Obstetric and Newborn Weak D-Phenotype RBC Testing and Rh Immune Globulin Management Recommendations: Lessons From a Blinded Specimen-Testing Survey of 81 Transfusion Services Open Access

This study presents results of an educational exercise devised by the CAP Transfusion, Apheresis, and Cellular Therapy Committee (formerly known as the CAP Transfusion Medicine Resource Committee) and provided to laboratories participating in the CAP Expanded Transfusion Medicine Exercise program (2017-B mailing, October 2017). Laboratories were ineligible for inclusion if they did not routinely support obstetric deliveries (≤5 per year) or did not submit any test results.

Reagent RBCs were obtained from a single group A donor selected to have an RhD phenotype negative or weakly reactive (≤2⁺) in the immediate-spin (IS) phase and strongly reactive in the antiglobulin test (AGT) phase, meeting the CAP-AABB-ACOG work group definition for serologic weak D phenotype.¹⁴ RHD variant genotyping was not available on the anonymized donor. Participating laboratories received 2 specimens containing 5% RBC suspensions from this donor's sample. Each of the 2 challenge samples had a clinical description. One specimen case history was a 24-year-old pregnant woman at 30 weeks gestation with placenta previa who needed 2 units of RBCs crossmatched (referred to as “maternal”). The other specimen case history was a newborn of a 32-year-old RhD-negative mother (referred to as “newborn”). These 2 samples were not described as being related to one another.

Laboratories were asked to perform RhD typing by their routine manual tube procedures and to respond to questions related to their evaluation and reporting of the results obtained. They submitted information about their reagent vendor, the IS reaction strength, and, if their procedures called for performing an AGT, its reaction strength. The laboratories were asked how they would report the RhD results to each patient's chart from the following 9 options: RhD-negative; RhD-negative weak D-negative; RhD-negative weak D-positive; RhD-indeterminate (or their laboratory's equivalent); RhD-indeterminate weak D-negative; RhD-indeterminate weak D-positive; RhD-positive weak D-negative; RhD-positive weak D-positive; and RhD-positive. The participants were also asked whether they used the term D^u either instead of or in addition to weak D. Participating laboratories provided basic demographics about their institutional characteristics and the number of annual obstetric deliveries.

To analyze the participant results, we considered IS-negative, AGT⁺ reactions, and IS weak-2⁺ reactions with or without AGT as weak D-reactive. To determine how laboratories interpret results, participants were asked about the clinical recommendations they would provide. For the maternal sample, participants were asked whether they would crossmatch RhD-negative RBCs and whether their institution would give RhIG to the mother if her baby typed 4⁺ for RhD. For the newborn, participants were asked whether and how they would screen his/her RhD-negative mother for excess FMH. Laboratories were asked their policies for RHD genotyping in obstetric patients with serologic weak or discrepant RhD typing.

Statistical analyses used 2-tailed Fisher exact tests for categorical frequencies and t tests for comparing annual delivery volumes. A significance level of P < .05 was used for comparisons. Analyses were performed with SAS 9.4 (SAS Institute, Cary, North Carolina).

Ninety-three laboratories subscribed to the CAP Expanded Transfusion Medicine Exercise program, but 12 were ineligible for this analysis because they had 5 or fewer annual deliveries (n = 11) or did not provide any data (n = 1), leaving 81 laboratories included (Table 1). Seventy laboratories (86%) (percentages rounded) were in the United States (n = 68) or Canada (n = 2). Sixty-three laboratories provided the annual number of deliveries at their hospitals, totaling 172 371 (mean, 2736; range, 300 to [in 1 facility] >9999). Because some participants did not answer all questions, the denominators of the responses were variable.

There was a diverse array of serologic reaction results, RhD typing interpretations, and RhIG recommendations employed. Seventy-nine laboratories provided results on the maternal challenge sample (Table 2). Among the 57 laboratories that obtained an IS-negative result, 41 (72%) stopped there and reported the type as RhD-negative, and 16 (28%) proceeded to AGT phase, in which all reactions were positive. Of 20 laboratories that initially reported weak-2⁺ IS reactions, 9 did not proceed to AGT and 11 did (all positive). Two laboratories obtained 3 to 4⁺ IS results and reported the serologic type as RhD⁺.

