Platelet Function Testing in Patients on Antiplatelet Therapy before Cardiac Surgery

Platelet function plays a pivotal role in occurrence of ischemic events after percutaneous coronary intervention. After vascular injury, circulating platelets adhere to the exposed collagen/von Willebrand factor and undergo activation. Thromboxane A₂ and ADP released by activated platelets amplify platelet activation and aggregation.²⁶  Continuous ADP-P2Y₁₂ receptor signaling sustains glycoprotein IIb/IIIa receptor activation and stable thrombus generation. Therefore, dual antiplatelet therapy is one of the cornerstones after coronary artery stenting.²⁷ 

Evidence from translational research studies suggests a “therapeutic window” concept for ADP-induced platelet reactivity in patients on dual antiplatelet therapy after percutaneous interventions (table 2).^28–30  Preoperative platelet function monitoring may identify those patients with high platelet reactivity to ADP who are candidates to proceed to surgery without delay, whereas those with low platelet reactivity may benefit from even longer waiting times than currently recommended.^19,23,24,29 

Aspirin reduces mortality and ischemic events and improves early venous graft patency after CABG.^31,32  Studies associating aspirin-induced platelet inhibition and coronary artery surgery–related bleeding are lacking. The recent Aspirin and Tranexamic Acid for Coronary Artery Surgery randomized controlled trial demonstrated that compared with placebo, preoperative aspirin neither reduced the composite of 30-day death and thrombotic complications (19.3% vs. 20.4%) nor increased major hemorrhage (1.8% vs. 2.1%), leading to reoperation, presumably because of 50% coadministration of tranexamic acid.^33,34  The most recent guidelines recommend perioperative continuation of aspirin and to restart aspirin as soon as there is no concern over bleeding risk (within 24 h) after isolated CABG (table 1).^17,18 

Accumulating evidence suggests an important relation between P2Y₁₂ receptor–induced platelet inhibition and coronary artery surgery–related bleeding and mortality. In a recent meta-analysis of around 25,000 patients undergoing cardiac surgery, patients with late preoperative discontinuation of clopidogrel (less than 5 days) as compared with clopidogrel-naive patients or patients with early drug discontinuation (5 days or more) had an increased relative risk of reoperation for bleeding (relative risk, 2.5; 95% CI, 1.92 to 3.25) and mortality (relative risk, 1.47; 95% CI, 1.25 to 1.72).⁸  When including observational trials only (n = 2,723), clopidogrel-treated patients had a relative decreased risk of immediate postoperative myocardial infarction as compared with clopidogrel-naive patients (relative risk, 0.63; 95% CI, 0.48 to 0.82), suggesting an anti-ischemic benefit of some degree of platelet inhibition.⁶  However, meta-analysis results are encumbered by variable P2Y₁₂ receptor inhibitor washout periods, inherent limitations in postoperative myocardial infarction diagnosis in acute coronary syndrome patients, the inclusion of mainly nonrandomized and retrospective studies, and substantial heterogeneity among studies.

In a nationwide registry including 2,244 acute coronary syndrome patients undergoing coronary artery surgery during dual antiplatelet therapy, Hansson et al.¹¹  demonstrated significantly less bleeding (Bleeding Academic Research Consortium level 4) in ticagrelor- versus clopidogrel-treated patients (12.9 vs. 17.6%). Within a washout period of less than 24 h, the incidence of bleeding was similarly high (38% vs. 31%), but ticagrelor-treated patients received significantly more red blood cells, platelets, and plasma transfusions. Notably, although there was a continuous decrease in bleeding with increasing days off clopidogrel, there was a sharp decline in bleeding after a 3-day washout period in ticagrelor-treated patients. Bleeding was associated with a 15-fold increased, unadjusted risk of 30-day mortality.^11,12 

