1. Introduction
International recommendations for provision of Rh immunoglobulin after first trimester bleeding events, such as induced, spontaneous and threatened abortion, vary by both gestational age and indication. Consensus is lacking because there is insufficient evidence to support or refute a need for immune-prophylaxis for Rhesus (Rh)-negative women in the first trimester [
1- Sperling J.D.
- Dahlke J.D.
- Sutton D.
- Gonzalez J.M.
- Chauhan S.P.
Prevention of RhD alloimmunization: a comparison of four national guidelines.
,
2- Chávez G.F.
- Mulinare J.
- Edmonds L.D.
Epidemiology of Rh hemolytic disease of the newborn in the United States.
,
3- Karanth L.
- Jaafar S.H.
- Kanagasabai S.
- Nair N.S.
- Barua A.
Anti-D administration after spontaneous miscarriage for preventing Rhesus alloimmunisation.
,
4Is Rh immune globulin needed in early first-trimester abortion? A review.
]. The American College of Obstetricians and Gynecologists (ACOG) states that giving a 50 mcg dose in early pregnancy loss “should be considered,” and recognizes the limited evidence we have to guide this clinical decision [
[5]American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 200: early pregnancy loss. Obstet Gynecol. 2018;132(5):e197–207.
]. Indeed, we do not yet know the earliest gestational age at which Rhesus-negative pregnant women can become sensitized to the Rh
o(D) antigen. Approximately 15% of women in the United States are Rh negative and as many as 35% of pregnancies end in abortion or miscarriage annually, with an additional 15% of ongoing pregnancies being complicated by some vaginal bleeding in the first half of pregnancy [
6- Garratty G.
- Glynn S.A.
- McEntire R.
- Retrovirus Epidemiology Donor Study
ABO and Rh (D) phenotype frequencies of different racial/ethnic groups in the United States.
,
7Jatlaoui TC. Abortion Surveillance—United States, 2016. MMWR. Surveillance Summaries. 2019;68.
,
8- Hasan R.
- Baird D.D.
- Herring A.H.
- Olshan A.F.
- Funk M.L.
- Hartmann K.E.
Association between first-trimester vaginal bleeding and miscarriage.
]. Screening of patients and prophylaxis with Rh immunoglobulin impose large systemic burdens on clinics and emergency departments. Patients paying out of pocket for abortion care bear an added financial burden, whether directly or amortized, for testing and receipt of a treatment with unclear benefit.
Rh-immunoglobulin works by decreasing the maternal exposure to Rh-antigens present on fetal red blood cells that enter the maternal circulation during bleeding events. Both the minimum Rh-antigenic exposure required for sensitization and immunization, and the natural history of Rh-antigen expression on fetal red blood cells (RBCs) throughout fetal development remain as gaps in knowledge in part because technologies have been insufficiently sensitive to quantify fetal RBC exposure at very low concentrations. Early studies of fetomaternal hemorrhage in the first trimester relied on the Kleihauer-Betke Technique (K-B), with an inadequate lower limit of detection of 4000 fetal RBCs per 10 million adult RBCs [
[9]- Pelikan D.M.
- Mesker W.E.
- Scherjon S.A.
- Kanhai H.H.
- Tanke H.J.
Improvement of the Kleihauer-Betke test by automated detection of fetal erythrocytes in maternal blood.
]. Additionally and critically, the K-B assay cannot distinguish between fetal RBCs and maternal F cells. Since maternal F cells are adult RBCs that express fetal hemoglobin, this limitation can lead to a false-positive K-B test result. Maternal F cell concentrations increase in patients with sickle cell disease, hemoglobinopathies and hereditary persistence of fetal hemoglobin, and during periods of physiologic stress, including early pregnancy. Maternal F cells do not cause Rh-sensitization because they are of maternal origin [
10- Dover G.J.
- Boyer S.H.
- Charache S.
- Heintzelman K.
Individual variation in the production and survival of F cells in sickle-cell disease.
,
11- Pembrey M.E.
- Weatherall D.J.
- Clegg J.B.
Maternal synthesis of haemoglobin F in pregnancy.