The 36 laboratories that obtained IS-negative, AGT positive results (n = 16) or weak-2⁺ IS results (n = 20) on the maternal challenge sample (serologic weak D phenotypes) collectively reported 9 different RhD interpretations. In addition to RhD-negative (n = 6), RhD-negative weak D-positive (n = 3), RhD⁺ weak D-positive (n = 7), and RhD⁺ (n = 13), there were 5 other combinations from 7 laboratories that included RhD-indeterminate or D^u+ terms as follows: RhD-negative D^u+, RhD-indeterminate, RhD-indeterminate weak D-positive, RhD-positive D^u+, and RhD-positive weak D-positive D^u+ (Table 2).

For the newborn challenge specimen, the proportions of IS-negative and IS-positive results were similar to the maternal specimen reports (Table 3). However, as expected, 57 of 60 laboratories (95%) with IS-negative reactions performed AGT to identify a weak D-reactive newborn (54 AGT⁺, 3 AGT-negative).

Intralaboratory serologic testing reproducibility could be examined because laboratories unknowingly tested the same RBCs twice for the “maternal” and the “newborn” specimen. Of the 80 laboratories testing both specimens, 69 (86%) reported the same overall IS results (53 positive, 16 negative). However, 11 laboratories (14%) had discrepant IS reactions (Table 4). Seven laboratories reported the maternal RBCs as IS-negative and the newborn RBCs as weak⁺ to 3⁺ IS positive, and 4 reported the maternal RBCs reacting weak-1⁺ and newborn RBCs negative. The “maternal-IS positive” laboratories were more likely to obtain a discrepant newborn result (7 of 23; 30%) compared with the “maternal-IS negative” laboratories (4 of 57, 7%, P = .01).

For the maternal challenge specimen, 13 of 79 responding laboratories (16%) would not have given the mother RhIG based on their testing and treatment procedures (Table 2). Ten laboratories had IS-negative AGT⁺ or weak-2⁺ IS results and interpreted the RhD typing as RhD-positive (n = 9) or D⁺ weak D⁺ (n = 1). The other 3 laboratories had strongly 3 to 4⁺ IS RhD-positive reactions (n = 2) or IS-negative testing with no AGT (n = 1). If the 41 laboratories with IS-negative, no-AGT results (conventional RhD-negative) and the 2 laboratories with 3⁺ to 4⁺ IS results (conventional RhD-positive) are excluded, 10 of 36 laboratories (28%) that sought and reported maternal serologic weak D phenotypes would not have recommended RhIG.

Laboratories typing the maternal specimen as IS-positive were less likely to give RhIG (15 of 22; 68%) compared with laboratories reporting an IS-negative result (51 of 57; 89%; P = .04) (Table 3). Within the laboratory group that reported the specimen to be IS-negative, the 16 laboratories that also did AGT (all AGT⁺) were less likely to give RhIG (11 of 16; 69%) compared with the 41 laboratories that did not perform AGT (40 of 41; 98%; P = .006) (Table 3).

Whether the laboratories would crossmatch RhD-negative RBCs for the maternal patient usually correlated with whether they would also give RhIG (75 of 79, 95%). Of 4 exceptions, 3 laboratories would give RhD-negative RBCs but not RhIG, and 1 would give both RhD-positive RBCs and RhIG. The likelihood of giving RhIG to the mother was not associated with the laboratory nationality (US versus non-US laboratory location) as follows: RhIG, 55 of 68 (81%) US versus 11 of 11 (100%) international (P = .19); or the following annual delivery volume: RhIG (n = 54), mean (SD) 2708 (1950), versus no-RhIG (n = 9), mean 2902 (3467) (P = .81).

