The Use of Platelet Antibody Testing at a Tertiary Medical Center.

Detection of platelet antibodies is often helpful in the diagnosis of platelet transfusion refractoriness (PR), neonatal alloimmune thrombocytopenia (NAIT), and post-transfusion purpura (PTP), but is controversial in idopathic thrombocytopenia purpura (ITP). We performed a retrospective chart review to determine the utilization and value of a solid-phase ELISA assay consisting of 6 wells coated with platelet glycoproteins (GP), and one well with HLA class I antigens (GTI PakPlus, Brookfield WI) at our medical center. For a 3 year period between June 2001 and June 2004, 183 tests for 163 patients were ordered for the following diagnostic questions: 151 tests for ITP, 19 for PR, 6 for NAIT, 2 for transfusion-related acute lung injury (TRALI), 5 for unclear reasons, and none for PTP. The mean patient age was 55 (range 12 – 94), 55% were male, 50% were inpatients, and the mean platelet count (excluding mothers with NAIT and a patient with TRALI) was 60x109/l (range 5–215 x 109/l). There were 48 patients who had 51 positive tests (28%), of which 61% were inpatients. The mean platelet count was 47 x 109/l (range 5–115 x 109/l). Only one or more GP wells were positive in 57% of positive tests, only the HLA class I well was positive in 27%, and ≥one GP well plus the HLA well were positive in 16%. To determine how results were used, we reviewed the medical records associated with each positive test. Out of 51 positive tests, no mention was made of the test order or result in 21 (41%). For 6 tests (12%), the order was noted, but no results were mentioned. In the other 24 tests, the result was used to make a diagnosis in 12 cases (3 NAIT, 3 PR, 6 ITP), and in 12 the result was used to confirm an already known diagnosis (1 NAIT, 4 PR, 7 ITP). Therapy was started before the test was completed on 17 occasions (4 PR, 13 ITP). Eighteen patients had repeat testing (38 tests). The interval of time between tests was 112 days on average (range 8–365), including 5 sets within 28 days. The results were unchanged in 11 patients (61%); in 5 patients GP wells changed from positive to negative; and in 2 patients the HLA well changed, including a patient with pre- and post-transfusion samples for evaluation of TRALI, and another in whom the test went from positive to negative. In 3 mothers evaluated for NAIT with positive results (1 HPA-1a, 2 HPA-5b), confirmation was obtained at a reference laboratory (Blood Center for Southeastern Wisconsin). A total of fifteen patients were evaluated for PR with 19 samples. Seventeen samples underwent parallel testing (HLA lymphocytotoxic antibody screen alone, or with platelet antibody detection at the reference laboratory); 13 results (76%) were confirmed. All four mismatches were between the HLA ELISA well and the HLA antibody screen: in 2 samples the PakPlus was positive but the HLA antibody screen was negative; in the other 2 a negative HLA well was accompanied by a high PRA (63%, 80%). These results demonstrated that 1) the ELISA assay was routinely negative in patients with platelet counts >115 x 109/l; 2) 53% of positive results were ignored in the medical records; 3) the test had a moderate reproducibility (61%), but could have been related to clinical changes in the patients; and 4) the yield of clinically useful information for patients with ITP (151 tests, 13 positive results utilized for clinical management) was low. For PR, the PakPlus is reliable but only qualitative, compared to the quantitative HLA antibody screen. The test appears most useful for rapid antibody identification in NAIT and is of limited clinical utility for ITP.