Lower Coated-Platelet Levels Are Associated With Increased Mortality After Spontaneous Intracerebral Hemorrhage
Background and Purpose—Coated-platelets are highly procoagulant platelets observed on dual-agonist stimulation with collagen and thrombin. Coated-platelet levels are decreased in patients with spontaneous intracerebral hemorrhage when compared with controls and inversely correlated with bleed volume. We sought to investigate whether coated-platelets are associated with increased mortality at 30 days after spontaneous intracerebral hemorrhage.
Methods—Coated-platelet levels were assayed in 95 consecutive patients with spontaneous intracerebral hemorrhage. The main outcome was mortality at 30 days according to coated-platelet levels at enrollment. Subjects were grouped into tertiles of the observed coated-platelet level distribution. Groups defined by tertile of coated-platelet level were compared using either ANOVA or a Kruskal–Wallis test for small group size for continuous measures and an exact Cochrane–Armitage trend test for categorical measures. Logistic regression was used to estimate the adjusted odds of death within 30 days associated with coated-platelet levels.
Results—Cumulative mortality at 30 days was 23% (22 subjects). Mortality at 30 days differed among the coated-platelet tertiles: 44% for the first tertile (lowest coated-platelet levels), 19% for the second tertile, and 6% for the third tertile (trend test; P=0.0004). Logistic regression examining the association between mortality and coated-platelet levels showed that the odds of death at 30 days in those with levels <27% (n=47) were 6.83× the odds for patients with levels ≥27% (95% confidence interval, 2.10–22.23).
Conclusions—These results support a link between impaired coated-platelet potential and outcome in intracerebral hemorrhage.
Coated-platelets are a previously unrecognized class of activated platelets, detected only after simultaneous stimulation with 2 agonists, thrombin and collagen, or thrombin and convulxin, a collagen receptor agonist from tropical rattlesnake venom.1,2 Initially referred to as COAT-platelets, an acronym for collagen and thrombin activated platelets3; this subset of activated platelets expresses surface phosphatidylserine, supports a robust prothrombinase activity, and retains high levels of several procoagulant proteins on the cell surface, including factor V, fibrinogen thrombospondin, fibronectin, and von Willebrand factor.1,2,4–6 Because thrombin generation is the central event in coagulation, coated-platelets are considered to be prothrombotic.2
In healthy controls, coated-platelet levels represent ≈32% of the entire platelet population.7 In previous studies, we found that patients with spontaneous intracerebral hemorrhage (SICH) have significantly lower coated-platelet levels shortly after the hemorrhagic event compared with normal controls, and these levels inversely correlate with the size of the bleed.8,9 Prompted by the link between lower coated-platelet levels and increased bleed volume in SICH,9 we initiated a prospective study investigating whether lower levels of coated-platelets measured at the time of the initial hemorrhage would be associated with increased mortality.
The study was approved by the Institutional Review Board of the University of Oklahoma Health Sciences Center. Individual informed consent was obtained for all study participants in accordance with the University of Oklahoma Health Sciences Center and the Oklahoma City Veterans Administration Medical Center rules and regulations. In cases where the patient was unable to provide consent, informed consent was obtained from the patient’s next of kin with legal power to provide such documentation.
Consecutive patients with SICH were recruited from the Neurology service at the University of Oklahoma Health Sciences Center and the Veterans Administration Medical Center for a period of 29 months. All patients were diagnosed with SICH by a board certified neurologist and underwent brain computed tomography at presentation.10 A neuroradiologist provided a definitive reading for every scan obtained. The neuroradiologist reading the imaging studies and the neurologist establishing the diagnosis and providing care for the patient during the admission were not aware of the coated-platelet measurements.
