Impact of Red Blood Cell Transfusion on Platelet Aggregation and Inflammatory Response in Anemic Coronary and Noncoronary Patients

The TRANSFUSION-2 Study (Impact of Transfusion of Red Blood Cell on Platelet Activation and Aggregation Studied With Flow Cytometry Use and Light Transmission Aggregometry)

Study population

The TRANSFUSION-2 study is a cross-sectional observational, prospective study conducted by the Allies in Cardiovascular Trials Initiatives and Organized Networks Group at Institut de Cardiologie, Pitié-Salpêtrière University Hospital (Paris, France). Patients with documented coronary artery disease or without coronary artery disease in whom an allogeneic RBC transfusion was prescribed were enrolled in the study. Inclusion criteria were: 1) age >18 years; 2) stable hemodynamic status; and 3) review of and agreement with the study protocol. Antiplatelet therapy with aspirin, clopidogrel, prasugrel, or ticagrelor was allowed. Exclusion criteria were: 1) previous RBC transfusion in the past 7 days; 2) hemodynamic instability with or without a cardiac assist device; 3) glycoprotein IIb/IIIa inhibitor administration within the past 7 days; 4) septic status at the time of transfusion; 5) use of steroidal and nonsteroidal anti-inflammatory drugs; 6) a low platelet count (<100 × 10⁶/l); and 7) concomitant transfusion of platelets. Written informed consent was obtained before participation, and this study was approved by the Pitié-Salpêtrière University Hospital Ethics Committee (Comité de Protection des Personnes Participants à la Recherche Biomédicale). The study was conducted and funded by the Allies in Cardiovascular Trials Initiatives and Organized Networks study group (http://www.action-coeur.org) and performed within the Institut National de la Santé et de la Recherche Médicale unit UMRS 937. A research grant was also obtained from Société Française de Cardiologie and Fédération Française de Cardiologie.

Data collection

All clinical and biological data from patients who provided informed consent were collected into a prospective, Web-based registry, as well as drug intake to evaluate drug-drug interactions. ABO type, rhesus group, number of units of recipient blood, and dates of collection and thawing were also obtained for each transfusion pack. Preparation of RBC packs was done according to French legislation (http://www.dondusang.net), and RBCs were obtained at our institution using the classic principle of centrifugation of total blood using saline-adenine-glucose-mannitol as the additive solution, followed and conserved between 2°C and 6°C for a maximum of 42 days.

Blood sampling and measurements

Blood was collected into Becton Dickinson (Franklin Lakes, New Jersey) 3.2% citrate Vacuette tubes after having discarded the first 2 ml to 4 ml of blood to avoid spontaneous platelet activation from a peripheral vein in the same fashion before and after transfusion. Blood samples were processed for platelet function testing within 2 h after drawing. Platelet reactivity was tested at baseline on blood drawn just before RBC transfusion, and a second blood draw was performed 12 to 36 h after the end of RBC transfusion. Plasma samples were obtained for each patient before and after transfusion by centrifugation of blood samples at 3,000 g for 20 min, and aliquots of 0.6 ml of plasma were stored at −80°C until assay.

Light transmission aggregometry

Platelet-rich plasma was obtained by centrifugation of citrated whole blood at 100g for 10 min at room temperature. Platelet-poor plasma was obtained by further centrifugation at 4,500g for 15 min. In vitro platelet aggregation in platelet-rich plasma was measured at 37°C using light transmission aggregometry (LTA) (model 490-4D; Chrono-Log Corporation, Kordia, the Netherlands) and was induced by 4 different agonists: 1) adenosine diphosphate (ADP) 20 μmol/l (Sigma-Aldrich, Saint Quentin Fallavier, France); 2) arachidonic acid (AA) 1.25 mmol/l (Sigma-Aldrich); 3) collagen 2 μg/ml (Helena Biosciences, Tyde and Wear, United Kingdom); and 4) thrombin receptor–activated peptide 20 μmol/l. Maximal platelet aggregation (MPA) and residual platelet aggregation (RPA) measured 6 min after the induction of aggregation were recorded for all agonists. Pre-specified criteria used to define nonevaluable samples were lack of sufficient signal, hemolysis, and platelet-rich plasma platelet count <150,000/ml and an unstable baseline signal.