The European coronary artery surgery registry is an observational multicenter study prospectively collecting data of 7,491 patients undergoing isolated on- or off-pump revascularization for stable coronary artery disease or acute coronary syndrome.⁹  In a subgroup of 2,311 patients undergoing isolated first-time coronary artery surgery treated with ticagrelor or clopidogrel, the incidence of bleeding stratified by the exact number of days P2Y₁₂ receptor inhibitor were withheld preoperatively was analyzed. After propensity score matching, the universal definition of severe perioperative bleeding sharply declined when ticagrelor was discontinued for 3 days as compared to 0 to 2 days (2.7% vs. 16%, P = 0.003). Even 24 h of waiting between last ticagrelor intake and surgery reduced bleeding by more than 50% (25% vs. 10.8%, P = 0.033). In clopidogrel-treated patients, severe bleeding gradually decreased over time with almost twofold greater bleeding with 0 to 3 days discontinuation versus 4 to 5 days (15.6% vs. 8.3%, P = 0.031).^7,10 

The key features of platelet function assays used in cardiac surgery are presented in table 3. Because the rationale is to assess antiplatelet response to P2Y₁₂ receptor blockers, ADP-induced platelet aggregation is the focus of the various assays. Light transmittance aggregometry is the accepted gold standard and measures the change in light transmittance in platelet-rich plasma after ADP stimulation. The results are reported as percentages of aggregation. Light transmittance aggregometry is a laborious method that requires skilled technicians more than 1 h of processing time.³⁵  The Multiplate analyzer assay (Dynabyte, Germany) measures electrical impedance in whole blood after adherence and aggregation of activated platelets on electrodes after stimulation by ADP. The results are presented as aggregation units. This assay requires around 500 µl of blood and minimal pipetting and provides rapid evaluation of platelet aggregation in less than 20 min.³⁶ 

The VerifyNow P2Y₁₂ assay (Instrumentation Laboratory, USA) is a point-of-care, turbidimetric-based optical detection system that assesses the antiplatelet response to P2Y₁₂ inhibitors in whole blood. The change in light transmittance after binding of ADP-stimulated platelets to fibrinogen-coated beads is recorded as P2Y₁₂ reaction units.^37,38  In this assay, a tube containing ~2 ml of blood is directly inserted into the device obviating the need for pipetting. The results are obtained within a few minutes.

Thromboelastography/thrombelastometry provide viscoelastic measurements of the evolving clot from the time of initial platelet-fibrin clot formation to clot strengthening to clot lysis.³⁹  Both TEG (Haemonetics Inc., USA) and ROTEM (Instrumentation Laboratory) employing kaolin or tissue factor are not suitable to measure the antiplatelet response to a P2Y₁₂ inhibitor. However, modification of the TEG assay to stimulate platelet aggregation with ADP and to generate fibrin in the absence of thrombin generation is possible by use of a specific PlateletMapping assay. In this assay fibrin is generated by reptilase and crosslinked by factor XIII. Thus, ADP-induced platelet-fibrin clot is formed in the absence of thrombin in heparinized blood.²³  TEG6s (Haemonetics Corp., USA) is a point-of-care based on microfluidics technology and is approved by the Food and Drug Administration in the United States. It is a cartridge-based assay that requires minimal pipetting, takes ~45 min, and provides the same information on platelet-fibrin clot formation as TEG described above.³⁹ 

Other assays include the Platelet Function Analyzer 100 and the Platelet Function Analyzer 2Y (Seimans, USA) that assesses P2Y₁₂ reactivity in whole blood under shear. Platelet Works is an assay that measures single platelet counts after agonist stimulation with ADP (Helena Laboratories, USA).^3,24 

As evidenced above, the underlying principles and agonist concentrations of platelet function assays differ, and a poor correlation has been reported between assays.³⁷  Moreover, platelet function assays are limited when platelet counts fall less than 100 × 10⁹/l.⁴⁰ 

Currently, there is not enough evidence from surgical patients to prefer one test over the other. Among these laboratory methods, TEG6s with the PlateletMapping assay may be more appropriate in coronary artery surgery patients because it can be performed at the bedside with minimum sample handling despite being an expensive test as compared with other assays ($125 compared to less than $100 for other assays). It provides detailed information on the viscoelastic properties of clot in addition to the response to P2Y₁₂ receptor blockers. Other assays provide antiplatelet response to P2Y₁₂ receptor inhibitors only. Another consideration would be the VerifyNow P2Y₁₂ assay, given its ease of use as a true bedside assay.

