<scp>RhD</scp> variants: Detected despite 3+ testing in gel

We read with interest “Evaluating RhD assessment by automated methodology: A potential “blind spot” for RhD variant identification” by Yurtsever et al.1 Like the authors, we have been concerned about missing D variants in females of childbearing potential by relying solely on automated gel testing. Based on literature using other automated gel platforms,2 we had previously used 2+ for Grifols D (IgM) gel testing as a cut-off for molecular evaluation. However, after the first three of five patients with 3+ D gel testing sent for molecular evaluation returned as needing to be treated as D negative, we re-evaluated our standard operating procedure. Over a 6–12-month period at two hospitals, 86 females of childbearing potential with ≤3+ testing (62 were 3+) by automated gel D testing and weaker than expected reactivity by tube testing (typically ≤2+ using Immucor anti-D series 4 (IgM and IgG)), had RHD molecular evaluations completed (by Grifols or Versiti). Eighty-five returned as weak or partial D and one returned as D positive. The majority (50/62, 81%) that tested 3+ by gel had a D variant that we would treat as D negative for transfusion and pregnancy purposes. Per the College of American Pathologists Transfusion Medicine Resource Committee Work Group,3 we treat any weak D variants that are type 1, 2, or 3 as D positive, as we do any that are type 4.1. We have elected to treat weak D type 4.0 as D negative out of an abundance of caution given multiple cases of anti-D reported4; we also treat any partial D that is not DAU0 as D negative. Figure 1A shows the breakdown of variant D results, excluding the single patient who had a conventional D allele: 21/85 (24.7%) were weak D type 1, 2, or 3; 29/85 (34.1%) were weak D type 4; 6/85 (7.1%) were weak D type 4.2; 9/85 (10.6%) were other types of weak D, and 20/85 (23.5%) were partial D. Most patients (63/85, 74%) were or had previously been pregnant, 15/85 (18%) had received transfusion, and one had a documented anti-D. Self-reported race, available in the electronic medical record, was compiled for the 85 patients with a variant D result; overall, 72% were Black, 21% were White, 1% was Asian, and 6% were unknown/other. Figure 1B shows the percentage of each variant D category from patients who self-identified as Black, with a disproportionate percentage needing to be treated as D negative. The strength of gel reactivity was evaluated in relation to the molecular results (Figure 2A). Of 10 patients with weak D type 1, eight tested 3+, one tested 2+, and one tested 1+. Of 10 patients with weak D type 2, two tested 3+, seven tested 2+, and one tested weak/1+. The one patient with weak D type 3 tested 3+. Of 29 patients with weak D type 4, 25 tested 3+ and four tested 2+. Of six patients with weak D type 4.2, five tested 3+ and one tested 1+. Of 9 patients with other weak D variants, seven tested 3+ and two tested 2+. Of 20 patients with partial D, 14 tested 3+, 4 tested 2+, and 2 tested 1+. Of note, three patients not included in this cohort had molecular RHD testing completed after we detected an anti-D; despite having 4+ gel testing, all returned as partial D. Tube testing results were predictably weaker than gel testing for patients evaluated using both methodologies; Figure 2B shows the strength of tube reactivity for patients with 3+ testing by gel that had tube testing completed (n = 54/62). Most notable were differences between gel and tube testing in the 20 patients with molecular testing showing partial D: of the 14 patients that tested 3+ by gel, 9/14 (62%) had tube testing that was negative or weakly positive, 3/14 tested 1+, and one tested 2+. Based on these data, we recommended that the manufacturer formally modify the Erytra settings across our hospital sites to flag any D gel test result that was ≤3+. If the patient is a female of childbearing potential, the medical laboratory scientist reports the result as D negative while awaiting tube testing. If tube testing is ≤2+ or if the automated gel testing is ≤2+, we send for RHD molecular testing. It is worth noting, though, that 5 patients in this cohort with 3+ tube testing returned as weak D type 4 or 4.2 and we are considering changing our tube cut-off for molecular testing to ≤3+. Once the molecular testing returns, the placeholder D negative entry in the blood bank information system is updated to reflect how the patient should be treated for transfusion/pregnancy purposes. An entry about the type of D variant is placed in the blood bank information system and a pathologist note is written in the electronic medical record with transfusion recommendations. If the variant should be treated as D negative, an entry is also added to the patient's problem list in the electronic medical record and if the patient is pregnant, the treating obstetrical team is notified. In sum, our data suggest caution when determining the cut-off strength of D reactivity using automated gel platforms (and serology in general)5 in females of childbearing potential, particularly for those who self-identify as Black given the high prevalence in our cohort with D variants. Had we not been aware of the potential “blind spot” for RhD variant identification, we may have incorrectly labeled more than 50 patients who typed as 3+ by automated gel testing as being D positive when—by our institutional guidelines—they should be treated as D negative for pregnancy and transfusion purposes. Failure to detect D variants and resultant lack of consideration of the need for Rh Immune Globulin in pregnancy (or transfusion with D positive red blood cells) may result in anti-D alloimmunization and put subsequent pregnancies at risk of hemolytic disease of the fetus and newborn. No funding supported this manuscript. The authors have disclosed no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.