Cis-AB showing discrepant results across different automated and manual methods: a case report and review of the literature

To the Editor,

ABO grouping is the most important pretransfusion testing that is directly related to the quality performance of immunohematology laboratories and patient safety [1]. Among the ABO subgroups, cis-AB is caused by a gene alteration resulting in a single glycosyltransferase enzyme with dual A and B glycosyltransferase activities; there exist various phenotypes of the cis-AB blood group, and these phenotypes are associated with various cis-AB alleles [2]. Although cis-AB is rare worldwide, it is relatively common in Northeast Asia, especially in Korea, with the reported frequency of 0.0354% (60/169,605) in all blood donors from southwestern Korea [3]. Because ABO subgroups including cis-AB may cause problems with either ABO discrepancy or mistyping, it is important to detect cis-AB accurately [2, 3]. We report a case of cis-AB, A₁B_w, that showed discrepant results in ABO grouping across different automated and manual methods. This case report was exempted from the approval of the Institutional Review Board of Konkuk University Medical Center, Seoul, Korea (IRB 2022-10-030) with waived informed consent.

A 39-year-old pregnant woman was referred to the obstetrics clinic for intrauterine growth retardation. In her medical record, she had no history of transfusion, transplantation, or any other comorbidities that might affect red cell activity in ABO grouping. Before this admission, ABO grouping was performed three times in our hospital immunohematology laboratory, and her blood group was constantly determined to be A⁺, using ORTHO VISION Max analyzer (Ortho-Clinical Diagnostics, Raritan, NJ, USA). On admission, however, it was notified that her blood group was determined to be cis-AB in a testing conducted in the other laboratory. In that laboratory, the flag sign “?” was detected by red cell reaction with anti-B in an automated method, using Qwalys-3 analyzer (Diagast, Loos, France) (Figure 1A). However, the result of confirmatory manual method was A⁺ using SIHDIA reagent (Shinyang Chemical, Seoul, Korea). Accordingly, another manual method was additionally performed using Diagast reagent (Diagast, Loos, France); due to the detection of small amount B antigen, the result of red cell reaction with anti-B was “weak+”, and the subtyping showed anti-A₁ “4+” and anti-H “1+” (Figure 1B).

Figure 1:

ABO typing results of the patient. (A) Automated method (Qwalys-3; Diagast, Loos, France), (B) manual method using Diagast reagent (Diagast, Loos, France) for anti-A, anti-B, anti-D, (C) automated method (ORTHO VISION Swift; Ortho-Clinical Diagnostics, Raritan, NJ, USA), (D) manual method using BioClone reagent (Ortho-Clinical Diagnostics) for anti-A, anti-B, anti-D.

With that information, the ABO grouping was rechecked in our hospital; however, the results of an automated method using ORTHO VISION Swift analyzer (Ortho-Clinical Diagnostics) and a manual method using BioClone reagent (Ortho-Clinical Diagnostics) were all A⁺ (Figure 1C and D). ABO genotyping was further conducted, and her blood genotype was confirmed to be cis-AB01/A102. During the admission, she received no transfusion.

Compared with conventional manual methods, automated methods for ABO grouping have been expected to increase work efficiency and reduce errors [4, 5]. Automated methods may show flag signs when there are mixed fields, specimen turbidity, mismatch results, and/or weak reactions, and these flag signs may trigger manual methods to confirm or recheck ABO grouping [6], [7], [8], [9]. Although automated methods are quite useful and reliable, there are often cases that may show ABO discrepancy or mistyping between automated and manual methods, and such a discrepancy may be observed even between automated methods with different principles [6], [7], [8], [9].

Table 1 summarizes 30 reported cis-AB cases (including our case) that showed discrepant results across automated and manual methods. To the best of our knowledge, 15 cis-AB cases out of these 30 cases were detected due to weak reactions or flag signs in automated methods. It is noteworthy that the results of four cases out of these 15 cases (including our case) showed flag signs or discrepancy in only one automated method, although two automated methods were applied [7]. These four cases were found in the circumstance of evaluation study, not in the real-hospital practice; it implies that these cases could have been missed, if only one autoanalyzer was applied. Regarding the other 15 cis-AB cases, they were mistyped typical AB in automated methods [7, 9]; these cases were detected due to the laboratories’ own policy to utilize both automated and manual methods simultaneously when ABO grouping is requested. In immunohematology laboratories that implemented automated methods, a generally accepted workflow would be to perform a confirmatory manual method when triggered by abnormal results in automated methods. In other words, unless automated and manual methods were used simultaneously, these cases would not have been detected.

Taken together, these 30 cis-AB cases imply that the implementation of automated methods itself cannot necessarily guarantee reducing errors in ABO grouping, especially for rare ABO subgroups such as cis-AB. It was also noteworthy that two cis-AB cases were not detected even though two automated methods with different principles were used [7]. Moreover, our case showed the possibility of mistyping cis-AB, regardless of using both automated and manual methods; among the five methods (two automated methods and three manual methods), three methods could not detect cis-AB. In addition to the different performance of automated methods, our case demonstrates that results of manual method may vary depending on the sensitivity of reagents.

In transfusion practice, it is essential to implement immunohematological tests with high analytical sensitivity for safe transfusion. However, it is challenging to evaluate the analytical performance of immunohematological tests for rare ABO subgroups because of their low frequency. If external quality control material for rare ABO subgroups is available, it would help evaluate and develop the analytical performance of immunohematological tests. Vigorous effort, including external quality assessment (EQA)/proficiency testing programs, has been made to improve the laboratories’ quality worldwide [10]. Compared with other laboratory tests, immunohematological tests may be assumed that the accuracy of their nominal results can be clearly assessed, the analytical methods used are the same worldwide, and commutability has no effect [10]. However, the present cases underscore a practical issue that rare ABO subgroups may not be detected depending on the composition of autoanalyzer technologies and reagent sensitivity even in high-performance laboratories. Therefore, immunohematology laboratories should pay special attention to the possibility of ABO mistyping, especially in cases of rare ABO subgroup like cis-AB. In addition to the current EQA program service, it would be meaningful to extend the EQA scheme for detecting rare blood subgroups across reference laboratories. Further studies would be also required to explore the analytical performance of various methods for rare blood subgroups.