We excluded patients with (1) concurrent use of anticoagulation, (2) prolonged coagulation studies, (3) dementia, (4) >96 hours between the onset of the symptoms and enrollment, (5) platelet transfusion before enrollment, or (6) hemorrhage secondary to trauma, tumor, hemorrhagic transformation of cerebral infarct, rupture of an aneurysm, or vascular malformation. These exclusion criteria were prompted by criteria for diagnosis of SICH, the observation that changes in coated-platelet production in animals after physiological manipulation require a minimum of 4 days to manifest,2 the potential confounding effect of platelet transfusion,11 and previously published coated-platelet abnormalities in Alzheimer disease.12
Smoking status, sex, race/ethnicity, age, and use of medications that may influence coated-platelet levels, such as selective serotonin reuptake inhibitors, 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors (statins), or antiplatelet medications,7,13 were recorded at the time of enrollment for each patient and reflected prehospitalization status and medication use. Additional clinical information recorded at the time of enrollment included a history of diabetes mellitus, hypertension, stroke, transient ischemic attack, traumatic brain injury, or end-stage renal disease as previous research has noted abnormal coated-platelet levels in these conditions.14–19 Hematologic parameters recorded included total platelet count, mean platelet volume, white-blood cell count, and hemoglobin, all obtained on the day of the coated-platelet measurement.
Mortality at 30 days, along with initial bleed volume, hematoma expansion, bleed location (lobar/deep/infratentorial), the presence of intraventricular hemorrhage, initial Glasgow Coma Scale and National Institutes of Health Stroke Scale scores, do not resuscitate status, and withdrawal of life support during admission were obtained through review of medical records at the time of discharge, medical records from outside facilities, if the patient was transferred to another medical facility, or progress notes from the outpatient stroke clinic or the primary care provider. Bleed volume was determined from the computed tomographic scan by using the ABC/2 formula in which A is the greatest diameter on the largest hemorrhage slice, B is the diameter perpendicular to A, and C is the approximate number of axial slices with hemorrhage multiplied by the slice thickness.20 All cases of hematoma expansion were established by the treating neurologist, confirmed by repeat brain computed tomographic studies, and recorded in the medical records.
After consent, 5 mL of blood was drawn into a plastic syringe containing 0.5 mL of acid citrate dextrose, and platelet rich plasma was prepared as previously described.2,21 Coated-platelets were assayed with 1 µL of platelet rich plasma in a 100-µL assay with the following reagents (final concentrations): 1.0 µg/mL of biotin–fibrinogen, 0.4 mmol/L of gly-pro-arg-pro-amide, 500 of ng/mL convulxin, 0.5 U/mL of bovine thrombin, 2 mmol/L of CaCl2, 1 mmol/L of MgCl2, 150 mmol/L of NaCl, and 10 mmol/L of N-(2-hydroxyethyl)-piperazine-N′-(4-butanesulfonic acid); pH 7.5. After 5 minutes at 37°, 0.8 µg of phycoerythrin–streptavidin and 0.5 µg of fluorescein isothiocyanate–abciximab were added. After an additional 5 minutes at 37°, the reaction was stopped with 0.2 mL of 1.5% (wt/vol) formalin in 150 mmol/L of NaCl, 10 mmol/L of N-(2-hydroxyethyl)-piperazine-N′-(4-butanesulfonic acid), pH 7.5. The percentage of abciximab-positive events (platelets) with bound biotin–fibrinogen was quantitated by flow cytometry. Results are reported as percent of cells converted to coated-platelets. Individuals performing the coated-platelet assay were not aware of the clinical diagnosis corresponding to the blood sample analyzed.
