Sepsis favors high-on-clopidogrel platelet reactivity.

High-on-treatment platelet reactivity (HPR) is associated with ischemic events in patients on antiplatelet therapy with a history of cardiovascular disease. On the other hand, recent data have associated sepsis with adverse cardiovascular events in patients admitted with bacteremia or respiratory infection. We aimed to assess P2Y12-mediated platelet reactivity (PR) during sepsis and recovery in patients under clopidogrel. This was a prospective observational study. Incoming patients presenting with signs/symptoms of sepsis already on a maintenance dose of clopidogrel of 75 mg qd for cardiovascular events were included in this study. Patients were assessed for their PR on presentation and following septic syndrome, using the VerifyNow point-of-care P2Y12 assay. Patients were excluded in the presence of evidence of noncompliance to antiplatelet regimen or in need of discontinuation during this study. Twenty-two septic patients on clopidogrel were included in this study (Supplemental Figure S1). Clopidogrel was administered for previous stroke, coronary, and peripheral artery disease in 27.3, 40.9, and 31.8% of patients, respectively. The main site of infection was respiratory tract followed by urinary tract, while the same amounts of gram-negative and -positive pathogens were isolated. HPR was noted in 77% and 29% of patients during sepsis and recovery, respectively, presenting a significant decrease in P2Y12 reaction units values during follow-up [240.7 ± 58.3 versus 179.5 ± 58.4, 95% CI (-102.7, -39.76), p = 0.0002]. Five patients died of infection, while no adverse cardiovascular events were noted in our study. Our study shows that sepsis may favor HPR, which is reversed when recovery occurs. This finding may underlie the adverse cardiovascular events in patients admitted with sepsis, possibly requiring alteration of antiplatelet regimen during the inflammation period.

Introduction P2Y 12 receptor antagonists, including clopidogrel, represent the cornerstone for the prevention of ischemic events in patients with cardiovascular disease. Interindividual variability in platelet response to clopidogrel has been commonly reported, while patients with high-on-treatment platelet reactivity (HPR) while receiving clopidogrel are at increased risk of adverse cardiovascular events [1]. On the other hand, sepsis represents a systemic inflammatory response to infection, recently associated with adverse cardiovascular events in patients admitted with community acquired bacteremia [2] or respiratory infection [3].

Methods
The aim of this study was to assess P2Y 12 -mediated platelet reactivity (PR) during sepsis and recovery in patients under clopidogrel. This was a prospective observational study performed at University Hospital of Patras between March 2015 and January 2016, approved by the Regional Research Ethical Committee (24958/19-12-2013) and conducted in accordance with the Helsinki Declaration. Incoming patients presenting with signs/symptoms of sepsis as per international criteria [4], already on a maintenance dose of clopidogrel of 75 mg qd for cardiovascular events, were included in this study. Patients with recent history of bleeding, inherited (e.g., Bernard-Soulier syndrome) or acquired defect of platelets (e.g., idiopathic thrombocytopenic purpura) during the last 3 months prior to recruitment, hematocrit <30 or >52%, platelets <119 or >502 K/μl, or evidence of poor adherence (as confirmed by thorough medical history) were excluded from this study. PR testing was performed with the VerifyNow (Accriva Diagnostics, San Diego, CA, USA) point-of-care P2Y 12 Correspondence: Karolina Akinosoglou, MD, PhD, Department of Internal Medicine and Infectious Diseases, University General Hospital of Patras, 5th Floor Rio 26504, Greece. E-mail: akin@upatras.gr Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/iplt Karolina Akinosoglou -This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. Angelos Perperis -This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. Spyridoula Theodoraki -This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. Dimitrios Alexopoulos -This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. KA designed the study, analyzed data, wrote, and revised the manuscript; AP performed platelet measurements; ST collected samples and assisted in manuscript revision; DA critically corrected manuscript, provided facilities, and shared protocols; and CAG designed the study and corrected the manuscript. assay, at the time of presentation and 28 days postrecovery. Patients who were required to temporarily discontinue clopidogrel during hospitalization or follow-up were withdrawn from this study, as well as patients that fulfilled any exclusion criteria between measurements. Results are reported as P2Y 12 reaction units (PRU), and a value >208 PRU was considered as HPR [5]. Data normality was assessed using the Kolmogorov-Smirnov and Shapiro-Wilk tests using an a of 0.05. A two-sample paired t-test and Wilcoxon signed-rank test where appropriate was used to compare PRU, WBC, CRP, mean platelet volume (MPV), and severity score values levels during sepsis and recovery. Correlations between variables on septic patients were tested by the Spearman rank correlation coefficient value.