For the newborn specimen, 74 of 79 responding laboratories (94%) would have given the mother RhIG based on their typing results (Table 3). Three laboratories typed the neonate's RBCs as IS-negative, AGT-negative for a false-negative AGT rate of 3 of 57 (5%). They also stated they would not have given RhIG. Of note, 3 other laboratories with IS-negative results did not perform AGT to attempt to identify a weak D-reactive newborn as is required in the CAP Laboratory Accreditation Program Transfusion Medicine Checklist (TRM.40780) and the AABB Standards for Blood Banks and Transfusion Services (standard 5.30.2), although these laboratories stated they would have given RhIG to the mother (Table 3).^19,20 

RhD serologic reactions varied by anti-D reagent manufacturer (Table 4). When testing the maternal specimen, laboratories using the 2 most common reagent vendors A and B had statistically different IS⁺ frequencies of 15 of 37 (41%) versus 1 of 21 (5%) (P = .003). However, the RhIG recommendation rates for laboratories using vendors A or B were not statistically different (27 of 36 [75%] versus 19 of 20 [95%], respectively, P = .08). The frequencies of IS result discrepancies between the maternal and the neonatal specimens did not vary by reagent vendor.

In 73 responding laboratories, 33 (45%) would screen for excess FMH in the mother of the newborn by Kleihauer-Betke testing, 6 (8%) by flow cytometry, and 34 (47%) by the fetal rosette test (Table 5). In the 64 laboratories reporting a neonatal serlogic weak D phenotype (IS negative to ≤2⁺ and AGT⁺), 29 (45%) would have performed a fetal rosette test.

The laboratories were not asked which fetal rosette test vendor they used. The US-based laboratories had 2 main commercial vendors. The Immucor FMH RapidScreen test (Immucor, Norcross, Georgia) instructions state “when the D antigen on the infant's RBCs requires a weak D test for detection, a test to detect FMH based on fetal hemoglobin is recommended.”²¹  Twenty-five of 52 laboratories (48%) with IS-negative, AGT-positive reactions would have used the fetal rosette test. According to the Ortho FETALSCREEN II test instructions (2015 and current 2018 versions; Ortho Clinical Diagnostics, Raritan, New Jersey), “when it is known that the infant is weak RhD positive, a test based on fetal hemoglobin is recommended.”²²  Six of 26 laboratories (23%) with “weak D⁺” in their RhD interpretation would have used the fetal rosette test.

Of 73 facilities providing information on their RHD genotyping practices for pregnant women with weakly reactive or discrepant serologic RhD typing, 39 (53%) recommended RHD genotyping in selected cases, 5 (7%) in all cases, and 29 (40%) did not recommend RHD testing.

Transfusion services are highly accurate in performing RhD typing on conventional RhD⁺ and RhD-negative specimens. Laboratories performing CAP proficiency testing routinely achieve ≥99.5% concordance with expected results.²³  The US Centers for Medicare and Medicaid Services require ABO and RhD blood typing proficiency testing to be 100% accurate (5 of 5 results in each survey), compared with 80% for other regulated analytes.²⁴  However, this study shows that weak D-reactive RBCs pose extra challenges in testing and interpretation for many laboratories and raises several critical concerns about transfusion service laboratory practices (Table 6). In the absence of RhD genotyping information, from 25% to 91% of weak D-reactive women in the US have RHD variants posing potential risk of anti-D formation, depending on the race and ethnicity of the patient population.^2,3  A significant percentage of laboratories in this study demonstrated inconsistent RhD typing results, potentially misleading RhD terminology, and RhIG eligibility practices that increased the risk of maternal RhD alloimmunization.

To avoid the issues found by this study, the key recommendations are as follows: (1) Do not routinely perform maternal antiglobulin anti-D typing when the initial D typing is negative; (2) avoid outmoded or misinterpretable RhD terminology (D^u, indeterminate, weak D-positive); and (3) for D-negative mothers, always perform antiglobulin anti-D typing on initially D-negative newborn specimens; always give RhIG when the newborn is weak D-reactive; and use quantitative testing for excess FMH (Kleihauer-Betke technique or flow cytometry) when the newborn is weak D-reactive. These points are elaborated below.

