The use of radiolabeled and fluorescein-labeled antiglobulins in assays to predict platelet transfusion outcome

Abstract

We used both radiolabeled and fluorescein-labeled antiglobulins in assays to detect antibodies against platelets in multiply transfused patients to determine the value of these tests in predicting the outcome of platelet transfusion in such patients. In 15 allosensitized patients, we studied 68 single-donor platelet transfusions, 43 (63%) of which had a poor outcome, defined as a corrected count increment (CCI), less than 10,000. The results obtained with either test were significantly correlated with the CCI following transfusion (p less than 0.001), but the assay using the radiolabeled antiglobulin had slightly better sensitivity, specificity, and predictive value. When the assays were used in combination, there was again significant correlation with the CCI of the transfusion, p less than 0.001. When both assays predicted failure of the transfusions, 31/31 (100%) such transfusions resulted in a CCI of less than 10,000, and when both assays predicted success of the transfusions, 14/15 (93%) such transfusions resulted in a CCI of greater than 10,000. Both assays are useful in predicting the outcome of the platelet transfusions; when the assay results were concordant, almost total predictive accuracy was obtained.

Efficacy of HLA‐matched platelet transfusions for patients with hypoproliferative thrombocytopenia: a systematic review

“…16 The patient population was largely an adult population as only two studies enrolled only pediatric patients. 24,26 Fifteen studies were prospective 13,16,[24][25][26][27][28][29][30][31][32][33][34]45,46 and 14 were retrospective. 12,17,22,[35][36][37][38][39][40][41][42][43][44]47 Twenty-one of the nonrandomized studies focused on patients with refractory thrombocytopenia, 12 Two full text articles identified by authors n = 2…”

“…Nine of the 29 studies did not specify the source of sample of patients, that is, how patients were recruited. 22,24,25,29,35,36,38,42,43 Eleven studies detailed sampling methods, 12,16,17,26,27,37,38,40,41,46,47 and one indicated that a random sample was selected but a clear definition of randomization was not provided. 25 The sample size was not predetermined in any study and the assessment of the outcomes was not blinded.…”

“…Fifteen studies clearly defined the eligibility criteria for inclusion of patients, 13,16,17,[26][27][28][29]35,[37][38][39]41,[45][46][47] and most provided clear definitions of outcome. 12,13,16,17,22,[24][25][26][27][28][29][31][32][33][34][35][36][37]39,[41][42][43][44][45][46] Three studies had acceptable control group and comparable characteristics, 16,28,35 five studies 17,22,29,35,47 described details of quality measures for the collection of data and laboratory tests (e.g., accuracy, reproducibility, calibration), and six analyzed confounding factors that potentially influenced the outcomes.…”