Results
Between March 2015 and January 2016, we identified 41 on-clopidogrel patients who presented with signs of sepsis. Twelve patients were excluded, following initial implementation of exclusion criteria, while another seven were withdrawn in due course due to the need of regimen temporary discontinuation during infection or follow-up (Supplemental Figure S1). Twenty-two patients were finally included in this study, and their characteristics and laboratory values upon admission and recovery are shown in Table I and Supplemental  Table SI. HPR was noted in 77% and 29% of patients during sepsis and recovery, respectively, presenting a significant decrease in PRU values during follow-up [mean ± SD of 240.7 ± 58.3 versus 179.5 ± 58.4, 95% CI (-102.7, -39.76), p = 0.0002] (Figure 1). A significant decrease in inflammatory indices, including WBC and CRP, as well as, severity scores was noted between sepsis and recovery; but not in PLT number, MPV, Hb, or modification of diet in renal disease (MDRD) six values (Supplemental Table SI

Discussion
This is the first study to show that on-clopidogrel septic patients exhibit HPR to a very high rate in contrast to PR when in healthy state. This agrees with previous observations showing that, endotoxemia blunts P2Y 12 receptor antagonist effects in healthy volunteers [6]. Mechanisms related to HPR in patients without sepsis include CYP2C19 loss-of-function variant alleles, resulting in reduced drug availability and diabetes mellitus [1,7]. It is possible that inflammatory response could result in CYP2C19 differential expression or posttranscriptional regulation, as noted in cancer patients [8]. On the other hand, it could be possible that increased platelet turnover previously noted during sepsis [9] could be responsible for between-doses-produced platelets not inhibited by clopidogrel. However, the former was not observed in our study, as indirectly reflected in MPV remaining within the normal range and not significantly altered during sepsis and postrecovery. HPR has been previously described in approximately 23% of cases using the method tested here [10]. However, there was a suboptimal clopidogrel response in up to 60% of diabetic patients [11]. Nevertheless, even though such comorbidity could explain HPR in our group, response to P2Y 12 inhibitor proved to be adequate (i.e., PRU < 208) following recovery, hence not supporting potential diabetes role in sepsis-related HPR. Similarly, PR has been previously associated with renal function, even though results have been conflicting [12,13]. In this context, one could argue that alterations in glomerular filtration rate during systemic inflammation could interfere with our results. Indeed, the estimated glomerular filtration rate as calculated by applying MDRD formula in our study group improved as time progressed to recovery; however, no significant difference was noted between two-time point measurements to imply potential role in HPR observed.
Changing to a more potent antiplatelet regimen could overcome HPR upon recognition [14]. It seems that both P2Y 12 receptor antagonists clopidogrel and ticagrelor potently reduce the peak levels of major proinflammatory cytokines, including interleukin-6 (IL-6), tumor necrosis factor-α, and CCL2, well known to drive systemic inflammatory response during sepsis. However, more potent ticagrelor also significantly reduces the peak levels of IL-8 and G-CSF and increases the peak level of anti-inflammatory IL-10, attenuating systemic inflammatory response [15]. This comes in line with clinical data associating ticagrelor with lower mortality related to infection and fewer deaths following sepsis than clopidogrel, in the PLATelet inhibition and patient Outcomes study [14,15]. It Table I  could be possible that sepsis favors platelet activation in a cytokine-dependent manner [16], potentially implicating differential regulation of CYP2C19 expression. Our study suffers limitations including small sample size. However, this size was powerful enough for us to detect significant differences between groups, even though it did not allow for multivariate analysis in order to investigate for other determinants of HPR on sepsis. Second, although P2Y 12 receptors were originally identified to be almost exclusive to platelets, they have now also been identified on a limited number of other cell types including lymphocytes and dendritic cells, the exact implication of which in sepsis remains to be seen [17]. Third, this study solely focused on PR in patients under clopidogrel. Our results cannot be extrapolated for prasugrel-or ticagrelor-treated patients, being under more potent regimens. Last, even though the design of our study controlled well for interindividual variation, change of environment (during hospitalization and discharge) and associated interventions cannot be ignored.
In conclusion, our study provides novel insights into potential mechanisms underlying adverse cardiovascular events in patients admitted with sepsis, possibly requiring alteration of antiplatelet regimen during inflammation period. Further prospective larger randomized studies are required to build on our results and establish potential direct causality.

Declaration of interest
Dr. Alexopoulos has received advisory board fees from AstraZeneca, Boehringer Ingelheim Bayer, and The Medicines Company; and lecture honoraria from AstraZeneca. No other conflict of interest is identified for this manuscript. The rest of the authors report no declaration of interest.