Phenotype-dependent effects of 2-mercaptoethanol pretreatment on ABO antibody reactivity by automated column agglutination

Objective To quantify phenotype-dependent effects of 2-mercaptoethanol (2-Me) pretreatment on ABO antibody reactivity measured by automated column agglutination and assess consistency across healthy participants and hematological patients. Methods The study included 104 healthy participants and 92 hematological patients, yielding 258 participant-antibody pairs and 516 repeated measurements. Antibody titer score was the primary mixed-effects-model outcome. Dilution-step changes were analyzed using the prespecified local weak-positive endpoint and, as a sensitivity analysis, the conventional ≥1+ endpoint. Crossings of 1:16, 1:32, 1:64, and 1:128 thresholds were summarized by phenotype, and available patient diagnoses, recent exposures, and immunological measurements were characterized. Results The 2-Me effect differed by antibody phenotype (treatment × phenotype, P < 0.001) but not by population (treatment × population, P = 0.164). Adjusted score reductions were 20.59, 12.81, 7.28, and 7.56 points for A/anti-B, B/anti-A, O/anti-A, and O/anti-B, respectively. Under the local endpoint, at least a one-step reduction or post-2-Me non-reactivity occurred in 90.9%, 72.1%, 37.1%, and 46.8%, respectively; the corresponding ≥1+ sensitivity estimates were 89.4%, 66.2%, 37.1%, and 50.0%. At the 1:16 threshold, 39/54, 25/59, 1/62, and 2/62 initially eligible pairs crossed below the threshold, respectively. Conclusion 2-Me-associated shifts in automated ABO antibody measurements were phenotype dependent and frequently crossed commonly used dilution values, especially for A/anti-B. These data quantify method-dependent changes rather than clinical benefit or analytical accuracy; laboratories should identify the pretreatment and endpoint method and should not transfer decision thresholds across methods without validation.