A number of small observational studies enrolling a total of ~2,000 patients undergoing cardiac surgery procedures during dual antiplatelet therapy suggested an inverse association between preoperatively measured platelet reactivity levels, as assessed by various platelet function assays and increased on-pump and off-pump cardiac surgery–related bleeding (table 4).^{23–25,41–54}  Kwak et al.⁴²  demonstrated an association between clopidogrel responsiveness and off-pump coronary artery surgery–related bleeding, irrespective of whether dual antiplatelet therapy was discontinued 1 or 3 days preoperatively. Patients in the highest tertile of platelet response to clopidogrel as assessed by the thromboelastography PlateletMapping assay (more than 76.5% inhibition) had higher 24-h chest tube output and higher transfusion rates as compared with patients in the other two tertiles (P < 0.001 for both). For postoperative transfusion requirements, 70% platelet inhibition was the optimal cutoff (area under the curve [AUC], 0.71; 95% CI, 0.647 to 0.868).⁴² 

Ranucci et al.⁴⁶  evaluated the association between platelet inhibition and bleeding in patients undergoing cardiac surgery within a mean of 4 days (± 3.2) of last P2Y₁₂ receptor inhibitor administration. As per institutional protocol, patients underwent urgent surgery irrespective of platelet function and underwent elective surgery if platelet function was acceptable (Multiplate, ADPtest, more than 31 units), even if discontinuation was less than 5 days. All patients received tranexamic acid. Notably, the incidence of bleeding remained substantially unchanged within a wide range of moderately decreased to normal platelet aggregation but rapidly increased with low platelet reactivity. The universal definition of severe perioperative bleeding occurred in 7.5% of the patients. The optimal cutoff was 22 units for the ADPtest, albeit with an AUC of only 0.62%. The negative and positive predictive values were 94% and 20%, respectively. The remarkably low positive predictive value underlines the multifactorial nature of surgery-related bleeding, which may be aggravated by consumption of coagulation factors including fibrinogen, residual heparin, and surgical sources. In a multivariable logistic regression analysis, the defined cutoffs of platelet inhibition remained significantly associated with bleeding; in addition, isolated coronary artery surgery resulted in less bleeding than other types of cardiac surgery.⁴⁶  Notably, the cutoff for the ADP test as identified by Ranucci et al.⁴⁶  was only slightly higher than “low platelet reactivity” associated with increased bleeding in patients with coronary artery disease after stenting.^28,29 

Similarly, Malm et al.⁵¹  identified an optimal cutoff of 22 units by Multiplate ADPtest (AUC, 0.73%; 95% CI, 0.63 to 0.84; P < 0.001) for universal definition of severe perioperative bleeding in 90 patients undergoing coronary artery surgery within 5 days of last ticagrelor intake. The positive predictive value of this cutoff was 63%, and the negative predictive value was 85%. All patients received tranexamic acid. Severe bleeding occurred in 36% of the patients. A multivariable analysis accounting for potential confounders, however, is lacking.⁵¹ 

The different positive/negative predictive values as reported by Ranucci et al.⁴⁶  and Malm et al.⁵¹  despite an identical optimal cutoff in ADPtest may be explained by the different incidence of severe bleeding (7.5% vs. 36%) and the different distribution of platelet inhibition inherent to different study design. In these two studies,^46,51  10% versus 42% of the patients had platelet function less than 22 units, respectively. The reported differences in positive predictive values (20% vs. 63%) emphasize the current limitations of predicting surgery-related bleeding merely based on measured platelet function and support the need for large-scale, well controlled validation studies with standardized perioperative management and bleeding endpoints to provide reproducible bleeding cutoffs.¹⁵ 

Recently, our group evaluated the association between tertile distribution of preoperative on-treatment platelet reactivity as assessed by various platelet function tests and bleeding in 149 high-risk patients (EuroSCORE, 8.4 ± 4.3) undergoing isolated on-pump coronary artery surgery within 48 h of last dual antiplatelet therapy with aspirin and a P2Y₁₂ receptor inhibitor.⁵²  All patients received tranexamic acid. The primary endpoint was perioperative erythrocyte loss (table 5) calculated as follows: (blood volume × preoperative hematocrit × 0.91) − (blood volume × hematocrit × 0.91 on postoperative day 5) + (ml of transfused red blood cells × 0.59). A factor of 0.91 was applied to correct hematocrit of peripheral blood sampling, and the factor 0.59 accounts for the average hematocrit of red blood cell units.^55–58 