Flow cytometry

Flow cytometry was performed <2 h after venipuncture for P-selectin measurement. The phosphorylation of vasodilator-stimulated phosphoprotein was measured within 24 to 48 h using a Beckman Coulter FC500 cytometer (Beckman Coulter, Villepinte, France).

To determine platelet P-selectin expression, blood samples previously activated by a dose of 10 or 20 μmol/l of ADP were mixed with saturated concentrations of anti-CD62p-PE (Beckman Coulter) monoclonal antibody and anti-CD41a–fluorescein isothiocyanate monoclonal antibody (Beckman Coulter). After staining with antibodies, samples were incubated for 30 min in the dark and diluted with 1 ml of Isoflow Sheath fluid (Beckman Coulter). Samples were immediately processed for flow cytometric analysis. To determine platelet CD62P expression, individual platelets were identified by size (forward and scatter) and anti-CD41a–fluorescein isothiocyanate immunofluorescence using a logarithmic scaled dot plot. P-selectin expression on the surface of platelets was defined as positive for anti-CD62P-PE. Variation in activation corresponded to the percent of gated platelets after activation by agonist minus the percent of gated platelets at rest. The level of P-selectin expression at rest was subtracted from the raw value of P-selectin expression after activation by ADP to compare the variation in delta of P-selection expression.

Vasodilator-stimulated phosphoprotein platelet reactivity index

Vasodilator-stimulated phosphoprotein (VASP) phosphorylation was measured using platelet VASP kits (Diagnostica Stago, Asnières, France) according to the manufacturer's instructions. Briefly, blood samples were incubated in vitro with ADP and/or prostaglandin E₁ (PGE₁) before fixation. The platelet population was identified on its forward-scatter and side-scatter distributions, and 5,000 platelet events were gated and analyzed for mean fluorescence intensity (MFI). The MFI corresponding to each experimental condition (PGE₁ and ADP + PGE₁) was determined to establish a ratio directly correlated with the VASP phosphorylation state. The VASP platelet reactivity index (PRI) was calculated from the MFI of each condition according to the formula: VASP PRI = [(MFI_PGE1 − MFI_PGE1+ADP)/MFI_PGE1] × 100. All measurements were performed in our research laboratory on thrombosis (the Institut National de la Santé et de la Recherche Médicale unit UMRS 937) according to previously published methods (14–17). The technicians were blinded to patients' identities and characteristics.

Biomarkers

Soluble CD40 ligand levels were measured using a human soluble CD40 ligand enzyme-linked immunosorbent bioassay (Biovendor, Karasek, Czech Republic). Tumor necrosis factor–alpha and interleukin-6 levels were measured using Quantikine sandwich enzyme immunoassays (AssayPro, St. Charles, Missouri) according to the manufacturer's instructions. C-reactive protein levels were quantified using commercially available enzyme-linked immunosorbent assay kits (Zymutest; HypenBioMed, Neuville-sur-Oise, France). Plasminogen activator inhibitor-1, d-dimer, and von Willebrand factor antigen concentrations were measured using Asserachrom kits (Diagnostica Stago). Tissue factor pathway inhibitor was measured using sandwich enzyme immunoassays (USCNK, Wuhan, China). All measurements were assessed in duplicate.

Study endpoints

The main endpoint was the relative change in MPA in response to ADP, expressed as the percent of change between baseline and post-transfusion measurement ([MPA_{post-transfusion} − MPA_baseline]/MPA_baseline). Other endpoints were the relative changes of platelet aggregation with other agonists or other platelet function tests and changes in biomarkers before and after transfusion ([post-transfusion value − baseline value]/baseline value).

Clinical follow-up

Patients were followed for 30 days through outpatient consults. Physicians noted the occurrence of major cardiovascular and cerebrovascular events, including death, stroke, recurrent myocardial infarction, urgent revascularization, and definite and probable stent thrombosis (per the Academic Research Consortium definition), as well as major and minor bleeding events according to the Bleeding Academic Research Consortium (BARC) definition (18).

Statistical analyses

On the basis of our previous study (19), at least 60 transfused patients were needed to yield 80% power with a risk for alpha error of 0.05 to demonstrate a difference of 10% relative change in MPA before and after transfusion.

