|Home | About | Journals | Submit | Contact Us | Français|
To determine the prevalence of platelet dysfunction, using an end-point of assembly into a stable thrombus, following severe injury. Background: Although the current debate on acute traumatic coagulopathy (ATC) has focused on the consumption or inhibition of coagulation factors, the question of early platelet dysfunction in this setting remains unclear.
Prospective platelet function in assembly and stability of the thrombus was determined within 30 minutes of injury using whole blood samples from trauma patients at the point of care employing thrombelastography (TEG)-based platelet functional analysis.
There were 51 patients in the study. There were significant differences in the platelet response between trauma patients and healthy volunteers such that there was impaired aggregation to these agonists. In trauma patients, the median ADP inhibition of platelet function was 86.1% (IQR: 38.6–97.7%), compared to 4.2 % (IQR 0–18.2%) in healthy volunteers. Following trauma, the impairment of platelet function in response to AA was 44.9% (IQR 26.6–59.3%), compared to 0.5% (IQR 0–3.02%) in volunteers (Wilcoxon non parametric test p<0.0001 for both tests).
In this study, we show that platelet dysfunction is manifest following major trauma, before significant fluid or blood administration. These data suggest a potential role for early platelet transfusion in severely injured patients at risk for postinjury coagulopathy.
Hemorrhage remains the leading cause of preventable death following trauma,1 and 25% of severely injured patients manifest evidence of coagulopathy on arrival to the emergency department (ED).2,3 Although the current debate on acute traumatic coagulopathy (ATC) has focused on disseminated intravascular coagulation (DIC)4,5,6 versus an acute endogenous coagulopathy mediated by activated protein C (aPC),7,8 the question of early platelet dysfunction remains obscure. The thrombocyte is particularly suspect in the context of the cell-based model of hemostasis, which highlights the critical interaction between the platelet, endothelium, and plasma factors during hemostasis.9 In spite of its importance, early recognition of platelet dysfunction is challenging as conventional plasma-based tests (aPTT, INR) are unable to determine platelet function and are insensitive to coagulopathy unless severely deranged.10 Although the complete blood count with differential provides a platelet count, this quantitative test does not provide an assessment of platelet function.11
Recently, point-of-care viscoelastic analyzers, including modified thrombelastography (TEG) with platelet mapping, have become available to rapidly identify and manage high-risk patients in the trauma bay.12 These same approaches can be employed to measure platelet function at the bedside.13 For example, identifying ADP receptor inhibition >60% in patients on antiplatelet medications identifies those at risk for developing bleeding complications during cardiac surgery,14,15,16 and even modest reductions in platelet function are associated with increased morbidity and mortality following trauma.17
We hypothesized that early platelet dysfunction is prevalent following severe injury, and can be evaluated in a point of care setting using thrombelastography (TEG)-based platelet functional analysis to measure the ability of platelets to assemble into a stable thrombus with different platelet activators.
This was a prospective observational multicenter study conducted at Denver Health Medical Center (DHMC), Denver, CO, and Memorial Hospital of South Bend (MHSB), South Bend, IN. The samples were collected during trauma activations by trained personnel on-call for a prospective study to evaluate the role of thrombelastography (TEG) in the management of postinjury coagulopathy. Of the trauma activations, patients age > 18, anticipated to receive a blood transfusion in the first 6 hr of hospital admission were enrolled in the study.18 Trauma team activation is the highest level response for patients at risk of critical injury. It is triggered prior to or upon patient arrival by emergency medical services (EMS) or the emergency physician for patients with 1) blunt and penetrating injuries with a pre-hospital systolic blood pressure less than 90mmHg, 2) penetrating gunshot wounds to the torso 3) stab wounds to the torso requiring endotracheal intubation, 4) amputation proximal to the wrist or ankle, 5) a Glasgow Coma Scale (GCS) less than 8 or respiratory compromised with presumed thoracic, abdominal or pelvic injury, 6) inter-hospital transfers requiring blood transfusion to maintain vital signs or 7) when the emergency medicine attending or chief surgical resident suspects the patient is likely to require urgent operative intervention.19 The subset evaluation focused on the assessment of platelet function using TEG-based platelet mapping.
Overall, a total of 51 consecutive trauma patients at risk for postinjury coagulopathy with field blood on arrival were enrolled in the study. One patient was excluded for end stage renal disease due to pre-existing platelet dysfunction, the second was an elderly woman excluded due to death from MI following a low speed MVC. Additionally, patients with isolated TBI were not included in the study, nor were patients with end stage liver disease.