This calculation accounts both for hidden blood loss and preoperative anemia, occurring in around 20% of patients undergoing cardiac surgery.⁵⁵  The secondary endpoint was bleeding at level 4 per the Bleeding Academic Research Consortium scale decreased with increasing tertiles of platelet reactivity as assessed by light transmittance aggregometry (% ADP-induced aggregation; first tertile, 0 to 13.7; second tertile, 13.7 to 25; third tertile, 25 to 100). Multivariable analysis identified tertile distribution of platelet reactivity, comorbidity, as reflected by the EuroSCORE, aspirin dose, and cardiopulmonary time as significant predictors of coronary artery surgery–related bleeding.⁵² 

Cardiopulmonary bypass induces hemostatic activation, causing platelet dysfunction.^4,59  In acute coronary syndrome patients undergoing on-pump revascularization during dual antiplatelet therapy with aspirin and ticagrelor (n = 74), Björklund et al.⁵⁴  evaluated whether postoperative platelet function testing, reflecting antiplatelet therapy and bypass-induced alterations, outperforms preoperative platelet function testing with respect to the prediction of bleeding. Poor postoperative platelet function was associated with severe bleeding with accuracy, however, being comparable to that of preoperative platelet function (ADPtest AUC, 0.75; 95% CI, 0.62 to 0.87 vs. 0.77; 95% CI, 0.65 to 0.89).

An individualized strategy for preoperative clopidogrel discontinuation based on preoperative platelet function testing has been demonstrated to reduce both preoperative waiting and bleeding.^23,24  The Timing Based on Platelet Function Strategy to Reduce Clopidogrel-Associated Bleeding Related to Coronary Artery Bypass Grafting (TARGET CABG) study, a nonrandomized prospective study, stratified clopidogrel-treated patients into time-based platelet function recovery groups as determined by preoperative thromboelastography PlateletMapping assay.²³  All patients were on aspirin therapy, which was continued until surgery. Surgery was scheduled within 1 day in those with a maximum amplitude_ADP of more than 50mm (high platelet reactivity), within 3 to 5 days in those with a maximum amplitude_ADP of 35 to 50 mm (intermediate platelet reactivity), and after 5 days in those with a maximum amplitude_ADP of less than 35 mm (low platelet reactivity). In the absence of a validated bleeding cutoff, we hypothesized that a cutoff associated with short- and long-term occurrence of ischemic events in previous studies involving patients treated with coronary stents would serve as a surrogate measure for adequate surgical hemostasis.⁶⁰  Mean 24-h chest tube drainage (primary endpoint) in clopidogrel-treated patients was 93% (95% CI, 81 to 107%) of the amount observed in clopidogrel-naive patients, and the total amount of red blood cells transfused (secondary endpoint) did not differ between groups (1.80 units vs. 2.08 units, respectively, P = 0.540). The total waiting period in clopidogrel-treated patients was 2.3 days (mean, 2.7 days/patient) and was 46% shorter than the current guideline-recommended waiting period. This was largely due to 31% of patients with high platelet reactivity who underwent CABG within 1 day after screening platelet function testing.²³ 

The benefits of individualized preoperative waiting were further corroborated by Mannacio et al.,²⁴  who evaluated the role of Innovance 2Y assay to define the optimal preoperative discontinuation period in patients undergoing off-pump coronary artery surgery. They compared patients undergoing individualized waiting targeted at complete recovery of platelet function as reflected by a closure time of more than 106 s with two propensity score–matched groups comprising 100 patients undergoing coronary artery surgery after the recommended 5 days clopidogrel washout period and 100 clopidogrel-naive patients. Overall, 20% of the patients had low platelet inhibition enabling surgery early after admission, and 20% displayed persistent platelet inhibition even after 5 days off clopidogrel. Individualized waiting significantly reduced bleeding (36-h chest tube drainage and number of transfusions), shortened preoperative waiting by 38%, and saved 280 days of total hospital stay as compared with guideline-based management.²⁴ 