].
In order to fill these persisting gaps in scientific knowledge, we calculated the threshold for Rh sensitization, tested a flow cytometry protocol below that threshold, and then piloted it in a cohort of women before and after first trimester uterine aspiration for spontaneous or induced abortion. Given that quantification of fetal RBCs through direct measurement in maternal blood is very challenging, we turned to earlier studies conducted on subjects injected with fetal RBCs to estimate a plausible threshold concentration of fetomaternal hemorrhage sufficient to cause maternal sensitization [
[12]Transplacental isoimmunization by fetal red blood cells.
]. Extrapolating from these data, we estimated the level of sensitivity that an assay would be required to attain to detect fetomaternal hemorrhage during the first trimester. We then quantified the level of fetal RBCs in women undergoing uterine aspiration using a flow cytometry based assay, which is more sensitive and specific than the K-B assay. We relied on dual staining with hemoglobin F (HgbF) and carbonic anhydrase to distinguish fetal from adult RBCs [
12Transplacental isoimmunization by fetal red blood cells.
,
13- Gielezynska A.
- Stachurska A.
- Fabijanska-Mitek J.
- Debska M.
- Muzyka K.
- Kraszewska E.
Quantitative fetomaternal hemorrhage assessment with the use of five laboratory tests.
,
14- Umazume T.
- Yamada T.
- Morikawa M.
- Ishikawa S.
- Kojima T.
- Cho K.
- et al.
Occult fetomaternal hemorrhage in women with pathological placenta with respect to permeability.
], and further distinguished maternal F cells from fetal RBCs by the brightness of staining for HgbF. We further modified the RBC processing and staining conditions, cell input number and data analysis of the assay to improve its performance in the detection of rare events [
[15]Measurement of human erythrocyte CAI and CAII in adult, newborn, and fetal blood.
]. We then piloted this modified flow cytometry protocol in a cohort of women before and after first trimester uterine aspiration for spontaneous or induced abortion. We hypothesized that the concentration of fetal RBCs would not be clinically significantly increased following uterine aspiration in the first trimester.
2. Materials and methods
2.1 Threshold calculation
We estimated the threshold for Rh sensitization based upon our analysis of the existing literature [
12Transplacental isoimmunization by fetal red blood cells.
,
16Technical issues: flow cytometry and rare event analysis.
,
17- Jakobowicz R.
- Williams L.
- Silberman F.
Immunization of Rh-negative volunteers by repeated injections of very small amounts of Rh-positive blood.
]. A study by Zipursky directly injected 15 nulliparous Rh-negative women with repeated doses of 0.1 ml fetal RBCs [
[12]Transplacental isoimmunization by fetal red blood cells.
]. Although 11 (73.3%) of these women did not develop antibodies, two of the fifteen women (13.3%) showed an antibody response after receiving two doses of 0.1 ml given six weeks apart. Two more women developed antibodies after four and five total doses, respectively. Based on these data, we used 0.1 ml as the minimum volume of fetomaternal hemorrhage capable of causing sensitization [
18Krevans JR, Woodrow J, Nosenzo C, Finn R. Patterns of Rh Immunization. International Society of Blood Transfusion 1965 (vol. 23, pp. 781–781). Karger Publishers.
,
19Management of alloimmunization during pregnancy.
]. The average reported female blood volume in early pregnancy is 4000 ml [
20American College of Obstetricians and Gynecologists. ACOG practice bulletin No. 181: Prevention of Rh D alloimmunization. Clinical management guidelines for obstetrician gynecologists. Obstet Gynecol 2017;130(2):e57–70.
,
21- Zipursky A.
- Pollack J.
- Neelands P.
- Chown B.
- Israels L.G.
The Transplacental passage of foetal red blood cells and the pathogenesis of Rh immunisation during pregnancy.
]. Hence our target fetal RBC concentration threshold is 0.1/4000 = 250 fetal RBCs per 10 million total RBCs.