For the maternal specimen, the frequency of IS⁺ reactions across 4 anti-D reagent vendors used in multiple laboratories varied from 5% to 41%. Each reagent vendor uses different monoclonal antibodies that yield well-known disparate reactions to partial-D RBC variants with modified epitopes.²⁵  In early research studies, anti-D antibody clones grouped in the same categories of epitope patterns among partial-D variant RBCs gave diverse reaction patterns in panels of weak D-variant RBCs.²⁶  In addition to monoclonal specificities, weak D RBC reactions were affected by RhD zygosity and the RhD-suppressive effect of Rh C antigen. More recently, Lucacevic Krstic et al²⁷  tested genotyped weak D obstetric specimens with 5 international commercial anti-D tube reagents. The rates of direct agglutination among these reagents varied from 54% to 100% versus 24 weak D type 1 RBC specimens and from 83% to 100% versus 18 weak D type 3 RBC specimens. These results and our findings indicate that cross-laboratory consensus in weak D typing is challenging. None of the currently approved US providers of immunohematology proficiency testing supply surveys devoted to weak D RBC typing.

Also notable was the 14% intralaboratory rate of discrepant IS reactions between concurrent typings of the same RBCs. Denomme et al²⁸  observed that in weak D specimens, small differences in manual tube typing techniques can contribute to variable results when testing is repeated on the same sample. The low frequency of weak D-reactive patients, less than 0.4% in most US and Canadian series, may also contribute to inconsistency when such specimens are encountered.^1,3,4,6,11 

Twenty-eight percent of laboratories that obtained an IS-negative reaction on the maternal specimen also performed an AGT. This frequency was similar to a 2012 CAP practice survey, when 30% of more than 3100 laboratories performed AGT for IS-negative pregnant women.¹⁶  All Food and Drug Administration–approved anti-D tube-method IgM/IgG reagents are expressly designed not to be extended to AGT in obstetric patients. These reagents are required to give a negative IS result with category DVI RBCs, the most common partial-D variant, so that those mothers and transfusion recipients are typed as RhD-negative and receive RhIG therapy and RhD-negative RBCs. In our exercise, laboratories were instructed to follow their normal procedures for tube testing, but the accompanying questions about weak D-reactive RBCs may have led some to seek a weak D phenotype. However, like the IS⁺ reactions, the IS-negative, AGT-positive results also were associated with a lower rate of RhIG treatment.

In our study, the complexities of comparative serologic reactions for weak D-reactive RBCs were compounded by the 9 different ways the RhD interpretations were expressed. Use of nonstandard terminology is confusing and hinders communications and medical judgment. The words “indeterminate” and “D^u+” in patient reports, employed by 7 of 79 laboratories here (9%), should be discontinued. Table 7 shows recommended immunohematologic terminology for serologic weak D phenotypes and discrepant D phenotypes. The term “weak D-reactive” helps convey this is a serologic phenotype, not a clinical recommendation, but it should not supplant the blood type of record.

Although not a limitation, it is important to note that this exercise was conducted by tube method. Many laboratories use other methods (gel, microplate), so the results should be taken in context as the findings of one method are not directly applicable to another. Table 8 provides suggested working definitions for weak D phenotypes in gel and microplate methods. These definitions may be refined as more data become available on RHD genotyping correlations.

For patient reports, obstetricians and other providers need to understand whether patients are considered RhD-positive or RhD-negative for a transfusion and pregnancy management perspective, and these terms should be used accordingly and in conjunction with RHD genotyping (Table 9). Transfusion services are also discouraged from using the terms “weak D-positive” or “weak D⁺” in medical record reports to avoid inappropriate clinical management of RhIG eligibility. This approach is consistent with recent guidelines from CAP-AABB-ACOG work group members to report the RhD blood type as RhD-positive or RhD-negative according to clinical management indications, aided by routine use of RHD variant genotyping to help determine risk for anti-D formation.⁸ 

Transfusion services should also consider management of the RhD type of weak D-reactive patient information in their information systems. This includes not only how the results are displayed but also what algorithms are used to control the Rh types of blood component transfusions. Information system vendors could assist in this regard by using appropriate language and decision protocols and moving away from outdated terminology.