Nineteen healthy volunteers were recruited at the Denver and South Bend centers. Data from 20 additional healthy volunteers were generously provided by Dr. William Heaton at North Shore-Long Island Jewish Health System. Healthy males and females over the age of 18 years were used as controls; exclusion criteria were: genetic bleeding disorders, pregnancy, oral contraceptive use, use of anti-platelet or anti-coagulant agents, and recent trauma. Trauma patients on antiplatelet medications (ASA or Plavix) as an outpatient are described separately in the results section. There were no patients on anticoagulants included in this study.
DHMC is a state-designated level I trauma center verified by the American College of Surgeons Committee on Trauma and the academic trauma center for the University of Colorado Denver. MHSB is a level 2 trauma center verified by the American College of Surgeons Committee on Trauma and is the academic trauma center for the Indiana University School of Medicine-South Bend on the Notre Dame Campus. Data collection and storage processes were in compliance with Health Insurance Portability and Accountability Act regulations and had been approved by both institutional review boards. Informed consent form was obtained from normal volunteers and waivers of informed consent were obtained for trauma patients under protocols approved by the respective internal review boards. Clinical data collected included: age, gender, injury mechanism and severity (ISS), systolic blood pressure on admission, as well as units of blood products transfused (PRBCs, fresh frozen plasma (FFP), platelets, cryoprecipitate) and crystalloid infusion within the first 6 hr. The Injury Severity Score (ISS) is derived from the Trauma Registry, calculated at the time of patient discharge by a trauma coordinator. The international normalized ratio (INR), activated partial thromboplastin time (aPTT), and base deficit, were performed in the central laboratory. We defined massive transfusion as > 10 units of PRBCs in 6 hr,20 as this time period has been previously shown to be an independent risk factor for multiple organ failure.21
Whole blood was collected in citrate (3.2%) or heparin (68 U/4 ml) in the field by paramedics or within 30 minutes of arrival to the ED if no field blood was available. Platelet mapping was performed within 2 hr of sample collection. Blood was collected from healthy volunteers and the test performed within 2 hr of sample collection. The mean time from performance of platelet mapping from the time of blood collection was approximately 30 minutes.
Citrated whole blood samples were analyzed at 37° C using a Model 5000 Thrombelastograph Haemostasis Analyzer (Haemonetics, Boston, MA). Kaolin was used as an activator. The following parameters were recorded from the temporal impedance tracings of the TEG: R time (min), angle (α, degrees), coagulation time (K, sec), maximum amplitude (MA, mm), clot strength (G, dynes/cm2) and lysis 30 min after MA (LY30, %). The significance of each parameter has been described.22 The TEG parameter, G, a global reflection of clot strength, was calculated from the amplitude (A), based on a curvilinear relationship: G= (5000×A)/(100−A).
Platelet function was determined using the TEG platelet functional analysis assay, measuring the reactivity of platelets to the activators arachidonic acid (AA) and ADP. This test was designed to monitor antiplatelet therapy, as aspirin exerts its anticoagulant effects via the AA pathway and clopidogrel acts via the ADP (P2Y12) pathway. The contribution of the P2Y12 (ADP) receptor, or the COX-1 (AA) pathway, to the clot formation is assessed by addition of ADP or AA, respectively.23 Platelet mapping is also utilized in cardiac surgery, with preoperative inhibition of ADP predicting the development of microvascular bleeding in patients on clopidogrel.13
Four tests were performed from each collected blood sample: MAThrombin, MAFibrin, MAADP, and MAAA. A heparinized blood sample was collected, with 360 μl of the blood sample placed into the pre-warmed cup of the TEG analyzer, followed by 10 μl of the prepared activator solution, comprised of reptilase/factor XIIIa/phospholipids. Reptilase catalyzes cleavage of fibrinogen to generate fibrin and is used to replace thrombin. Next, either ADP (2 μM, final concentration, MAADP), or arachidonic acid (1 mM final concentration, MAAA) was added as a platelet agonist. To determine the fibrin contribution to the clot, independent of platelets, the activator solution was added to 360 μl of heparinized blood (MAFibrin). The maximum hemostatic activity (MAThrombin) was measured using a kaolin activated whole blood sample collected in citrate, rather than heparin. The citrated whole blood sample was inverted gently 5x and placed on its side, undisturbed, for 30 min before running TEG analyses. Samples were analyzed within 2 hr of blood sample collection. The percent platelet stimulation in response to either the ADP or AA agonist was calculated using the following equation: (100 − [(MAADP(AA) − MAFibrin)/(MAThrombin − MAFibrin)× 100].23 Figure 1 shows a normal TEG platelet mapping tracing (1a), as well as a tracing of a representative trauma patient from our study (1b).