Data were analyzed using SAS (SAS System for Windows, version 9.4; SAS Institute Inc, Cary, NC). Descriptive statistics were used to summarize the distribution of coated-platelet levels and other patient and clinical variables. Means were compared between patients who did and did not die within 30 days using a 2-sample t test, and medians were compared using a Wilcoxon signed-rank test for skewed distributions. Proportions were compared between patients who did and did not die within 30 days using a χ2 test or Fisher exact test when the expected frequency count was low (at least 20% of the expected counts <5 or any expected counts of 0). Subjects were grouped into tertiles of the observed coated-platelet level distribution. Comparisons among groups defined by tertile of coated-platelet levels were made using ANOVA or a Kruskal–Wallis test for small group sizes and for continuous measures and an exact Cochrane–Armitage trend test for categorical measures. Logistic regression, with adjustment for individual covariates, was used to estimate the adjusted odds of death within 30 days associated with coated-platelet level. For regression modeling purposes, coated-platelet level was dichotomized at the median given the small number of deaths that were observed among patients with coated-platelet measures in the highest tertile (n=2 deaths). Estimates from the regression model are presented as the odds ratio (OR) followed by the 95% confidence interval for the OR. A confounding variable was defined as one that on adjustment in the model, the OR for death associated with low coated-platelet measures changed by >10%.22 Mediating factors, such as bleed volume or hemorrhage severity, that may represent mechanisms linking coated-platelet levels to mortality, were not considered as possible confounding factors of the association between coated-platelet levels and mortality. Potential confounding factors included age; sex; medication use, including statin therapy, antiplatelets, and selective serotonin reuptake inhibitors; laboratory measures, including platelet count, white-blood cell count, hemoglobin, and mean platelet volume; previous medical conditions, including diabetes mellitus, hypertension, and previous stroke/transient ischemic attack; days from onset; and smoking. Given the small number of deaths within 30 days (n=22), the model was adjusted for each potential confounding factor 1 at a time. A 2-sided, Bonferroni-adjusted α-level of 0.0167 was used when making pair-wise comparisons among the coated-platelet tertile groups but was otherwise set at a 2-sided level of 0.05.
A total sample size of 90 subjects was targeted to achieve 81% power to detect odds of death at 30 days of ≥2.0 associated with a 1-SD decrease in coated-platelet levels. This calculation assumes a SD of 13% in coated-platelet levels, a mortality rate of 25% at 30 days, and a 2-sided α-level of 0.05. The event rate and SD estimates were based on previously published research data for mortality at 30 days and coated-platelet levels in SICH. Sample size calculations were performed using PASS software.23
Between January 1, 2012 and May 30, 2014, 187 patients were screened for SICH and 99 were enrolled. Of these, 4 without adequate blood samples were excluded from final analysis. Reasons for ineligibility and exclusion are presented in Figure 1. Of these 95 patients analyzed, 30 (32%) were military veterans, resulting in a larger percentage of men in our study as a result of the composition of the US armed forces during the time these veterans served.
Demographic variables, coated-platelet levels, relevant comorbidities, and pertinent medications for all patients are listed in Table 1.
Mortality at 30 days was 23% (22 of the initial 95 patients), median bleed volume on admission was 17 mL (interquartile range, 6 to 35 mL), with a range between 1 and 166 mL, and hematoma expansion was noted in 9 patients (9%). In all patients, the primary cause of death was listed as intracerebral hemorrhage.
Subjects were categorized into tertiles based on the observed coated-platelet distribution (<21%, 21%–33%, and >33%). The cumulative incidence of mortality at 30 days according to coated-platelet tertile is presented in Figure 2. Death at 30 days was more common among patients in the lowest tertile of coated-platelets (14 deaths, 44%) compared with patients in the middle tertile (6 deaths, 19%) and the highest tertile (2 deaths, 6%) of coated-platelets (trend test; P=0.0004; Figure 2A). Pair-wise comparisons, using an adjusted α-level of 0.0167, suggest that the risk of death at 30 days is higher among patients in the lowest coated-platelet tertile compared with those in the highest tertile (P=0.001) and that the risk does not differ significantly between subjects in the middle and highest tertile groups (P=0.15) or between subjects in the lowest and middle tertile groups (P=0.058).
Table 2 summarizes the distribution of patient characteristics, including the use of medications that may influence coated-platelet levels (selective serotonin reuptake inhibitors, statins, or antiplatelet medications),7,13 at baseline according to coated-platelet tertile. There were no significant differences among the groups in terms of baseline characteristics.
