Effect of Cardiopulmonary Bypass Machine on the Platelet Function Testing.

Background: Bleeding during coronary artery bypass surgery is a leading cause of mortality. Several factors have been associated with bleeding, platelet dysfunction being the most signicant. Objective: to assess the effect of cardiopulmonary bypass machine (CPB) during cardiac surgery on platelet function using Platelet Function Analyzers (PFA-100), and Multiplate Electrode Aggregometry (MEA), and correlating that with a drop in Hemoglobin (Hb). Methods: Whole blood samples were collected preoperative and sixty minutes intraoperatively of different cardiac procedures utilizing (CPB) and tested for platelet function by PFA-100 and MEA. Complete blood count was measured using an automated hematology analyzer. Results: A signicant difference was found between pre- and intraoperative ADP and EPI measurement in PFA-100, where preoperative PFA-ADP values displayed the ability to predict the intra-op drop in Hb (P– value 0.01, correlation coecient 0.4699). At the same time, pre-op MEA- Ristocetin and TRAP showed an inverse correlation with an intra-op drop in Hb (-0.31 and -0.36). Conclusion: The current study reported signicant changes in platelet dysfunction in cardiac surgeries with CPB, measured by two modalities PFA-100, and MEA. While PFA-100 and MEA both detected the changes in platelet dysfunction due to CPB, PFA-100 results were sensitive and positively predicted intra-op Hb drop as compared to MEA. There was a signicant change in Hb one hour into the CPB, indicating that platelet transfusion might help decrease Intra- and postoperative bleeding independent of the platelet count as they are dysfunctional. PFA-100 results can be relied upon for distinction of high-risk cardiac surgery patients for bleeding and can be used for clinical decision making to improve patient outcome.

Thus, emphasizing the need for evidence-based critical evaluation of the diagnostic performance of PFA-100 and MEA for platelet dysfunction in the clinical setting.
Several other factors are linked to morbidity and mortality associated with CPB, but none singularly has been precisely proven (4). The routine procedure of CPB still requires immediate therapeutic improvements and optimization strategies to deal with pre-and intraoperative complications, such as functional haemostatic defects leading to its hall-mark of bleeding (5). Bleeding incidence after CPB varies from 5-25%, and platelet dysfunction is possibly the most plausible yet controversial explanation for coagulopathy related to CPB (6). Platelets, the smallest anucleate fragments in the blood, are activated during CPB and play a crucial role in hemostasis and immune response, speci cally systemic in ammation, tissue repair, and regeneration (7,8). Quantitative, as well as qualitative aspects of platelet dysfunction, participate in thromboembolism and microvascular injury (9). Platelets communicate and build a number of interactions with a variety of blood proteins like coagulation factors and endothelium, thus imparting paramount contribution in hemostasis, directly and indirectly (10). However, in CPB related hemostatic defects, functional abnormalities of platelets are extrinsic rather than intrinsic. This is depicted by optimal platelet reactivity in CPB in-vitro and a minimum amount of degranulated circulating platelets (11).
Owing to the paramount importance of platelet function in bleeding diathesis with its subsequent in uence on therapeutic strategies and patient outcome, there is a dire need to address two major aspects of cardiac surgery. Firstly, the different modalities of cardiac surgery utilizing CPB and its effect on platelet function; and secondly the evaluation of the diagnostic performance of point-of-care testing devices dedicated to the measurement of platelet function, such as PFA-100, MEA, and VerifyNow System (2,(12)(13)(14)(15).
The current study aimed to evaluate two methods PFA-100 & MEA side by side comparing and identifying which one is more sensitive in predicting the risk of intra-op bleeds. We presumed that the use of sensitive, time saving and e cient test for prediction of high-risk patients would govern accurate clinical decisions, leading to better patient outcome and reduced mortality associated with CPB.