All statistical analyses were performed using SAS vs. 9.3 for Windows (SAS institute, Cary, NC). Chi-square test or the Fisher Exact Test were used for comparisons of categorical variables. For normally distributed variables, data were expressed as mean and standard error of the mean (SEM) and the t-test was used for comparisons. For non-normally distributed variables, data were expressed as median and interquartile range (IQR) and the Wilcoxon non-parametric test was used for comparisons. Specifically, the two main variables of interest (platelet reactivity activator AA and activator ADP) were found to have a non-normal distribution. Multivariate analysis employed logistic regression models for which we built binary outcomes (dependent variables) using two different cutoffs, i.e., the median value (50% inhibition of AA platelet stimulation; and 75% inhibition of ADP platelet stimulation) and a 90% cutoff. Independent variables that were associated with the outcomes with p<0.35 in the unadjusted bivariate analysis were selected to be included in the logistic regression models. Specifically, the following independent variables were tested for inclusion: The dependent variables: (>50% inhibition of AA platelet stimulation: >75% inhibition of ADP platelet stimulation.
The independent variables: ED BD> 8 mEq/L, ED SBP < 70 mmHg, >1 PRBC 0–6hrs, >1 FFP 0–6hrs, >1 PLT 0–6 hrs, ISS>25, blunt mechanism, penetrating mechanism, and presence of head injury. A stepwise selection procedure was used to define best predictors. Goodness of fit of the models was assessed using the C-statistic, which reflects the area under the receiver operating characteristics curve.
There were 51 trauma patients enrolled in the study over a 7-month period, 39 healthy volunteers were used as controls, and 6 patients on antiplatelet therapy will be described elsewhere. Patients in the study group averaged 44 years of age, and 63% were male. Mechanism of injury included 86% blunt trauma, and 20% of patients with an associated head injury. Patients arrived with a mean GCS of 11.9, mean ISS of 19 (mean NISS 27, mean AIS Head of 1). Overall mortality was 4% (2/49); one death was due to traumatic brain injury, the other death was due to exsanguination in the ED.
The patients arrived with a mean systolic blood pressure of 118.3 mmHg, mean hemoglobin of 13 (± 0.3 g/dL), and mean base excess of −6.4 (± 0.84 mEq/L), a mean platelet count of 232 (± 13.2 × 103/μl), and a mean INR of 1.14 (± 0.03). 55% of the patients (27/49) received at least 1 unit of PRBC in the first 6 hours. Patients who received blood, received, on average, 4.37 ± 0.6 units of PRBC, 1.35 ± 0.3 units of FFP and 1.23 ± 0.2 apheresis units of platelets in the first 6 hours. Demographics and on-arrival labs are summarized in Table 1.
In trauma patients, the median ADP inhibition of platelet function was 86.1% (IQR: 38.6–97.7%), compared to 4.2 % (IQR 0–18.2%) in healthy volunteers. Following trauma, the impairment of platelet function in response to AA was 44.9% (IQR 26.6–59.3%), compared to 0.5% (IQR 0–3.02%) in volunteers (Wilcoxon non parametric test p<0.0001 for both comparisons, Figure 2, Figure 3). A representative case of a patient with profound platelet dysfunction following life-threatening trauma is displayed in Figure 4. The coagulopathy was reversed with early platelet transfusion.
Although BD ≥ 8 mEq/L at ED was coupled to increased attenuation of the stimulatory effects of AA and ADP, these effects were only significant for ADP-mediated platelet function (59.6 vs. 97.0 % inhibition of ADP function: BD < 8 vs. BD > 8; Wilcoxon p= 0.013. 35 vs. 49.9 % inhibition of AA function: BD<8 vs. BD > 8, Wilcoxon p= 0.078).