2.2 Flow cytometry
The detection of rare events is improved through the use of multiple antigens in flow cytometry. In order to distinguish maternal F cells from fetal RBCs and adult RBCs, we chose to use staining for both HgbF and carbonic anhydrase, using IQ Products’ Fetal Cell Count kit (IQ Products, Netherlands, Cat. No. IQP-363) [
22Blood volume changes in normal pregnancy.
,
23- Porra V.
- Bernaud J.
- Gueret P.
- Bricca P.
- Rigal D.
- Follea G.
- et al.
Identification and quantification of fetal red blood cells in maternal blood by a dual-color flow cytometric method: evaluation of the Fetal Cell Count kit.
]. Red blood cells from a nulliparous adult female (control subject who had an elevated maternal F cell fraction) and from second trimester cord blood were collected in K
2EDTA tubes and counted after washing twice with washing buffer provided in the Fetal Cell Count Kit (IQProducts, Cat# IQP-363). We modified the manufacturer’s protocol to adequately fix and permeabilize large numbers of cells needed to detect rare events of fetal RBCs. For the RBC experiments, 4.6 × 10
7 cells (control RBC or control RBC plus cord blood) were fixed for 45 min and permeabilized for 3 min and then stained with anti-human carbonic anhydrase conjugated to FITC and anti-human fetal hemoglobin conjugated to PE. In experiments with fetaltrol, commercially prepared fetal RBC + adult RBC mixture, R&D Systems Cat#FH101, FTL3, 1.8 × 10
7 fetaltrol cells were stained in a similar manner.
Next, we tested the modified manufacturer’s protocol for linearity and stability of fetal RBC and maternal F cell detection. For assay linearity, we serially titrated cord blood into adult nulliparous female blood in a 1:10 ratio to a concentration of 1:100,000. To test sample stability post-staining, the ratio between cord blood and control blood was adjusted to 1:10,000. The resulting cell mixtures were stained and then stored for 24, 48, and 72 h prior to running flow cytometry. Negative controls included running control RBCs (nulliparous adult female) without cord blood at all three time points. We also ran flow cytometry on commercially prepared fetaltrol as a positive control. Next, we tested the stability of fetal RBCs in samples that were incubated for different lengths of time prior to mixing, fixation and staining. Cord blood and adult blood were stored at 4 °C for 1, 3, 5, 7 and 13 days prior to staining. For the pre- and post-staining stability experiments, RBCs were kept at 4 °C in the dark until flow cytometry was performed. Up to 5 million events per sample were acquired on an LSRII flow cytometer (BD Biosciences, San Diego, CA) and analyzed using FlowJo (Treestar Inc, v. 10.0.8). For analysis, singlets were gated based upon forward vs. side scatter area (using a stringent gate), followed by visualization of HgbF vs. carbonic anhydrase staining on bivariate plots.
2.3 In vivo cohort
We performed an in vivo pilot study to assess the use of flow cytometry in a clinical cohort, approved by the University of Pennsylvania Institutional Review Board. We invited women undergoing uterine aspiration for induced or spontaneous abortion before 12 completed weeks gestation to participate. We excluded women with hemoglobinopathies or vaginal bleeding prior to enrollment. We collected demographic and clinical data on all participants. We performed phlebotomy before and after uterine aspiration. Rh status was determined for all participants and Rh-negative patients were treated with Rh immunoglobulin according to clinical protocol, following post-procedural phlebotomy. Samples were stained and run on the flow cytometer in batches within 72 h of collection. We chose a sample size of convenience for this exploratory study, reflecting the lack of reliable published estimates of antigen exposure in early pregnancy and the levels needed to induce Rh-sensitization. A comparison of fetal RBC and maternal F cell fractions before vs. after uterine aspiration was modeled using a repeated measures Poisson regression, including the total number of cells as an offset, and assuming a random intercept to account for clustering within each woman. All modeling was done in Stata v15 (STATA Corp., College Station TX). Data in figures were visualized using Prism v.8 software (GraphPad, San Diego, CA).