In addition to the interpretive variability and use of nonstandard terminology among participants, the survey identified disparate practice recommendations for RhIG administration. The AABB Technical Manual advises that weak D-reactive women should be candidates for RhIG unless their RHD genotype is determined not to be at risk for alloimmunization.²⁹  Detailed recommendations on RHD genotypes and RhIG eligibility were provided by the CAP-AABB-ACOG work group.^8,14 

In this study, 28% of laboratories (10 of 36) would not have recommended RhIG for a weak D-reactive mother of an RhD⁺ newborn. Among these 10 laboratories, 5 reported IS-negative AGT⁺ typing and 5 obtained IS reactions 2⁺ or less (Table 2). In past years, weak D-reactive obstetric patients were often considered RhD⁺ and not candidates for RhIG.^{7, 15,16}  However, severe and even fatal Rh hemolytic disease of the fetus and newborn has occurred in alloimmunized weak D-reactive women.³  It also should be noted that an unusually large FMH from an RhD⁺ fetus before delivery may resemble a serologic weak D phenotype in a D-negative mother, although mixed-field typing would be expected.³⁰ 

Similarly, disparate practices were identified by the survey in FMH testing methods. The fetal rosette test is contraindicated to screen for excess FMH when the newborn's RBCs are weak D-reactive.³¹  The monoclonal anti-D reagents used in fetal rosette tests may not react well with weak D-reactive RBCs, causing false-negative results. However, 47% of laboratories would have used the fetal rosette test for the mother of the newborn with weak D-reactive RBCs in our exercise. Weak D-reactive RBCs have stimulated anti-D after transfusion to RhD-negative patients, and severe anti-D hemolytic disease of the fetus and newborn has been reported in weak D-reactive babies.^32–34  Adequate RhIG coverage is needed after a weak D-reactive newborn delivery. The relative infrequency of weak D-reactive newborns in D-negative mothers may lead to less familiarity with proper procedures. All laboratories performing RhD typing on newborns for maternal RhIG therapy should address weak D-reactive newborns in their standard operating procedures and staff training, to ensure that quantification of fetal RBCs is performed on the mother's postpartum specimen instead of fetal rosette testing.

This study has several limitations. The laboratory sample size (n = 81) was relatively small. However, some aspects of our findings have been seen in past CAP surveys, including maternal RhD typing in AGT phase after negative IS¹⁶  and frequent omission of RhIG in weak D-reactive mothers.¹⁵  Clerical errors can occur when laboratories transfer test results or other data to survey response forms. The RHD genotype of the RBC donor used in this study would have been of interest but was not available. However, the RBC serologic reactions were typical for weak D-reactive RBCs, and our findings for RhD interpretations and RhIG management would be generally applicable for any weak D-reactive RBC specimen of unknown genotype. Many laboratories in this study likely performed automated RhD typing methods as their primary technique, which may have limited their experience with tube-based serologic testing of weak D-reactive RBCs. This might have accounted for some of the testing issues observed. On the other hand, a similar exercise with weak D-reactive RBCs across automated methods with gel matrix and microplate techniques would probably reveal even more serologic and interpretative complexity. This study was conducted after the first CAP-AABB-ACOG recommendations in 2015¹⁴  but before their second set of recommendations in 2020,⁸  so further implementation of best practices may have occurred since then.

Sixty percent of the surveyed laboratories were obtaining RHD genotyping on some or all of their obstetric weak D-reactive patients. This provides evidence that recent recommendations are becoming more widespread, but further adoption is encouraged.

In conclusion, the results of this study indicate that obstetric patients with serologic weak D RBCs in themselves or their newborns may get inconsistent or inappropriate blood bank testing and confusing RhD blood type reports. These problems increase the risk of inadequate maternal RhIG therapy and anti-D alloimmunization. Although RHD genotyping is increasingly used for RHD variant identification, further education and benchmarking studies are needed to improve laboratory practices for serologically weak D-reactive patients. We provide recommendations for appropriate testing, consistent immunohematologic terminology, and RHD genotype–guided management of RhIG therapy and RBC transfusions.