Background

The direct antiglobulin test (DAT) is frequently ordered by both hematologists and non-hematology medical professionals in the workup of possible allo- or autoimmune hemolytic anemias. Frustratingly, positive DAT results in the absence of other abnormal blood hemolysis biomarkers can be challenging to interpret, as published data on the frequency of these "false-positive" DATs is limited. We sought to characterize the frequency of "false-positive" DATs and describe potential associations with co-morbid diseases, in order to better define the diagnostic accuracy of the DAT for allo- or autoimmune hemolysis.

Methods

After obtaining appropriate Institutional Review Board approval, we identified all positive DAT results from Oregon Health & Science University during the calendar year between January and December 2019. Electronic medical records of the patients associated with these positive DAT results were individually reviewed, with collected data including patient demographics, specific pattern of DAT positivity (IgG, C3 or both), reason for obtaining DAT (suspicion for autoimmune hemolysis or transfusion, or as part of preparatory workup for possible blood product transfusion), hemolysis biomarkers (CBC, reticulocytes, LDH, haptoglobin, total and direct bilirubin), recent blood product transfusion or IVIG administration within the preceding 6 months, and co-morbid illnesses. True positive DAT results were considered those concurrent with at least one other positive blood hemolysis biomarker, or those in patients with a pre-existing diagnosis of autoimmune hemolytic anemia or Evans syndrome.

Results

182 positive DAT tests among 134 individual patients were identified within the specified time period. 65 of these were obtained in the ambulatory setting, and 117 as inpatient. The majority of DAT results were positive for IgG only (165 of 182, 90.7%), with 17 of 182 (9.3%) positive for both IgG and C3, and no instances of isolated C3 positivity. The median values for hemolysis biomarkers were as follows: LDH 307, total bilirubin 0.9, and haptoglobin 80. Of the 134 patients with positive DAT results, 48 (36%) had at least one other positive or elevated blood biomarker of hemolysis, and 31 (23%) had two or more positive biomarkers. 55 records (41%) had no other associated findings of hemolysis, and 31 patients had not had hemolysis biomarkers drawn. 53 of 134 patients (40%) had DATs checked as part of alloantibody workup prior to potential blood product transfusion. The calculated rate of DAT positivity concurrent with other positive hemolysis biomarkers was 79 of 134 (59%), and DATs without other signs of hemolysis were seen in 55 of 134 (41%). Among patients with DAT results without signs of hemolysis, common co-morbid illnesses/conditions included recent blood product transfusion, multiple myeloma, acute myeloid leukemia, chronic lymphocytic leukemia, myelodysplastic syndrome, hypothyroidism, and cirrhosis. The majority of patients had received blood product transfusion or IVIG within the preceding 6 months (94 out of 134 [70%]).

Discussion

In this single-center retrospective analysis, we found a substantial rate of DAT positivity for IgG, or combined IgG and C3, without other laboratory biomarkers of hemolysis. While many of these non-hemolysis-related positive DAT results could be considered false positives, a substantial proportion of patients had also received blood product transfusion or IVIG within the previous 6 months, and thus a large fraction of included patients may have had positive DATs related to previous alloimmunization and lingering "bystander" hemolysis or agglutination of autologous red blood cells. In addition, co-morbid hepatic, autoimmune or hematologic diseases were common among this cohort, corroborating findings in previously published analyses. In summary, we found a clinically meaningful rate of false-positive DAT results in both ambulatory and inpatient settings. We posit that a positive DAT result, in isolation, need not drive extensive hemolysis workup, given the frequent association with prior blood product transfusions and numerous co-morbid illnesses.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.