Severity of the anatomic injury correlated with increased inhibition of the AA and ADP stimulation of platelet function; however, similar to the BD, only the inhibition of ADP function was significant (58.9 vs. 97.2 %ADP Receptor Inhibition: ISS < 25 vs. ISS > 25, Wilcoxon p= 0.003. 41.8 vs. 49.8 %AA Receptor Inhibition: ISS 25 vs. ISS > 25, Wilcoxon p= 0.14). Of note, these patients had a normal platelet count on arrival (182 × 103/μl), and normal INR (1.2). Of the patients with traumatic brain injury, the median inhibition of ADP function was 89.4% and the inhibition of AA function was 40.1 %, despite normal platelet counts of 214 × 103/μl, and a normal INR of 1.12.
PRBC transfusion requirement in the first 6 hr appeared to be associated with the early platelet dysfunction, and the inhibition of ADP function correlated with PRBC transfusion in the first 6 hr (59.6 vs. 96.1% inhibition of ADP function: no RBC/0–6 hrs vs. ≥ 1 RBC/0–6 hrs, Wilcoxon p= 0.025, 41.6 vs. 47.9% inhibition of AA function: no RBC/0–6hrs vs. ≥ 1 RBC/0–6 hrs, Wilcoxon p= 0.236).
For ADP>75% inhibition two predictors emerged: ED BD> 8 mEq/L (OR:7.4, 95% CI: 1.4–40.2) and requirement of the transfusion of at least 1 plasma unit within 0–6 hr postinjury (marginally significant OR: 5.1, 95% CI: 0.9–28.9). C-statistic for the % inhibition of AA function model was 0.62, and for the % inhibition of ADP stimulation was 0.74. Using a 90% cutoff for the ADP inhibition revealed two significant predictors, namely ED SBP < 70 mmHg (OR 12.6, 95% CI 1.7–92.4) and requirement of at least 1 RBC in the first 6 hr (OR=8.9; 95% CI 1.7–46.7). C statistic was 0.75, suggesting a moderate goodness of fit. No predictors for AA emerged using either cutoff (50% and 90%)
Six patients on antiplatelet medications (ASA or Plavix) were enrolled in the study. The inhibition of platelet function in stabilizing stable viscoelastic clots was comparable and not additive to the previously described trauma subgroups, with a median inhibition of ADP stimulation of 68.2% (IQR: 56.5–79.5%), and a median inhibition of AA stimulation of 72.5% (34.9–85%).
Despite vigorous investigation, the pathogenesis of acute traumatic coagulopathy remains elusive, although most research has focused on the fluid phase of hemostasis. In this study we show that platelet dysfunction, using the end-point of the viscoelastic stability of the thrombus, is manifest following major trauma, before significant fluid or blood administration. Tissue injury and hemorrhagic shock appear to be dominant risk factors for the platelet inhibition present immediately after injury. The mean platelet count on arrival was 232 (± 13.2 × 103/μl), demonstrating that the platelets are dysfunctional despite ample number.
Platelets play a central role in hemostasis, yet there have been relatively few studies which have evaluated the significance of platelet function in trauma.24 Analysis of combat casualties in the Vietnam conflict first raised the question of early platelet dysfunction in this setting.4,25 Subsequently, preemptive platelet transfusion to attenuate microvascular bleeding in a civilian setting of massive transfusion was suggested, although a prospective randomized controlled trial failed to show a survival benefit. 8 Since that time, attention has been given to prophylactic plasma transfusion in patients at risk for massive transfusion, with recent retrospective studies suggesting that preemptive platelet transfusion may attenuate postinjury coagulopathy.26,27
The results of our study are consistent with the concept that coagulopathy may be initially mediated by platelet hyperactivation in trauma, rendering platelets unresponsive to subsequent stimulation. As widespread ADP release into the systemic circulation resulting from tissue injury and hypoperfusion can lead to pronounced coagulation impairment,28 the platelet abnormality may be an early and important derangement postinjury, with the earliest abnormality of hemostasis revealed by the substantial reduction of platelet reactivity at the ADP site (Figure 2).