Tables 3 and 4 summarizes hematologic parameters (total platelet count, mean platelet volume, white-blood cell count, and hemoglobin), bleed volume, time from onset to sampling, Glasgow Coma Scale score, National Institutes of Health Stroke Scale score, bleed location (lobar/deep/infratentorial), presence of intraventricular hemorrhage, presence of hematoma expansion, and mortality at 30 days according to coated-platelet tertile. Patients in the lowest tertile of coated-platelets (<21%) had a larger bleed volume, higher white-blood cell count, lower Glasgow Coma Scale score, higher National Institutes of Health Stroke Scale score, and were more likely to have intraventricular hemorrhage than those in the middle and highest tertiles. Hematoma expansion was confirmed by imaging studies in 9 patients and was found to occur more commonly (8/9) in patients within the lowest tertile of coated-platelets (Figure 2B). A summary of the association between demographic characteristics, clinical parameters, hematologic measures, hemorrhage severity indicators, and mortality at 30 days is presented in Table I in the online-only Data Supplement.
Logistic regression, with adjustment for individual covariates, was used to estimate the adjusted odds of death within 30 days associated with coated-platelet levels. Given the small number of deaths among patients in the highest tertile group (n=2), 2 groups defined by the median coated-platelet level were considered, with coated-platelet groups defined relative to a cut-point of <27% versus ≥27%, representing the median value for coated-platelet levels in our entire SICH population. Among the 47 patients with coated-platelets <27%, 18 (38%) died within 30 days compared with only 4 (8%) of the 48 with coated-platelets ≥27%.
Based on an unadjusted model, the odds of death within 30 days for patients with coated-platelet levels below the median (<27%) were 6.83× the odds for patients with higher levels of coated-platelets (≥27%; OR, 6.83; 95% confidence interval, 2.10–22.23; P=0.0014). The confidence intervals are wide because of the relatively small number of patients who died within 30 days (n=22). Regardless of the other factors that were considered in the regression model, the association between low coated-platelet levels and increased odds of early death remained. The only variable that met our definition of a potential confounder,22 resulting in at least a 10% change in the estimated OR, was time from onset. After adjustment for time from onset, the odds of death within 30 days for patients with coated-platelet levels below the median (<27%) were 6.00× the odds for patients with high levels of coated-platelets (≥27%); OR, 6.00; 95% confidence interval, 1.80–19.93; P=0.0035). Thus, although the estimated OR was reduced by >10%, the association between lower coated-platelet levels and increased odds of death within 30 days remained high and statistically significant.
In this first prospective study of coated-platelets in SICH, we found that lower coated-platelet potential, measured shortly after the occurrence of the intracerebral bleed, is associated with increased mortality at 30 days (Figure 2A). More specifically, a 6-fold increase in the odds of death within 30 days was noted in patients with coated-platelet levels <27% when compared with those with levels ≥27%, even after adjusting for potentially confounding variables.
These results are consistent with previous research from our group showing an association between lower coated-platelet levels and the presence of SICH and an inverse linear correlation between coated-platelet levels and bleed volume in patients afflicted by this condition.8,9 Lower levels of coated-platelets were also noted in patients with nonlacunar stroke with early hemorrhagic transformation as compared with those without such a complication.21 In addition, a link between lower coated-platelet levels and increased mortality at 30 days was noted in patients with subarachnoid hemorrhage.24
The current findings also support previously published data from human and animal studies showing impaired coated-platelet generation under hemorrhagic conditions. Brooks et al25 have shown that the presence of a genetically determined deficiency of coated-platelet production in German shepherd dogs results in a severe bleeding diathesis, whereas Daskalakis et al26 recently demonstrated impaired coated-platelet generation in a significant proportion of individuals with platelet-based bleeding diathesis of unknown cause. Furthermore, Saxena et al27 have also shown that coated-platelets seem to affect the bleeding phenotype in severe hemophilia. Interestingly, the levels noted by Daskalakis et al26 in patients with impaired coated-platelet synthesis (<20%) are similar to the levels noted for patients with SICH in the lowest tertile of coated-platelet levels (<21%) and with the highest mortality (Figure 2A).