The 2017 European Society of Cardiology guidelines on Patient Blood Management for Adult Cardiac Surgery recommend antifibrinolytic therapy (class I level A) and perioperative treatment algorithms for the bleeding patient based on point-of-care testing to reduce the number of transfusions (class IIa level B). However, specific cutoff values for the viscoelastic assays to trigger therapeutic interventions are not provided by these guidelines.¹⁷  Moreover, the cutoffs vary in the randomized and observational studies that are included in two meta-analyses including around 8,000 patients and the multicenter Transfusion Avoidance in Cardiac Surgery trial supporting these guidelines.^3,61–63 

In the Transfusion Avoidance in Cardiac Surgery trial, Karkouti et al.³  demonstrated the benefits of an algorithm sequentially implemented in 7,402 consecutive patients undergoing cardiac surgery in 12 hospitals. The algorithm is based on postcardiopulmonary bypass sponge weight as an objective measure of bleeding and Plateletworks and ROTEM as an objective in vitro assessment of coagulation performed after rewarming and suggests a reevaluation of ROTEM after each treatment step unless severe bleeding occurs. As compared with 3,555 patients who underwent surgery during the control phase, implementation of the algorithm in 3,847 patients was associated with a reduced relative risk of transfusion of red blood cells (relative risk, 0.91; 95% CI interval, 0.85 to 0.98) and platelets (relative risk, 0.77; 95% CI, 0.68 to 0.87) and major hemorrhage (relative risk, 0.83; 95% CI, 0.72 to 0.94) after adjusting for hospital, time of algorithm implementation, and prespecified patient-specific risk factors such as demographic variables, type and urgency of surgery, cardiopulmonary bypass time, and use of cell salvage and tranexamic acid.³ 

Although algorithms reduced transfusions, neither the above referenced meta-analyses nor the Transfusion Avoidance in Cardiac Surgery study demonstrated an impact on mortality.^3,61,62  However, subgroup analysis of the Transfusion Avoidance in Cardiac Surgery study demonstrated that bleeding requiring at least five erythrocyte transfusions was associated with an increased relative risk (relative risk, 41; 95% CI, 13 to 132) of mortality.¹  This dose-dependent relationship between bleeding and outcome underscores the importance to prevent severe bleeding and to immediately treat emerging coagulopathy by targeting the respective pathways.² 

In patients on aspirin alone (n = 76) or aspirin and a P2Y₁₂ receptor inhibitor within 5 days before coronary artery surgery (n = 179), Aggarwal et al.²⁵  demonstrated the benefits of a transfusion algorithm including platelet function testing and functional fibrinogen. Patients were randomized to preoperative platelet function monitoring with either Multiplate (Group A) or TEG PlateletMapping (group B) with pre- and postpump assessment of functional fibrinogen or routine care (group C). Patients were managed according to a standardized protocol including transfusion triggers and tranexamic acid use. In cases of postpump microvascular bleeding in group A and B patients, preoperatively assessed platelet inhibition triggered transfusion of one or two platelet pools, based on two arbitrarily defined cutoffs (Multiplate ADP test less than 30 and less than 40 units, respectively; PlateletMapping assessed platelet inhibition more than 60 and more than 70%, respectively). Group C patients underwent routine clinical care based solely on a postpump TEG kaolin and kaolin-heparinase tests. In patients on P2Y₁₂ receptor inhibitors, platelet function monitoring guided management and extended TEG evaluations significantly increased transfusion of platelets but reduced overall transfusions, and reexploration for bleeding as compared with control. This beneficial effect was accompanied by a cost savings of almost 50%.²⁵ 

The Society of Cardiovascular Anesthesiologists Continuous Practice Improvement Subcommittee recently published a summary of recommendations for blood management in cardiac surgery providing a “best practice advisory that can be easily adopted by clinicians.”⁶⁴  This advisory suggests a stepwise escalating approach starting with complete heparin reversal followed by assessment of platelets, fibrinogen and coagulation factor deficiencies, and continuation of antifibrinolytic therapy beyond the operation room in case of microvascular bleeding. In patients on P2Y₁₂ receptor inhibitors, preoperative discontinuation is recommended, and the possibility of performing preoperative platelet function testing to time surgery is addressed.⁶⁴ 