The phenomenon of “exhausted platelet syndrome” has been described previously in patients with various medical conditions, including chronic transplant rejection and thrombotic thrombocytopenic purpura, revealing an acquired defect due to circulation of exhausted platelets following prolonged activation in-vivo. 29 In trauma, previous clinical studies have demonstrated attenuation of platelet stimulation to ADP agonism for patients in hemorrhagic shock,30 and following head injury,31 with the diminished platelet response to ADP more prominent in non-survivors than survivors following life-threatening injury.32
Additionally in our study, PRBC transfusion requirement as a surrogate for shock correlated with increasing severity of platelet dysfunction. Acquired platelet dysfunction has been linked to blood transfusion requirements in other fields. For example, point of care platelet function testing has demonstrated utility in predicting blood loss following cardiopulmonary bypass,14,33 with a recent study finding preoperative inhibition of the platelet stimulatory function of ADP > 60% to have a 92% positive predictive value for the development of microvascular bleeding in patients taking clopidogrel.13 Indeed, we found a more substantial effect on the ADP receptor than the AA receptor on coagulation. This finding is consistent with previous reports in the cardiac surgery literature showing that ADP correlates more strongly to surgical bleeding than AA.16 The cardiac literature has shown that quantification of acquired platelet dysfunction can help predict bleeding risk, and it is likely that identification of platelet dysfunction in trauma may have important clinical applications and prognostic significance as well.
With the limitations of conventional coagulation tests in trauma, and increasing use of platelet function analyzers in point of care settings in cardiac surgery, thrombelastography was used as a global assessment of coagulation integrity, with TEG platelet mapping as an assessment of platelet function.34,35 These types of platelet function analyzers support the multidisciplinary approach required to manage a massively transfused patient,36 which becomes a major advantage when compared to conventional tests. Rapid identification and treatment of postinjury platelet dysfunction has negative predictive value as well. A representative case demonstrating improved thrombus formation as a result of goal-directed blood component therapy is shown in Figure 4.
Traditionally, studies of platelet function in such individuals have been performed using Light Transmittance Aggregometry (LTA). LTA is a labor intensive and time consuming test of platelet function that has been considered the gold standard of platelet function assays.37 However, many new platforms have been developed to improve upon some of the drawbacks of LTA testing, i.e. recent studies have demonstrated that platelet aggregation tests can predict increased bleeding risk during and after cardiac surgery. TEG platelet mapping and Multiple Electrode Aggregometry (MEA) have demonstrated good correlation with LTA.38,39,40
This study has several limitations. In this project, we evaluated platelet function in severely patients at risk, defined as trauma activations in the ED. The 4% overall mortality in this cohort has been the typical annual mortality at our centers, but clearly reflects a broad range of injured patients. Future studies will need to focus on those patients who develop coagulopathy. The near-complete inhibition of ADP-mediated platelet aggregation in the most severely injured patients is noteworthy, however, and warrants further investigation. Although platelet dysfunction is present immediately following injury, the duration of platelet dysfunction has not been quantified. Furthermore, as the majority of trauma patients receive blood products (including platelets) during resuscitation, future studies comparing injured patients to a control group who received a transfusion for other reasons would be one way to clarify the role of blood transfusion as an independent contributor to platelet dysfunction. Ongoing studies are investigating the mechanism(s) for early platelet dysfunction as part of a large randomized prospective trial, with the clinical implication being that early presumptive platelet transfusion may be a part of the Massive Transfusion Protocol (MTP).
In summary, significant differences in AA and ADP stimulation of the platelet contribution to clot formation between trauma patients and healthy volunteers were delineated. Therefore, platelet dysfunction may be a sensitive biomarker for the development of coagulopathy and other complications following life threatening injury. The ADP pathway appears to be more dominant than the AA pathway in mediating the platelet dysfunction, and we conclude that platelet dysfunction is present early following injury, before significant fluid or blood transfusion, and continues during the resuscitation period. Shock, tissue injury, and blood transfusion are independently associated with the increasing degree of platelet dysfunction, and further studies to define the use of point-of-care platelet function assays to guide blood component therapy are underway.
The authors would like to thank Jan Howard, MSN, Arsen Ghasabyan, MPH, Harsha Musunuru BS, Robert Cassady BA and Patrick Davis BA for their efforts in collecting data for this study. The authors would like to thank Wm. Andrew Heaton, M.D., Senior Director of Blood Bank and Transfusion Services, North Shore-Long Island Jewish Health System, Rakija Galloway-Haskins, NSLIJ Health System, and Michael Dioguardi, Senior Research Associate Haemonetics Corporation for helping provide data from healthy controls.
Disclosure information: Dr Walsh received speakers fee from Vida Care and consulting fees from CSL Behring and TEG analyzer from Haemonetics. Dr Moore received research support from Haemonetics as principal investigator. All other authors have nothing to disclose. Reagents for this study were received from Haemonetics, and Enzyme Research Laboratories.
NIH Disclosure: Supported in part by NIH Grants T32-GM008315 and P50-GM49222 (Denver) and NIH Grant HL013423 (Notre Dame).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.