“…Author, year PLT count increment Prospective Marktel, 2010 24 IR, LR, SDP HLA matched CCI > 4.5 in 74% Median increment: 43.5 ¥ 10 9 /L with HLA matched vs. CCI > 4.5 in 59% Median increment: 36.5 ¥ 10 9 /L for HLA 1 mismatch vs. CCI > 4.5 in 26% and Median increment: 6.25 ¥ 10 9 /L with RDP; p < 0.01 for HLA vs. RDP, p = 0.02 for random vs. HLA mismatch, p = 0.16 for HLA vs. HLA mismatch IR, LR non-HLA-matched RDP CCI > 4.5 in 74% Median increment: 36 ¥ 10 9 /L Levin, 2003 25 No correlation between HLA antibodies by ELISA, LCT, LIFT + PIFT, and < 20% 1-hr recovery Positive ELISA and PIFT (p = 0.04) and LIFT + PIFT (p = 0.03) associated with 16-hr recovery <10% Petz, 2000 16 Mean 24-hr PPR: SDP, LR, HLA: 21 Ϯ SEM 4%, p = ns vs. random SDP,LR, CXM: 23 Ϯ SEM 4%, p = 0.04 vs. random SDP, LR, ASP: 24 Ϯ SEM 3%, p = 0.007 vs. random LR, SDP: 15 Ϯ SEM 1%, p < 0.01 for ASP vs. random Hogge, 1995 26 11/16 (69%) with LCTABS had 2¥ CCI with HLA matched PLTs compared to RDP p < 0.01, 2/8 (25%) with no LCTABS had response to HLA-matched PLTs p = ns Friedberg, 1994 27 SPRCA CXM was better predictor than HLA for 1 hr, Median CCI Ն 7.5 ¥ 10 9 /L for HLA CXM compatible vs. 0 for HLA CXM incompatible, p < 0.007, Mean CCI for HLA SDP: A + BU 6.1 ¥ 10 9 /L vs. BX + C 3.55 ¥ 10 9 /L vs. SDP, 0, p < 0 Moroff, 1992 45 1-hr CCI Ն 7.5 ¥ 10 9 /L, HLA 54% vs. CXM 48%, p = ns, 24-hr CCI Ն 4.5 ¥ 10 9 /L, HLA 42%, CXM 23%, p < 0.05, 1-hr CCI A, BU 11.0 ¥ 10 9 /L vs. BX 6.0 ¥ 10 9 /L vs. C 9.0 ¥ 10 9 /L vs. D 8.0 ¥ 10 9 /L, p = ns, 1-hr CCI Ն 7.5 ¥ 10 9 /L + 24-hr CCI Ն 4.5 ¥ 10 9 /L in 53% of HLA-matched CXM compatible vs. 45% with HLA-matched CXM incompatible, p = ns Bishop, 1988 46 Mean 1-hr CCI: HLA antibody Grade 0, 15.5 ¥ 10 9 /L vs. Grade 1, 11.6 ¥ 10 9 /L; Grade 2, 8.9 ¥ 10 9 /L; Grade 3, 5.5 ¥ 10 9 /L; Grade 4 mismatch, 5.0 ¥ 10 9 /L Murphy, 1987 28 NR Ware, 1985 29 1-to 2-hr CCI HLA A match: 6,640 Ϯ 7,290 HLA B match: 7,892 Ϯ 6,857 HLA C match: 8,435 Ϯ 9,820 HLA D match: 5,855 Ϯ 9,027 Random: 18,415 Ϯ 5,386 Dahlke, 1984 30 CCI from HLA A3 mismatch to A1 or A11 vs. A-and BU-matched PLTs was less, p < 0.001, A1 or A11 mismatched to A3 vs. A and BU associated with small increments p < 0.001, In B5 group B18, BW16 higher CCI (p < 0.01), Lower CCI with B7 and BW21 Low CCI with B5 to B15 and B17 (p < 0.01), Low CCI B27 to B7 (p < 0.01) and with B8 and B14 bidirectionally and B12 and BW21 bidirectionally (p < 0.05) Hester, 1978 31 Median 1-hr CCI for afebrile patients with two antigens shared = 12.0 ¥ 10 9 /L vs. 8.0 ¥ 10 9 /L for febrile patients (p = 0.01) Median 1-hr CCI for patients with two antigens shared = 12.0 ¥ 10 9 /L vs. Յ8.0 ¥ 10 9 /L in patients with one antigen, no antigen, or unknown (p < 0.01) No difference of CCI for one antigen match vs. no antigen match Median CCI related to number of antigens shared and not specific antigen of haplotype Macpherson, 1979 32 CCI > 5.0 ¥ 10 9 /L in 40% HLA A match vs. 55% in BX p = ns, vs. 14% in C p < 0.05, vs. 21% in D Duquesnoy, 1977 13, 21 1-and 24-hr recovery 55%-75% and 40% for HLA A and B, HLA C and D recovery less than A...…”

Predictive value of enzyme‐linked immunoassay platelet crossmatching for transfusion of platelet concentrates to alloimmunized recipients

“…More recently, immunoglobulin binding assays have been considered the most sensitive tests available. These include immunofluorescence [16][17][18][19][20], radiolabelled antiglobulin tests [ 18,19,21], complement lysis inhibition assay [22] , radiolabelled staphylococcal protein A assay [23], enzyme immunoassays [20,24], and solid-phase red cell adherence test [25]. The predictive values for platelet crossmatches have been determined for all methods except complement lysis inhibition tests and enzyme immunoassays.…”