing that in various populations, NIPT for RHD typing is possible, especially if additional paternal testing is incorporated in the typing algorithm, as indicated by reports from Argentina, China, and Japan.^42,75,77 

A neonate was born after an uncomplicated pregnancy of a first-time mother (G1P0). The nurse noticed that the neonate showed several petechiae and significant bruising on the trunk and limbs. A full blood count was requested, revealing an isolated and severe thrombocytopenia of 10 × 10⁹/L, prompting further investigation. Cerebral ultrasonography revealed an ICH. Blood samples were taken from the mother, the biological father, and the neonate to assess the possibility of FNAIT. The mother was found to be HPA-1a negative, with the presence of anti–HPA-1a antibodies, whereas the father and neonate were both HPA-1a positive, showing an HPA-1ab genotype. The parents were informed that in future pregnancies, there is a 50% chance that the fetus will again be HPA-1a positive. Therefore, NIPT with cff-DNA to determine the HPA status of the fetus was recommended in any subsequent pregnancy.

One month later, the sister of this HPA-1a–negative mother attended a preconception counseling appointment to inquire about her own risk of having a child with FNAIT. In accordance with our protocol in such cases, HPA-1a typing was requested and she was also found to be HPA-1a negative. Additionally, because the HLA-DRB3∗0101 type is correlated with an increased risk of alloimmunization, this typing was also determined with a positive test result.⁷⁸ Given these findings, she is considered to have an elevated risk of developing FNAIT and was also advised to pursue noninvasive genotyping in any future pregnancies to determine the fetus’s HPA status.

FNAIT is a very rare condition and many aspects remain unclear. Different countries have varying guidelines on whether to treat suspected cases of FNAIT during pregnancy. Unlike HDFN, in which Doppler ultrasound can noninvasively assess whether the fetus suffers from anemia, there is no similar method for detecting fetal thrombocytopenia. The only direct approach, cordocentesis, is not recommended as a routine diagnostic tool because of the high risk of bleeding complications in the fetus with low platelet counts.^5,79 Currently, the primary treatment option for managing potential FNAIT cases is administering IV immunoglobulins to the mother (with or without steroids), aiming to reduce antibody levels in the fetus and thereby preventing bleeding complications.^5,79,80 Although IV immunoglobulin generally has mild side effects, usually only headaches, it may rarely lead to thrombotic complications, aseptic meningitis, renal failure, and hemolytic anemia.⁷⁹ Despite a calculated 98.7% success rate in preventing ICH in the systematic review from Winkelhorst et al, the quality of evidence is limited by inadequate control groups and the heterogeneity of the trials performed.^79,81 Currently IV immunoglobulin is thus still being used off-label. In some countries it is combined with corticosteroids.⁷⁹ To start timely antenatal treatment, NIPT for HPA genotyping of the fetus plays a crucial role in guiding management decisions (Figure 2).

In general, HPA antigens are encoded by single-nucleotide variation change. Similar to the RBC antigens, the first developed genotyping assays by RQ-PCR showed limitations in achieving conclusive test results at early gestational ages. However, ddPCR and NGS techniques make reliable typing from gestational week 11 onward possible.^{21,24,26-35,37,71} A recent review discusses the various platforms used for HPA-1a genotyping early in pregnancy.^35,82 

All key learning points from this paper are summarized in Table 1, outlining challenges and solutions encountered along the way.

Introduction of routine fetal RHD typing to target anti-D prophylaxis in RhD-negative women was at the start of this century an opportunity to obtain experience with NIPT and served as a driver to develop assays for fetal blood group antigen typing in alloimmunized women. The first available technology, RQ-PCR, proved to be a useful and a robust technical platform for reliable fetal RHD screening in nonimmunized women but had drawbacks in typing alloimmunized women for a broader panel of blood group antigens.^{18,21,24,38,41,45} However, new technologies such as ddPCR, and as most-promising platform, NGS-based typing, now allow for accurate diagnostic fetal typing as early as 11 weeks of gestation. These technologies incorporate sufficient controls to ensure a qualitatively correct test result based on sufficient amounts of cff-DNA, reducing the risk of false-negative results. The use of NGS-based typing is particularly valuable, because it allows for the analysis of multiple gene sequences that encode specific blood group antigens and individual identification single-nucleotide polymorphisms.³¹ Some countries continue monitoring pregnancies after negative cff-DNA results with titers and Doppler measurements. However, with the current accuracy, it has been proven that this is redundant.^83,84 Currently, some laboratories use assays without information on the cff-DNA concentration and if the test predicts an antigen-negative fetus, they opt to repeat testing at a later gestational age, extending the period of uncertainty and still carrying a risk of false negatives.⁸³ To use NIPT for RBC and HPA antigens in a clinical setting, we would like to reiterate the recommendations from Clausen et al, that clinicians should be aware of the usually low number of cases used to validate the assays and test performance related to false-negative and false-positive results.⁴⁴ Furthermore, clinicians should be informed about the laboratory’s strategy for confirming the isolation of sufficient cff-DNA. Finally, it is important that clinicians are informed about the detection of RHD variants and how the laboratory translates this information into predictions of fetal RhD-antigen positivity. When fetal genotyping is the qualifier to start therapy, such as IV immunoglobulin in FNAIT, and in rare cases in HDFN, but also for FcRn blockers in future, preventing false-positive results is essential.

The authors thank C. Ellen van der Schoot and Claudia Folman for their support in coreading the manuscript.

Contribution: R.M.v.t.O., E.J.T.V., and M.d.H. contributed equally to writing and editing of the manuscript.

Conflict-of-interest disclosure: E.J.T.V. is the principal investigator for a phase 2 trial (NCT03842189) and phase 3 trial (NCT05912517) of a new drug for the treatment of hemolytic disease of the fetus and newborn which is sponsored by Janssen Pharmaceuticals. M.d.H. discloses a consulting fee for Janssen Pharmaceutical as well as a research grant, sponsored by Janssen Pharmaceuticals. R.M.v.t.O. declares no competing financial interests.

Correspondence: Masja de Haas, Medical Affairs, Sanquin, Plesmanlaan 125, 1066 CX Amsterdam, The Netherlands; email: m.dehaas@sanquin.nl.