4. Discussion
In March, 2019, the National Abortion Federation (NAF) released new guidance stating that its members may consider foregoing Rh testing and immunoglobulin administration for patients undergoing induced abortion by uterine aspiration up to 8w0d and by medication abortion up to 10w0d [
]. The American College of Obstetricians and Gynecologists (ACOG) recommends giving a 50 mcg dose in early pregnancy loss, but reworded the recommendation from “should receive” to “should be considered” in the 2018 interim update, in recognition of the limited evidence we have to guide this clinical decision [
[5]American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 200: early pregnancy loss. Obstet Gynecol. 2018;132(5):e197–207.
]. Other healthcare organizations continue to recommend Rh immune-globulin administration to Rh-negative women with bleeding in the first trimester, while acknowledging insufficient evidence to guide management [
1- Sperling J.D.
- Dahlke J.D.
- Sutton D.
- Gonzalez J.M.
- Chauhan S.P.
Prevention of RhD alloimmunization: a comparison of four national guidelines.
,
19Management of alloimmunization during pregnancy.
]. This standard of care imposes large systemic burdens on clinics, emergency departments, and patients for an intervention with unclear benefit. The paucity of evidence to inform such recommendations is underscored by the variation in international obstetrical society guidelines for provision of Rh immunoglobulin at <12 weeks gestation for miscarriage, threatened abortion and molar pregnancies [
[1]- Sperling J.D.
- Dahlke J.D.
- Sutton D.
- Gonzalez J.M.
- Chauhan S.P.
Prevention of RhD alloimmunization: a comparison of four national guidelines.
]. Epidemiologic data suggest that there are clinical situations in early pregnancy where Rh immunoglobulin is unnecessary. Wiebe and colleagues compared databases in Canada, where Rh immunoglobulin is routinely given for all vaginal bleeding events prior to 12 weeks gestation, and The Netherlands, where Rh immunoglobulin is not given before 10 weeks in early pregnancy loss nor before 7 weeks in induced abortion [
[25]- Mark A.
- Foster A.M.
- Grossman D.
- Prager S.W.
- Reeves M.
- Velásquez C.V.
- et al.
Foregoing Rh testing and anti-D immunoglobulin for women presenting for early abortion: a recommendation from the National Abortion Federation's Clinical Policies Committee.
]. They found a statistically significantly higher rate of antibody detection in Canada 4.21 per 1000 (95% CI: 4.12–4.30) vs. The Netherlands 4.03 per 1000 (95% CI: 3.93–4.12). Such indirect evidence underscores the importance of developing clinical metrics for evaluating the utility of Rh immunoglobulin in early pregnancy.
We tested a flow cytometry assay that is sensitive enough to quantify fetal RBC frequencies below 100 in 10 million total RBCs. Because there can be delays in getting samples processed by the lab (for example, if a sample arrives after hours or on a holiday or weekend), understanding the stability of the assay for clinical use is important. The fetal cells showed no decrease in concentration in samples stored prior to staining, even at 13 days, suggesting that fetal cells are stable in whole blood. Further laboratory studies that control specifically for ABO compatibility and type would strengthen these findings. Fetal RBCs are more resistant to degradation than adult RBCs, so overestimation of fetal RBCs is more likely than underestimation. However, multiple factors affect the clearance of fetal RBCs from maternal circulation, including the ABO blood type. The optimal timing of sample collection following fetometernal hemorrhage to capture maximum exposure to fetal RBCs is not known, but is likely within three hours [
[26]- Wiebe E.R.
- Campbell M.
- Aiken A.R.
- Albert A.
Can we safely stop testing for Rh status and immunizing Rh-negative women having early abortions? A comparison of Rh alloimmunization in Canada and the Netherlands.
].
Dual-color flow cytometry has an additional advantage over the K-B and single-color immunophenotyping methods of being more specific, as it can distinguish fetal RBCs from maternal F cells based upon carbonic anhydrase staining [
[22]Blood volume changes in normal pregnancy.
]. Maternal F cells cannot cause maternal sensitization, so their concentrations are clinically irrelevant, but their inadvertent detection and misclassification can lead to falsely elevated results with the K-B method or a flow cytometry assay that lacks anti-carbonic anhydrase antibodies. A proportion of our samples revealed maternal F cells above the K-B detection threshold of 4000 per 10 million RBCs, and we surmise that earlier studies using K-B may have detected maternal F cells rather than fetal RBCs. This is plausible, given that distinguishing fetal RBCs from adult RBCs that express fetal hemoglobin has high interobserver variability and low reproducibility [
27Mollison's blood transfusion in clinical medicine.