In our study, we also found that patients with SICH in the lowest tertile of coated-platelet levels (<21%) had significantly larger bleed volumes and were more likely to experience hematoma expansion than those in the middle and lowest tertiles (Tables 3 and 4; Figure 2B). These patients also had a modestly significant increase in white-blood cell counts. The finding of larger bleed volumes was expected, given the previously reported inverse linear correlation between coated-platelets and bleed size in SICH.9 We suspect that the presence of 8 of 9 cases (89%) of hematoma expansion, a major and potentially preventable complication in SICH,10,28 in patients in the lowest tertile of coated-platelets is also linked to the correlation between lower coated-platelet levels and larger bleed volume.9
The main characteristic of coated-platelets centers on the presence of bound procoagulant proteins on their surface.1,2 The presence of these procoagulant proteins is responsible for the prothrombinase activity associated with coated-platelets but not with noncoated-platelets.1 Thus, a deficiency in coated-platelet generation may be one of the elements that negatively affects the hemostatic process when responding to a vascular injury.29 Therefore, the mechanism underlying our main findings likely reflects the consequences of inadequate levels of the prothrombotic coated-platelet population in the setting of a hemorrhagic insult, such as SICH. Although support for this hypothesis is provided by both animal and human studies,25,26 recent data from Coluci et al30 may also open the possibility of targeted interventions aimed at alleviating this situation by increasing platelet procoagulant potential. These authors have shown for the first time that in vivo use of 1-deamino-8-d-arginine vasopressin (desmopressin) in patients with mild platelet-based bleeding diatheses markedly enhanced the ability to form coated-platelets.30 As a result, future studies aimed at examining the potential role of desmopressin in patients with SICH identified by the lowest coated-platelet potential and the highest mortality risk may be warranted.
Limitations of this study include a relatively small number of deaths in our cohort, a larger percentage of men than women, the exclusion of patients receiving platelet transfusion, and the potential exclusion of patients with severe hemorrhage who died before a diagnosis of SICH was established. Nevertheless, these results demonstrate a strong association between lower coated-platelet levels and 30-day mortality in patients with SICH, setting the stage for additional studies examining possible interventions aimed at increasing platelet procoagulant potential and evaluating the effect of these interventions in patients most likely to benefit from novel treatment options.
We thank Leslie Guthery and Paul Friese for help with patient enrollment, data entry, and coated-platelet assays.
Sources of Funding
This work was supported by a grant from the Department of Veterans Affairs Clinical Science Research & Development Service (Award number, 1I01CX000340).
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.009068/-/DC1.
- Received February 9, 2015.
- Revision received April 16, 2015.
- Accepted April 27, 2015.
- © 2015 American Heart Association, Inc.
- Remenyi G,
- Szasz R,
- Friese P,
- Dale GL.
- Alberio L,
- Safa O,
- Clemetson KJ,
- Esmon CT,
- Dale GL.
- Prodan CI,
- Vincent AS,
- Padmanabhan R,
- Dale GL.
- Prodan CI,
- Vincent AS,
- Dale GL.
- Morgenstern LB,
- Hemphill JC 3rd,
- Anderson C,
- Becker K,
- Broderick JP,
- Connolly ES Jr,
- et al
- Charania R,
- Smith J,
- Vesely SK,
- Dale GL,
- Holter J.
- Kothari RU,
- Brott T,
- Broderick JP,
- Barsan WG,
- Sauerbeck LR,
- Zuccarello M,
- et al
- Maldonado G,
- Greenland S.
- 23.↵PASS 2014. NCSS, LLC [computer program]. Kaysville, Utah. 2014.
- Prodan CI,
- Vincent AS,
- Kirkpatrick AC,
- Hoover SL,
- Dale GL.
- Pecci A,
- Balduini CL.
- Colucci G,
- Stutz M,
- Rochat S,
- Conte T,
- Pavicic M,
- Reusser M,
- et al