Cross-match-compatible platelets improve corrected count increments in patients who are refractory to randomly selected platelets

Abstract

Background

Cross-match-compatible platelets are used for the management of thrombocytopenic patients who are refractory to transfusions of randomly selected platelets. Data supporting the effectiveness of platelets that are compatible according to cross-matching with a modified antigen capture enzyme-linked immunosorbent assay (MAC-ELISA or MACE) are limited. This study aimed to determine the effectiveness of cross-match-compatible platelets in an unselected group of refractory patients.

Materials and methods

One hundred ABO compatible single donor platelet transfusions given to 31 refractory patients were studied. Patients were defined to be refractory if their 24-hour corrected count increment (CCI) was <5×10⁹/L following two consecutive platelet transfusions. Platelets were cross-matched by MACE and the CCI was determined to monitor the effectiveness of platelet transfusions.

Results

The clinical sensitivity, specificity, positive predictive value and negative predictive value of the MACE-cross-matched platelets for post-transfusion CCI were 88%, 54.6%, 39.3% and 93.2%, respectively. The difference between adequate and inadequate post-transfusion 24-hour CCI for MACE cross-matched-compatible vs incompatible single donor platelet transfusions was statistically significant (p=0.000). The 24-hour CCI (mean±SD) was significantly higher for cross-match-compatible platelets (9,250±026.6) than for incompatible ones (6,757.94±2,656.5) (p<0.0001). Most of the incompatible cross-matches (73.2%) were due to anti-HLA antibodies, alone (55.3% of cases) or together with anti-platelet glycoprotein antibodies (17.9%).

Discussion

The clinical sensitivity and negative predictive value of platelet cross-matching by MACE were high in this study and such tests may, therefore, be used to select compatible platelets for refractory patients. A high negative predictive value demonstrates the greater chance of an adequate response with cross-matched-compatible platelets.

Introduction

Refractoriness to platelet transfusion is a difficult clinical problem that can compromise the supportive care of thrombocytopenic patients requiring frequent platelet transfusions. As many as 30–70% of patients with malignant haematopoietic disorders, chemotherapy-induced marrow aplasia or marrow transplantation are reported to become refractory to random donor platelet transfusions^1,2. The corrected count increment (CCI) is the most widely used surrogate marker for evaluating refractory patients’ responses to platelet component transfusions and is generally defined as a 1-hour CCI of less than 7,500 to 10,000³; recovery of <20% after 1 hour or <10% after 16 hours^4–6, or the occurrence of two platelet count increments below 5×10⁹/L at 18–24 hours after transfusion of ABO-compatible random donor platelets that have been stored for <72 hours^5,7. Refractoriness may be due to clinical, patient-related factors, (e.g., fever and sepsis) product-related factors or immunological causes. Immunological platelet destruction, mediated by alloantibodies directed against antigens on platelets, is frequently the principal or an important contributing factor in platelet refractoriness^8,9. There is a general correlation between alloimmunisation to human leucocyte antigen(s) (HLA), human platelet antigen(s) (HPA) and clinical platelet refractoriness^10–15. This correlation provides the basis for blood product selection strategies in the management of refractory conditions^2,16.

The strategies used for the management of these patients are either identification of anti- HLA/HPA antibodies in the patients’ serum and providing platelet components which are negative for the concerned antigens (antibody-specificity prediction or “antigen-negative approach”); transfusion of HLA-identical or best-matched platelets; or transfusion of cross-match compatible platelets^16,17. The first two approaches require building a panel of thousands of HLA-typed potential aphaeresis donors, which can be a mammoth task. In contrast, cross-match-compatible platelets are readily available, less expensive and allow matching for platelet-specific antigens. Another advantage is that strong HLA antibodies that are directed against antigens not present on the chosen platelet product and thus irrelevant for platelet transfusion outcome are not detected. This approach is said to be reasonably predictive of post-transfusion platelet count increment^18,19.

A variety of methods have been used for platelet cross-matching (PLT-CM). The commonly used techniques include an antigen capture enzyme-linked immunosorbent assay (ELISA)^20,21, flow cytometry^22,23, platelet immunofluorescence (PIFT)^23,24 and a solid phase red cell adherence (SPRCA) assay^18,25,26. Each of these techniques has its limitations. Assays such as the Immunobead assay²⁷ and the monoclonal antibody-specific immobilisation of platelet antigens (MAIPA) assay²⁸ have been considered as possible reference methods; however, these assays are time-consuming and not easily adapted for routine use. Very few studies in the literature have objectively assessed the benefit of providing cross-matched platelets as opposed to uncross-matched platelets or cross-match-incompatible platelets for transfusion support^{18,19,29–31}.