Means and standard deviations are used to report the results. Data were analyzed using paired Student t tests. A confirmatory analysis using the Wilcoxon rank test was also performed to assess statistical significance in case of a possible non-Gaussian distribution. Correlations between RBC storage duration and platelet aggregation data were tested using the Pearson test. Statistical tests were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, California). All tests were 2 sided, with a statistical threshold for significance of 0.05.

Study population and RBC transfusion

Sixty-one patients were included in the TRANSFUSION-2 study. Baseline characteristics are presented in Table 1. Thirty-three patients had suspected ACS treated with both P2Y₁₂ inhibitors and aspirin, and the remaining 28 patients were hospitalized for other cardiac diseases treated with aspirin only or without any antiplatelet agent. Our study population was a high-risk population: one-fifth had diabetes, and 40.9% had renal insufficiency. Two-thirds of the patients were treated with aspirin and proton pump inhibitors. The 33 patients with ACS treated with dual-antiplatelet therapy received clopidogrel 75 mg (n = 28), and 5 patients received prasugrel 10 mg. The mean baseline hemoglobin level before transfusion was 7.8 ± 1.1 g/dl, and the distribution of ABO blood types was 55.7% (n = 34) in group O, 29.5% (n = 18) in group A, 11.4% (n = 7) in group B, and 3.3% (n = 2) in group AB. The vast majority of the patients received 2 packs of RBCs (80%), leading to a mean and significant rise of 2.3 g/l in hemoglobin. The repartition of ABO blood type of the transfused packs matched perfectly the patients' blood types, according to the transfusion rules. Patients included in this study did not receive any other blood products (plasma or platelet concentrates) during the study period, and there was no substantial modification of the transfused patients' medications between the 2 sampling periods (<24 h).

Effect of RBC transfusion on platelet aggregation measured by LTA

Results of platelet aggregation measured by LTA before and after transfusion in response to ADP are presented in Figure 1. Compared with baseline values, transfusion resulted in an 11.6% relative increase in MPA and a 10.8% increase in RPA after platelet activation by ADP (p = 0.004 and p = 0.005, respectively). The results of LTA in response to other agonists are shown in Table 2. Platelet reactivity was also increased by 11.8% for MPA and 12.7% for RPA with thrombin receptor–activated peptide activation (p = 0.04 and p = 0.02, respectively), but no significant difference was found after activation by a nonspecific agonist (collagen) or by AA, exploring the Cox-1 inhibition pathway.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Platelet Reactivity Before and After RBC Transfusion Measured by MPA in Response to ADP 20 μmol/l and by RPA in the Entire Study Population

ADP = adenosine diphosphate; MPA = maximal platelet aggregation; RBC = red blood cell; RPA = residual platelet aggregation. *p < 0.05 for comparison with baseline.

Effect of RBC transfusion on platelet activation measured by flow cytometry

Results for VASP PRI and changes in P-selectin expression are reported in Figure 2. The effect of transfusion on platelet aggregation was confirmed by a relative increase of 20.7% (p = 0.002) in the VASP PRI, a specific marker of the P2Y₁₂ receptor and ADP pathway. Conversely, there was a relative increase of 8.3% in P-selectin expression, a marker of platelet activation, after transfusion of RBCs when 10 μmol/l ADP was used as an agonist (p = 0.07). This difference was blunted by the use of 20 μmol/l ADP (2% increase; p = 0.11).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Variation in Platelet Aggregation and Activation Measured by Flow Cytometry With the VASP PRI and With P-Selectin Expression With 10 μmol/l ADP and 20 μmol/l ADP Measured at Baseline and After Transfusion in the Entire Study Population

ADP = adenosine diphosphate; PRI = platelet reactivity index; VASP = vasodilator-stimulated phosphoprotein. *p < 0.05 for comparison with baseline.

Effect of RBC storage duration and blood type on platelet activation

The median length of conservation of the RBC packs before administration was 18.8 days (interquartile range: 11 to 26 days; maximum 36 days, minimum 3 days). We found no correlation between RBC storage duration and the change in platelet aggregation or platelet activation with LTA (r = 0.07 for ADP, r = 0.03 for collagen, r = 0.17 for AA, r = 0.03 for thrombin receptor–activated peptide). Similar results were obtained for P-selectin expression (r = 0.03 for both ADP conditions). However, we found a significant correlation between the increase in VASP PRI after transfusion and the duration of conservation (r = 0.34, p = 0.009).