,
28- Nance S.J.
- Nelson J.M.
- Arndt P.A.
- Lam H.T.
- Garratty G.
Quantitation of fetal–maternal hemorrhage by flow cytometry: a simple and accurate method.
]. Additionally, K-B has historically been used to assess the F cell concentration in non-pregnant patients with sickle cell disease, hemoglobinopathies, hereditary persistence of fetal hemoglobin and their responses to medications, physiologic stress and pregnancy [
[29]50 years Kleihauer-Betke Test.
].
We found a statistically significant increase in the concentration of fetal RBCs present in the maternal blood after uterine aspiration, suggesting that the procedure does in fact result in the transfer of low numbers of fetal RBCs into the maternal circulation and that our flow cytometry assay is sufficiently sensitive to quantify this change. In spite of the increase in fetal RBCs, this flow assay also shows that the concentration of fetal RBCs post-procedure was clinically insignificant as it remained below the reported threshold for sensitization (250 fetal RBCs per 10 million total RBCs) in all study participants [
[12]Transplacental isoimmunization by fetal red blood cells.
]. In fact, even if we assumed that the level of fetal RBCs in the population were 195 per 10 million cells, the probability of seeing a woman with 250 or more fetal RBCs per 10 million cells is less than 0.0001 (0.01%) (see Methods for calculations). Nearly half of our participants exhibited fetal RBCs pre-procedure (18/37), consistent with Hollenbach et al., who found that 60% of their participants at gestational ages 6–22 weeks exhibited fetal RBCs pre-procedure [
[30]- Hollenbach S.J.
- Cochran M.
- Harrington A.
“Provoked” feto-maternal hemorrhage may represent insensible cell exchange in pregnancies from 6 to 22 weeks gestational age.
]. We excluded women with any vaginal bleeding, while 26% of the Hollenbach cohort reported some prior bleeding; however, there was no correlation in their study between bleeding and presence of fetal RBCs. Additionally, we used a flow cytometry protocol capable of not just detecting fetal RBCs and distinguishing them from maternal F cells, but also capable of separately quantifying maternal F cells.
There are several limitations to our study. The threshold calculation was made using small studies that would be logistically and ethically challenging to reproduce today. Human immune systems are complex and heterogenous, and individual responses to antigenic exposure may vary in subtle ways including by pregnancy status, ABO compatibility/incompatibility, genetic predisposition and timing and level of exposure. There are also limits to our understanding of fetal RBC development throughout gestation and the natural history of maternal F cell production in pregnancy. Additional antigens that distinguish fetal RBCs from adult RBCs would be helpful in the removal of various sources of noise and false positives that can obscure rare event detection by flow cytometry, and is an important future research direction. Additionally, our cohort study is limited by its small sample size. Larger studies powered to detect differences by gestational age, procedural and pregnancy characteristics, and patient demographics are warranted.
This study also has several strengths. The threshold calculation was performed using the most conservative published inputs. The flow cytometry protocol was tested using second trimester cord blood, which contains fetal RBCs that do not yet express carbonic anhydrase. The protocol is capable of separately quantifying fetal RBCs and maternal F cells to very dilute concentrations.
Our study suggests that fetal RBC exposure in the first trimester is insufficient for maternal Rh-sensitization. Larger studies of pregnant women are warranted to identify both the earliest gestational age and the clinical circumstances under which fetal RBC exposure warrants Rh immunoglobulin treatment. This evidence can then be used to inform universal guidelines for Rh immunoglobulin in the first trimester, reduce cost and improve care.
Article info
Publication history
Published online: March 02, 2020
Accepted:
February 24,
2020
Received in revised form:
February 24,
2020
Received:
August 11,
2019
Footnotes
☆Declaration of competing interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Copyright
© 2020 Elsevier Inc. All rights reserved.