At our centre patients requiring platelet transfusions routinely receive buffy coat-reduced random donor platelets (RDP). Patients who are refractory to RDP transfusions and can afford the cost of single donor platelets (SDP) are transfused ABO-compatible SDP collected by automated plateletpheresis procedures; no other option is currently available for these refractory patients.

Hence, a study was planned to evaluate the clinical usefulness of PLT-CM using commercially available modified antigen capture ELISA (MACE) I and II (GTI, Brookfield, WI, USA). These are phase III assays and do not require any additional infrastructure. To the best of our knowledge there are no published reports of PLT-CM by MACE. The purpose of the study was to determine whether platelet cross-matching by MACE I and II (PLT-CM-MACE) can effectively identify platelet units that will improve the post-transfusion platelet counts in patients who are refractory to randomly selected platelets, by an objective evaluation of transfusion outcome of cross-match-compatible vs incompatible platelets.

Materials and methods

Study plan

The study was conducted in the Department of Transfusion Medicine at Sanjay Gandhi Postgraduate Institute of Medical Sciences, a superspeciality teaching institute in northern India, between July 2010 and June 2012 after approval from the Institute’s Research and Ethics committee. The study included 100 ABO-compatible SDP transfusions, cross-matched by MACE I and II and transfused within 6–12 hours of collection to 31 patients refractory to platelet transfusions. The patients were defined to be refractory if their 24-hour CCI was <5×10⁹/L following two consecutive platelet transfusions^5,7. Patients with clinically overt sepsis or splenomegaly were excluded. The PLT-CM was done subsequent to the SDP transfusion; until then the patient’s pretransfusion serum sample and SDP sample were stored at 4°C. Testing was done according to the manufacturer’s instructions. The CCI after administration of PLT-CM-compatible vs PLT-CM-incompatible SDP to the clinically refractory patients was studied at 24 hours. The 1-hour CCI was not assessed because of problems with sample collection due to transfusions at odd times.

During the study period, a prophylactic platelet transfusion strategy was used at our institute: the standard platelet component for all recipients was non-leucoreduced, unpooled RDP. In selected cases, including patients with recurrent non-haemolytic febrile transfusion reactions and those undergoing haematopoietic stem cell transplantation, bedside leucofiltration was performed. The platelets were also irradiated with gamma ray irradiation at a dose of 25 Gy according to standard AABB indications to prevent transfusion-associated Graft-vs-Host disease.

Patients’ data

The patients’ demographic and clinical data, i.e., age, gender, diagnosis, presence of splenomegaly, fever, infection or bleeding at the time of SDP transfusion, the ABO and Rh D blood group of recipient and donor, body weight and height of the patient, platelet count before and 24 hours the platelet transfusion and platelet count in the SDP bag were recorded. Bleeding was categorised according to the World Health Organization (WHO) bleeding scale³². Blood component transfusions and response details for previous platelet transfusions were also recorded. The patients’ blood volume and body surface area were calculated using standard formula³³.

Measurement of platelet transfusion outcome

The 24-hour CCI was the primary response variable and was calculated as follows:

$$24-\text{hour\hspace{0.17em}CCI}=\frac{({\text{PLT}}_{\text{post}}-{\text{PLT}}_{\text{pre}})\times \text{BSA}}{\text{PLT}/\text{unit}}$$

where PLT_post: platelet count (/μL) 24 hours after transfusion, PLT_pre: pre-transfusion platelet count (/μL), PLT/unit: donor unit platelet count (×10¹²) and BSA: body surface area (m²). Responses were termed as inadequate if the 24-hour CCI was <5×10⁹/L.

Platelet cross-match assays

Platelets were cross-matched using commercially available MACE I and MACE II kits (GTI, Brookfield, WI, USA). MACE I is a qualitative, solid-phase ELISA designed to detect and distinguish IgG antibodies to HLA class I antigens and to epitopes on the platelet glycoprotein (GP) IIb/IIIa. MACE II is a qualitative, solid-phase ELISA designed to detect and distinguish IgG antibodies to epitopes on the platelet GP Ia/IIa, Ib/IX and IV. The procedures and interpretation were conducted in accordance with the manufacturer’s instruction manual and compatibility status was noted. If incompatibility was noted against any one or more of the specificities detected by PLT-CM-MACE I and II, the SDP was termed as incompatible. The clinical sensitivity, specificity, positive predictive value and negative predictive value of PLT-CM-MACE were calculated, as shown in Table I. The results of platelet cross-matching were termed as true positive, true negative, false positive or false negative depending upon whether the 24-hour CCI was adequate or inadequate (Table I). The clinical sensitivity of a test is the chance with which the test diagnoses all true positive cases, i.e., in this case all cases with an inadequate response. The significance of a particular antibody specificity responsible for cross-match incompatibility was calculated by multiple regression analysis.