There was no clear impact of ABO blood type on platelet reactivity (MPA with ADP 20 μmol/l measured by LTA), as no major difference was found for blood type (+5.89% for group O [n = 34], +6.05% for group A [n = 18], −1.25% for group B [n = 7], and +9.7% for group AB [n = 2]; p = 0.77 for trend).

Effect of RBC transfusion on inflammatory and thrombotic biomarkers

Results of measurements of inflammatory and thrombotic biomarkers and changes before and after transfusion are reported in Table 3. We did not find any significant changes between the pre-transfusion and post-transfusion measurements.

Subgroup analysis

In an exploratory subgroup analysis, we evaluated the results in patients treated with P2Y₁₂ inhibitors compared with those not receiving P2Y₁₂ inhibitors. As expected, coronary patients treated with P2Y₁₂ inhibitors had lower baseline values of platelet reactivity whatever the test used (data not shown). Platelet reactivity after RBC transfusion was predominantly increased in coronary patients treated with P2Y₁₂ inhibitors when measured by LTA (relative increases of 20.4%, p = 0.002, and 15.6%, p = 0.03, for MPA and RPA, respectively) compared with noncoronary patients not receiving P2Y₁₂ inhibitors (relative increases of 4.7%, p = 0.20, and 0.7%, p = 0.40, for MPA and RPA, respectively).

Short-term outcomes

Clinical follow-up was obtained in all patients at 1-month follow-up. The mean length of stay in the hospital was 12 days for transfused patients. Three patients, all from the coronary group, died during the hospitalization period, 1 from a fatal myocardial infarction, 1 from gastrointestinal bleeding, and 1 from septic shock. The rate of major cardiovascular and cerebrovascular events was 11.5% (n = 7). The rate of additional bleeding (per the BARC definition), not counting the initial RBC transfusion, was 9.8% (n = 6), including 2 BARC 3 events and 4 BARC 1 or 2 events.

Study limitations

The lack of a control group should be acknowledged as a theoretical limitation of our study. Several controls, such as anemic or nonanemic patients or healthy volunteers receiving RBC transfusion or another vascular solution, and treated with different regimens of antiplatelet agents, would have been necessary, but such situations are difficult to conceive in clinical research and our previous in vitro model of RBC transfusion in nonanemic healthy volunteers support our findings. Another limitation is whether this effect of RBC transfusion could have been differentiated from that of treatment interruption, for which another study design would have been needed. However, we believe that our study raises platelet recovery as an issue in the context of RBC transfusion in stable patients and that its effect only adds to the increase in platelet reactivity after RBC transfusion. The subgroup analysis of the ACS group was meant only to be exploratory and must be interpreted with caution, as this study was not powered to conduct for such an analysis. Finally, the magnitude of the increase in platelet reactivity found before and after RBC transfusion in our studies, both in vitro and in vivo, can be seen as modest but of similar magnitude to that found in pharmacodynamic studies between carriers and noncarriers of the loss-of-function allele cytochrome P450 2C19*2 (24,25), an effect well correlated with an increase in ischemic events (26,27).

We acknowledge that the increase in platelet reactivity observed in our study is of moderate magnitude, but we believe that it may be a plausible explanation for the detrimental effect of RBC transfusion found in clinical studies (4,6,10,12). In patients with ACS, an increase in platelet reactivity due to the transfusion itself, or potentialized by the cessation of antiplatelet agents, seem to lead to a clinical situation of high risk for ischemic events, especially when added to the detrimental physiological impact of blood loss itself. A recent meta-analysis supports the detrimental effect of blood transfusion in high-risk situations such as myocardial infarction, but there is a lack of analysis of potential confounders such as treatment cessation (13,28). The potential role of the P2Y₁₂ pathway highlighted in our study suggests that interruption of clopidogrel, prasugrel, or ticagrelor in particular may exacerbate the detrimental effects of transfusion, while aspirin interruption may be less of a concern considering the data from our study on the AA pathway.