Methods:
Patient selection: The study was approved by the Institutional Review Board of College of Medicine, King Saud University, Kingdom of Saudi Arabia. Samples were collected after informed consent indicating the purpose of the study and the right of the participant to withdraw at any time without any obligation towards the study team. Blood samples from (n = 34) adult patients were recruited from KKUH, having routine normal coagulation tests prior to the different cardiac procedures, including valve replacement, coronary artery bypass etc. were collected at two points in time; at the beginning of the CPB and 60 min into CPB. All patients going for CPB procedure were advised to stop antiplatelet therapy ve days prior to the procedure.
Before starting the CPB for every patient, 20 cc blood from central-line access point was collected for complete blood count (CBC), and platelet function tests. The second time-point was 60 min after the start of CPB. The two blood samples at each point in time were collected as one sample in the EDTA tube for CBC, and in the Lithium tube for platelet function tests. Citrated blood in sodium citrate tubes was aspirated. Platelet and blood products were transfused as needed as per the discretion of the cardiac surgeon.
Platelet function testing: PFA-100 analyzer and MEA were used to measure the extent of platelet dysfunction pre-and intra-CPB. Complete blood counts including HCT, Hb and PLT counts were measured using a hematology analyzer (Advia 120; Bayer, Tarrytown, NY, USA), and platelet function analysis by PFA-100 (Siemens Healthcare Diagnostics, Malvern, PA, USA). Disposable cartridges on high shear rates were used for PFA-100, containing aperture inside a membrane coated with cADP or cEPI, which induced platelet activation leading to adhesion and aggregation, resulting in occlusion of aperture. The time needed to occlude aperture and stop the blood ow is called Closure Time (CT). The longer is the closure time; the more platelet dysfunction is expected (16).
Whole blood samples from these patients were also tested by a Multiplate analyzer (Dynabyte, Munich, Germany). Aggregation and adhesion of platelets on electrode surfaces are converted to electrical impedance which is measured by Multiplate analyzer. The test cuvette contained platelet agonists; aspirin (ASP: 15 mM), (TRAP: 1 mM), collagen (COL) (100 lg/mL), arachidonic acid (ADP 0.2 mM), and ristocetin (RISTO: 10 mg/mL). Platelet adhesion after activation results in aggregation on metal sensor causing electrical impedance between electrode wires, which was measured continuously for six minutes and analyzed after conversion to arbitrary units.
We calculated the most important parameter, which is the area under the aggregation curve (AUC). It is affected by the total height of the aggregation curve as well as by its slope and is best suited to express the overall platelet activity, and expressed as AUC*min (14). Lower value indicates more platelet dysfunction De nitions Thrombin receptor-activating peptide (TRAP-6); a potent platelet activator and stimulates platelet aggregation via stimulation of platelet thrombin receptors. Aspirin or Clopidogrel cannot block this action. This means TRAP allows GpIIb/IIIa antagonists to be detected even in samples containing Aspirin (ASP) or Clopidogrel. Adenosine diphosphate (ADP) stimulates platelet activation by the ADP receptors. In COL test, Collagen activates the collagen receptor, which in turn activates the platelets via arachidonic acid production. Ristocetin (RISTO) platelet aggregation test is based on the binding of RISTO-vWF (von Willebrand factor) complexes leading to the aggregation of platelets as well as their activation. Aspirin ® inhibits this activation. This explains the ASP sensitivity of RISTO test.

Data Analysis
Descriptive statistics were computed as a baseline means, standard deviations (SD) and minimum and maximum values for continuous variables. Preoperative variables were compared to intra-operative variables for PFA-100 and MEA. We used skewness-kurtosis normality tests to nd variables with nonnormal distribution. Student t-test compared signi cance in mean values of normally distributed variables, while the Wilcoxon t-test compared signi cance in mean values of non-normally distributed variables. We used the software STATA v.13.0 (Stata Corp., College Station, TX, USA) in our analysis. A statistical signi cance threshold of P < 0.05 was adopted.

Results
Our study compared the effect of CPB on platelet function by using pre-and intra-operative platelets function parameters for 34 patients using PFA-100 analyzer and MEA. Data for two patients were removed due to erroneous PFA-100 values leading to 32 patients' data used for this study. Although all patients going for CPB procedure were advised to stop antiplatelet therapy ve days prior to the procedure, we found 7 patients presenting with high PFA-100, suggesting some residual effect of antiplatelet.
We presumed that pre-and intra-operative measurements should show signi cant change towards worsening of platelet function. Pre-and intra-operative PFA-100, MEA and CBC measurements are summarized in Table- (14).
Next, we tried to answer if PFA-100 pre-op values related to intra-op drop in Hb value, as in, do we expect lower Hb count for elevated PFA-100 values. We assumed two approaches to nd the answer: rst, we look for a correlation between PFA-100 variables and intra-op Hb; secondly, we run a statistical t-test which shows a signi cant change in Hb groups based on a dichotomy of PFA-100 variables. The binary variable was labelled 'abnormal' when both ADP and EPI values exceeded the normal range and labelled 'normal' otherwise.
We used Pearson's correlation and found a weak correlation coe cient between PFA-ADP and intra-op drop in Hb (0.4699), which is illustrated as a scatter plot in Figure- Considering the gender, mean Hb levels in female patients 75 ± 0.5 (m: 73, r: 69-84 g/L) were signi cantly lower (a 16% decrease, P-value 0.0009) than the male patients 87 ± 1.0 (m: 87, r: 71-108 g/L). The same pattern was observed in HCT levels.