Table I.

Assessment of usefulness of platelet cross-matching.

Platelet cross-match	Response to single donor platelet transfusion
	Inadequate CCI	Adequate CCI	Predictive value
Positive	TP	FP	TP / TP+FP
Negative	FN	TN	TN / TN+FN
	Clinical sensitivity	Clinical specificity
	TP / TP+FN	TN / TN+FP

TP: true positive; FP: false positive; TN: true negative; FN: false negative.

Statistical analysis

Microsoft Excel 2007 (Lucknow, India) was used for data analysis. A chi-square test was used to calculate the relation between transfusion response and results of platelet cross-matching. Multiple regression analysis was done to determine whether a particular antibody specificity could predict transfusion response better than others. P values of less than 0.05 were considered statistically significant.

Results

Patients’ characteristics

The clinical data of the 31 patients and 100 SDP transfusions included in the study are summarised in Table II. The median age of the patients was 32 years (range, 6 to 75 years). There were 20 male and 11 female patients. Twenty-one (67.7%) of the patients were suffering from aplastic anaemia and remaining 10 (32.3%) had acute myeloid leukaemia. As shown in Table II these patients received from one to more than ten PLT-CM SDP transfusions as part of the study. Seventy-three (73%) of the SDP transfusions included in the study were prophylactic transfusions whereas 27 (27%) were transfused for management of WHO grade III bleeding episodes.

Table II.

Patients’ demographics and platelet transfusion details.

Parameter		N. of patients (%)
Diagnosis	Aplastic anaemia	21 (67.7%)
	Acute myeloid leukaemia	10 (32.3%)
Gender	Male	20 (64.5%)
	Female	11 (35.5%)
Blood group	B RhD positive	12 (38.7%)
	A RhD positive	8 (25.8%)
	O RhD positive	6 (19.4%)
	AB RhD positive	3 (9.7%)
	B RhD negative	1 (3.2%)
	O RhD negative	1 (3.2%)
Episodes of SDP transfusion evaluated	One	12
	Two	3
	Three	5
	Four	5
	5–10	5
	>10	1

SDP: single donor platelets.

Prior to inclusion in study the patients had been transfused with a mean number of 14 (range, 3–25) packed red cell units, 81.6 (range, 20–172) RDP units and nine (range, 0–68) fresh-frozen plasma units.

There were 256 days of bleeding during the study observation period. Most were grade 1 episodes (115 days, 44.9%); the remaining were grade 2 (86 days, 33.6%) and grade 3 (55 days, 21.5%). There were no WHO grade 4 bleeding episodes.

PLT-CM-MACE and outcome of single-donor platelet transfusions

Table III shows the results of PLT-CM and response to SDP transfusion as adequate and inadequate as determined by the CCI. PLT-CM-MACE was incompatible (positive) in 56 out of 100 (56%) SDP transfusions. Of these 22 (39.3%) had inadequate 24-hour post-transfusion CCI (true positive; positive predictive value) and 34 (60.7%) had adequate post-transfusion CCI (false positive). As shown in Table III, PLT-CM-MACE was compatible for 44 (44%) of the SDP transfusions. Of these, 41 (93.2%) resulted in adequate 24-hour post-transfusion CCI (true negative, negative predictive value). As evident from Table III, PLT-CM-ELISA was positive in 22 out of 25 (88%) SDP transfusions with inadequate 24-hour post-transfusion CCI (clinical sensitivity); however, the test was negative in only 41 out of 75 (54.6%) SDP transfusions with an adequate 24-hour post-transfusion CCI response (clinical specificity). Thus the clinical sensitivity, specificity, positive predictive value and negative predictive value of PLT-CM-MACE for post-transfusion CCI were 88%, 54.6%, 39.3% and 93.2%, respectively. The difference between adequate and inadequate post-transfusion CCI for MACE cross-matched-compatible vs incompatible SDP transfusions was statistically significant (chi-square value with Yates’ correction=12.175, p=0.000). A positive PLT-CM-MACE predicted poor platelet recovery at 24 hours (odds ratio [OR]=8.843; 95% confidence interval [95% CI]=2.23–40.8; p=0.000). The 24-hour CCI (mean±SD) was also significantly higher for cross-match-compatible platelets (9,250±3,026.6) as compared to cross-match-incompatible platelets (6,757.94±2,656.5) (two-tailed, unpaired t-test, p<0.0001). In our study 21 (63.6%) of the 33 SDP transfusions given to female patients were found to be incompatible, whereas only 35 (52.2%) of the 67 SDP transfusions given to male patients were incompatible.