Figure 1 summarizes:

1. Current recommendations addressing preoperative platelet function testing to time surgery in patients on P2Y₁₂ receptor inhibitors

2. A stepwise intraoperative use of ROTEM/TEG in postpump microvascular bleeding

3. The possibility of preoperatively identification of drug-induced platelet inhibition to specifically target hemostasis pathway in postpump microvascular bleeding as successfully demonstrated by Agarwal et al.^{15,17,18,25,63–65} 

Although time to clot formation differentiates between coagulation factor deficiencies and residual heparin, clot strength differentiates between low fibrinogen and low platelet count, thus enabling quick and targeted hemostatic therapy. However, thresholds for bleeding cutoffs may vary depending on institutional expertise.¹⁷ 

Suggested thresholds for platelet transfusion in post cardiopulmonary bypass bleeding are a platelet count of less than 50,000/μl and less than 100,000/μl in severe ongoing bleeding and/or evidence of platelet dysfunction.⁶⁴  In the latter case, desmopressin maybe considered, although without robust effect on transfusion requirements.⁶⁴  The efficacy of platelet transfusion in antagonizing P2Y₁₂ receptor inhibitors with respect to type of P2Y₁₂ receptor inhibitor and discontinuation period was recently addressed in a comprehensive review.⁶⁶ 

Because severe surgery-related bleeding is clearly linked to increased morbidity and mortality, the prevention of its occurrence and the immediate correction of coagulopathy presents a pivotal clinical goal. Although current guidelines only issue class IIa or IIb recommendations for preoperative platelet function assays, there is accumulating evidence of an association between platelet inhibition and increased coronary artery surgery–related bleeding.

Preoperative platelet function testing may identify the patient at risk for increased surgery-related bleeding, may individually guide the timing of elective surgery after antiplatelet therapy cessation, and may trigger targeted postpump hemostatic therapy when incorporated in an intraoperative blood management concept together with viscoelastic assays. However, a universal cutoff point of platelet function associated with bleeding, although demonstrated in patients after coronary artery stenting, has not been well defined and validated so far in the surgical patient. Moreover, the optimal assay for its measurement also remains elusive. Confounders like comorbidity, the experience of the treating physicians, transfusion practice, cardiopulmonary bypass time, and use of antifibrinolytics all affect the severity of surgery-related bleeding in addition to platelet inhibition.

Larger prospective studies employing consensus-based bleeding endpoints are needed to establish the utility of platelet function testing in patients presenting for cardiac surgery during antiplatelet therapy. These studies are also needed to establish a potential therapeutic window for on-treatment platelet reactivity that limits both bleeding and myocardial ischemia. An individualized preoperative waiting time and a targeted intraoperative blood management concept are expected to substantially shorten the length of hospital stay, reduce costs, and improve patient outcomes.

Support was provided solely from institutional and/or departmental sources.

Dr. Gurbel reports receiving grants from the National Institutes of Health (Bethesda, Maryland), BayerHealthCare, LLC (Hanover, New Jersey), Medicure Inc. (Winnipeg, Canada), Instrumentation Laboratory (Bedford, Massachusetts), US WorldMeds, LLC (Malvern, Pennsylvania), Haemonetics (Braintree, Massachusetts), Amgen (Thousand Oaks, California), Idorsia Pharmaceuticals (Cherry Hill, New Jersey), Ionis Pharmaceuticals (Carlsbad, California), Janssen (Raritan, New Jersey), and Merck (Kenilworth, New Jersey); receiving honoraria and payment for lectures and consultations including service on speakers’ bureaus from BayerHealthCare, LLC, Janssen, Merck, UptoDate (Waltham, Massachusetts), and Medicure Inc.; and holding patents in the area of personalized antiplatelet therapy and interventional cardiology. Dr. Tantry reports receiving honoraria from Medicure Inc., AstraZeneca (Wilmington, Delaware), and UptoDate. The other authors declare no competing interests.