Discussion
Bleeding during cardiac surgery procedures makes allogeneic blood transfusion an essential requirement to maintain hemostatic blood components (13); therefore, accurate and timely assessment of patients in need of transfusion is crucial. To meet these requirements, evaluation of platelet function along with platelet count intra-operatively is pivotal. PFA-100 and MEA are two commonly used modalities to monitor platelet function and assess the need for transfusion and blood ow cessation (18). Intraoperative transfusions play a crucial role in the patient outcome; therefore, sensitive and e cient platelet function test should be performed for predicting high-risk patients. The current study compared the predictive value and sensitivity of PF-100 and MEA for the identi cation of high-risk intra-operative patients using the intra-operative drop in Hb as a reference outcome.
The PFA-100® works on the principle of shear stress during vascular injury and provides quantitative data related to high shear force-associated platelet dysfunction with whole blood, and provides accurate and quantitative in-vitro measurements (18,19). In the current study, a signi cant difference (P < 0.05) in platelet function was observed pre-and intra-CPB through PFA-ADP (125.9 ± 69.9 vs 205 ± 74.9), as well as EPI measurements (166.8 ± 72.5 vs 214.2 ± 42.7) with an increase in platelet dysfunction after starting CPB. However, con icting results have been reported previously regarding intra-operative bleeding and its association with platelet dysfunction assessed by PFA-100 (20). In agreement with our ndings, a positive relationship between the extent of bleeding and platelet dysfunction was reported by Ostrowsky et al. (21) and Raman et al. (22). On the other hand, and contrary to our ndings, Fattorutto et al. (23) and Forestier et al. (24) proved that PFA-100 results couldn't distinguish patients at low risk of bleeding from patients who had signi cant bleeding; excluding the high-risk CPB patients. In cardiothoracic procedures involving CPB, speci city and sensitivity of PFA-100 were 94% and 85%, respectively, in identifying response to platelet transfusion. Our PFA-100 results clearly predicted the outcome based on preoperative analyses as measured by correlation with intra-op drop in Hb values via Pearson correlation. Beside the Pearson co-relation test, we used a different yet meaningful approach to distinguish high and low-risk patients by generating binary variables based on values of PFA-100 ADP and EPI. The binary variable is considered abnormal when both ADP and EPI values were out of normal range. This approach is useful as some of the platelet inhibitors such as aspirin and Clopidogrel affect EPI during the test. In such a scenario, there was a signi cant difference between Hb levels of abnormal and normal groups.
MEA is another frequently used modality for platelet function analysis, which can be opted for the nearpatient setting to get prompt results. MEA has certain advantages over PFA-100, including measurement of platelet function in whole blood in the presence of other blood cells for timely assessment (25). Although, in conformity with PFA-100 results, signi cant changes in platelet function were observed through MEA in pre-and intra-CPB, as indicated by a decreased aggregation depicted through less area under the curve. Meyer et al. showed the signi cance of MEA performed at the start of a surgical procedure to improve the management of blood product transfusion (26). A positive correlation between the preoperative values of ADP from MEA and total blood transfusion during surgical intervention was reported in the study of Mengistu et al. (27). However, in the current study, we found the diagnostic performance of MEA as comparatively weaker than that of PFA-100, taking an intra-op drop in Hb as an outcome. The role of MEA in lowering the number of transfusions has not been justi ed previously in the literature. Apropos, in the current study, MEA couldn't provide such bene ts due to lower sensitivity as compared to PFA-100. Contrary to the previous studies(3), we found that PFA-100 is sensitive to detect the changes in platelet dysfunction and can be relied upon to distinguish high-risk patients for bleeding.
PFA-100 also showed high variability pre-and intra-CPB in ADP and EPI; while in MEA, only ADP and RISTO show greater variability intra CPB as compared to pre-CPB values; depicting the higher sensitivity level of PFA-100.
The novelty of the current study lies in the fact that intra-operative transfusion management has been studied considering the intra-op drop in Hb as a primary outcome, which is easy to measure on-site as compared to calculations based on units transfused post cardiac surgery. Limitations of the current study lie in the fact that MEA and PFA-100 results have not been compared with the gold standard platelet function test: Light transmission aggregation(3).
This study does have a limitation of correlating the bleeding to drop in the Hb after 60 minutes of the procedure, which might re ect uid priming the CPB or retrograde priming. Still, the timing of this measurement might be of value to limit the variability and complexity of different procedures, affect surgeons' preferences and allow the analysis of platelet dysfunction in a broader range of cardiac procedures.

Conclusion
The current study reported signi cant worsening in the platelet function 60 min intra-operatively during CPB, measured by two modalities PFA-100 and MEA, which might be reversed by platelet transfusion during surgery. We found PFA-100 to be more sensitive and predictive of intra-op drop in Hb as compared to MEA,    Scatter plot between pre-op PFA-100 EPI and intra-op drop in Hb values.