Table III.

PLT- CM-MACE and single donor platelet transfusion outcome.

Platelet cross-match by MACE N=100	Response to SDP Transfusion
	Inadequate CCI (Poor response) N=25	Adequate CCI (Good response) N=75
Incompatible (N=56)	22*	34
Compatible (N=44)	3*	41
p value*		0.000

PLT-CM: platelet cross-match; MACE, modified antigen capture ELISA; SDP: single donor platelets; CCI: corrected count increment;

^*P value calculated by the Chi-square test.

Antibody specificities in cross-match-incompatible cases and their significance

As shown in Figure 1, out of 56 incompatible SDP, 25 (44.6%) were incompatible by MACE I, 21 (37.3%) by both MACE I and MACE II and 10 (17,9%) were incompatible by MACE II alone. Forty-one out of 56 (73.2%) of the incompatible cross-matches were due to anti-HLA antibodies, either alone in 18/56 (32.2%) instances or in combination with anti-platelet GP antibodies in 23/56 (41%) cases. Anti-platelet GP antibodies alone were responsible for 15/56 (26.8 %) of the incompatible cross-matches (Figure 1).

Figure 1.

Antibody specificities responsible for platelet cross-match incompatibility.

*Incompatible only for HLA (N=18, 32.2%); ** Incompatible for both HLA and PLT-GP (N=23, 41%); ***Incompatible only for PLT-GP (N=15, 26.8%); # MACE II specific anti-PLT-GP = Antibodies against platelet glycoprotein (GP) Ia/IIa, Ib/IX and V.

Table IV shows the frequency of antibody specificities in PLT-CM-MACE incompatible cases and its significance in predicting the post-transfusion response as calculated by multiple regression analysis. Only anti-HLA and anti-IIb/IIIa incompatibilities were found to be significant predictors of an inadequate post-transfusion response in this study.

Table IV.

Antibody specificity in cross-match-incompatible cases as predictors for platelet transfusion response.

Antibody specificity	Frequency in total PLT-CM-incompatible cases (N=56) N (%)*	Significance for inadequate post-transfusion response P value
Anti-HLA	41 (73.2)	0.038
Anti-IIb/IIIa	15 (26.7)	0.002
Anti-Ia/IIa	27 (48.2)	0.129
Anti-Ib/IX	14 (25)	0.813
Anti-GPIV	7 (12.5)	0.092

P value calculated by multiple regression analysis.

^*In many instances more than one antibody was responsible for incompatibility.

Discussion

This study was performed to assess whether PLT-CM-MACE can effectively identify platelet units that will improve the post-transfusion platelet count in patients who are refractory to randomly selected platelets. As various patient- and product-related factors influence the outcome of platelet transfusions we tried to minimise their influence by strict study inclusion criteria and transfusing the ABO-compatible SDP within 6 to 12 hours of collection. In the present study the clinical sensitivity, specificity, positive predictive value and negative predictive value of PLT-CM-MACE for post-transfusion CCI were 88%, 54.6%, 39.3% and 93.2%, respectively. Out of 44 PLT-CM-MACE-compatible SDP transfusions given to refractory patients in this study, 41 (93.2%) resulted in adequate CCI. The difference between adequate and inadequate post-transfusion CCI for MACE cross-match-compatible vs incompatible SDP transfusions was statistically significant (p=0.000). Table V summarises the study population, methods used and the results of published platelet cross-matching studies for comparison. In the study by Levin et al.²⁴, 184 serum samples from 95 randomly selected thrombocytopenic patients with haematological malignancies were cross-matched by platelet immunofluorescence tests (CM-PIFT). Of the patients with positive platelet cross-matches poor platelet recovery was found in 12 (46%), i.e. the positive predictive value of CM-PIFT was 46%. Of the patients with negative CM-PIFT, poor recovery was demonstrated in 36%, i.e. the negative predictive value was only 64%. This contrasts with the PLT-CM-ELISA using MACE I and II in present study, which had a negative predictive value >90%. The positive predictive value in our study was 39.3%. As explained by Levin et al.²⁴, IgG bound to transfused platelets does not necessarily lead to enhanced platelet destruction, whether it is alloantibodies, immune complex or other non-specific binding of immunoglobulins to platelets. This could also be due to PLT-CM incompatibility resulting from antibody specificities that do not necessarily mediate platelet destruction; as demonstrated in the present study, only anti-HLA and anti-GP IIb/IIIa were significant predictors of poor response.

Table V.

Review of published studies on evaluation of platelet cross-matching.

Study N.	Reference	Method used	Study population	Interpretation
1	Gelb AB, Leavitt AD34	SPRCA	475 CM compatible platelet transfusions to 66 refractory patients	CM compatible platelets significantly improve the mean CCI for approximately one half of patients who are refractory to random donor platelets even when the patients are not preselected for having alloimmunisation to explain their refractory state.
2	Levin MD et al.24	CM-PIFT	184 CM platelet transfusions to 95 patients (158 CM-compatible and 26 [14%] CM-incompatible)	The CM-incompatible platelets did not predict poor platelet recovery (OR 0.66, p=0.33).
3	Rebulla P et al.26	SPRCA	569 CM compatible platelet transfusions given to 40 refractory patients	Platelet count increments were significantly higher (p<0.05, t-test) than those observed in the same patients given 303 transfusions from random donor platelet pools.
4	Witta AP, Nambiar A35	SPRCA	738 CM compatible platelet transfusions given to 71 patients	Only 41% of the compatible units improved 1 hour CCI to >7,500. However this CCI response was still significantly higher than comparable random donor platelets units for these patients.
5	Sayed D et al.22	Flow cytometric platelet CM	60 ABO compatible SDP units given to 39 patients	CM-compatible platelets associated with good response in 51.4% of transfusion events. Non- cross-matched platelets, poor response in 83% of events.
6	Present study	PLT-CM-MACE	100 ABO compatible SDP given to 31 clinically refractory patients	CM-compatible platelets associated with adequate CCI in 93.2% of transfusion events, CM-incompatible platelets associated with poor response in 39.3% of transfusion events.

CM: cross-match; PIFT: platelet immunofluorescence test; SPRCA: solid phase red cell adherence assay; PLT-CM-MACE: platelet cross-matching by a modified antigen capture ELISA.

In our study the 24-hour CCI (mean±SD) was significantly higher for cross-match-compatible platelets (9,250±3,026.6) than for cross-match-incompatible ones (6,757.94±2,656.5) (p<0.0001). Similar observations have also been made by other authors^18,26,34. In a study by Rebulla et al.²⁶ 40 refractory patients were given 569 cross-match-compatible platelets (cross-matched by a commercially available immunoadherence assay, Capture P; Immucor, Norcross, GA, USA). Post-transfusion increments were significantly higher (p<0.05) than those observed in the same patients given 303 transfusions from random platelet pools.

In the present study 63.6% of the SDP transfusions given to female patients were found to be incompatible as compared to 52.2% given to male patients. In the study by Rebulla et al. women were more frequently refractory to platelet transfusions than were men²⁶.

In our study 56% of SDP transfusion were found to be incompatible; thus, in order to provide compatible platelets for refractory patients a large inventory of SDP would have to be available. This was also demonstrated by the study by Rebulla et al.²⁶, in which a median of 70 donors (range, 12–141) were tested to provide each cross-match-compatible transfusion.

In this study 73.2% of the incompatible cross-matches were due to anti-HLA antibodies; either alone in 18/56 (32.2%) instances or in combination with anti-platelet-GP antibodies (23/56, 41%). Anti-platelet-GP antibodies alone were responsible for 15/56 (26.8%) of incompatible cross-matches. In the study by Rebulla et al.²⁶ 85% of the patients had anti-HLA antibodies while only 15% had both anti-HLA and anti-HPA antibodies; in comparison, there was an unexpectedly high proportion of platelet-specific antibodies in our study (N=38, 67.8%). As the autocontrol test undertaken as part of the test was negative, these antibodies are likely to be specific HPA antibodies.

As only anti-HLA and anti-GP IIb/IIIa were found to be significant predictors of poor response in our study, the cause of poor specificity of PLT-CM-MACE in our study could be detection of antibody specificities directed against platelet GP Ia/IIa, Ib/IX and IV by MACE II. This aspect needs to be further evaluated.

In conclusion, our study suggests that PLT-CM-MACE may be a useful tool for selecting effective platelets from a local inventory for patients who are refractory to randomly selected SDP transfusions. Consequently we feel that for blood banks and transfusion services not equipped to perform flow cytometry or other standard methods for the detection of platelet antibodies, MACE is a good strategy for providing platelet transfusion support to refractory patients. However as the specificity and positive predictive value of the test is low, it is not recommended as a routine measure for all SDP transfusions.