Fibrin Clot Strength in Patients with Diabetes Mellitus Measured by Thrombelastography

Background Patients with diabetes mellitus (DM) exhibit increased risk of recurrent myocardial infarction. Maximal clot strength measured by thrombelastography (TEG) is a risk factor for recurrent ischemic events. We hypothesized that diabetic subjects exhibit increased fibrin clot strength in platelet-poor plasma and that glycemic control correlates with maximal fibrin clot strength. Methods We collected plasma samples from subjects with known or suspected coronary artery disease undergoing cardiac catheterization (n = 354). We measured kaolin-activated TEG in platelet-poor citrate plasma. Time to fibrin formation (R), clot formation time (K), and maximal fibrin clot strength (MA) were recorded. Results Plasma fibrin MA was increased among subjects with DM (n = 152) as compared to non-DM (n = 202) (37.0 ± 8 versus 34.1 ± 8 mm; p < 0.001). Hemoglobin A1c (HbA1c) (ρ = 0.22; p = 0.001) and fibrinogen (ρ = 0.29; p < 0.001) correlated with fibrin MA. In multivariable regression analysis, DM remained significantly associated with plasma MA after adjustment for fibrinogen level (p = 0.003). Conclusions Subjects with diabetes mellitus exhibit increased maximal fibrin clot strength measured by TEG in platelet-poor plasma.

Thrombelastography (TEG) is an established clinical laboratory method for analysis of coagulation and fibrinolysis and allows for quantitative measurement of the dynamic processes of clot formation and clot strength [12,13]. We have previously demonstrated the feasibility of measurement of plasma fibrin clot strength by TEG in platelet-poor plasma and correlation with whole blood fibrin TEG measurements [14]. Elevated whole blood as well as platelet-poor plasma maximal clot strength by TEG (TEG-MA) has been associated with an increased risk of ischemic events after coronary stenting [15][16][17]. Subjects with high plasma clot strength showed a 3.8-fold increased risk for cardiovascular death or myocardial infarction after percutaneous coronary intervention [17].
We hypothesized that subjects with DM exhibit increased fibrin clot strength (TEG-MA) in platelet-poor plasma as compared with controls and that extent of glycemic control correlates with maximal fibrin clot strength.

Materials and Methods
2.1. Patients. The study protocol was approved by the Indiana University Institutional Review Board for Research. All enrolled individuals gave written informed consent. Subjects with known or suspected coronary artery disease who were referred for cardiac catheterization were included in this study. Blood samples were obtained prior to cardiac catheterization or hospital discharge. Plasma samples were stored as part of a biobank registry in patients with cardiovascular disease (Genetics and Cardiac Catheterization (GENCATH) study).

Blood
Samples. Whole blood was collected by peripheral venipuncture in 3.2% citrate after discarding of initial blood draw. We obtained platelet-poor plasma (PPP) by centrifugation at 2000 ×g for 15 min. Plasma was stored at −80°C and thawed immediately before analysis.

Subjects with Diabetes
Mellitus. Hemoglobin A1c (HbA1c) was recorded if there was a value present during chart review within three months of the blood sample collection. Within the insulin-dependent subpopulation of the diabetic group, the total daily administered insulin dose was recorded and this was defined as the sum of shortacting plus long-acting insulin as obtained from the patient's medication list.

Thrombelastography (TEG).
Kaolin-activated TEG (Haemonetics, USA) was performed in all subjects (n = 354) using platelet-poor citrate plasma samples according to the manufacturer's instructions. Citrate plasma was mixed with kaolin and then loaded in a heparinase-coated cup containing 20 ml of CaCl2 [18]. Thrombelastography was stopped after time to fibrin formation (R), clot formation time (K), and maximal clot strength (MA) were recorded. In a subset of subjects (n = 74), kaolin-activated TEG was also performed in the same fashion using whole blood citrate samples.
2.6. Statistical Analysis. We conducted statistical analysis with SPSS 24 (IBM, USA). We defined statistical significance as a p value less than 0.05, with all tests being conducted two sided. All the values are mean ± SD. We used unpaired, two-sided Student's t-test to compare normally distributed continuous data between two groups after assessing normal distribution of continuous data by the Kolmogorov-Smirnov test. We used the χ 2 test to compare categorical variables. To calculate correlation between thrombelastography parameters (MA, R, and K) and other continuous variables, Spearman's rho was used. Multivariable linear regression analysis was performed with TEG-MA as the dependent variable and diabetes mellitus as the covariate with forward stepwise adjustment for remaining clinical variables and fibrinogen level.

Discussion
The results of our study demonstrate that subjects with diabetes mellitus have increased fibrin clot strength (MA) as measured by TEG in platelet-poor plasma compared to  nondiabetics. The association of diabetes mellitus with plasma fibrin clot strength persisted after adjustment for fibrinogen, which influences strength of clot formation measured by TEG. We have recently demonstrated that elevated fibrin clot strength measured by TEG in platelet-poor plasma is associated with increased risk of recurrent myocardial infarction and death after PCI, similar to the findings observed in studies using whole blood TEG assays [17].
In whole blood, maximal clot strength (MA) measured by TEG is the sum of the additive forces generated by aggregating and contracting platelets incorporated into the growing, initially soluble fibrin network, which is cross-linked and stabilized by FXIIIa [13,20]. In a large cohort study of subjects with coronary artery disease by Lev et al., DM was also independently associated with whole blood (platelet-fibrin) clot strength measured by TEG [21].
In the subgroup of patients in our study who had both whole blood and plasma TEG studies performed, there was a moderate correlation between both measurements for TEG-MA and to a lesser degree for TEG-R and TEG-K. While plasma MA was significantly higher in diabetic subjects, we did not find a significant difference in whole blood MA in our study. This may be due to the smaller number of subjects with whole blood TEG assays completed, as well as possible differences in platelet contribution to clot strength in whole blood TEG between subjects with and without diabetes.
Fibrinogen is an important determinant of fibrin structure and of cross-linked fibrin network strength, and both fibrinogen and factor XIIIa activity influence maximal fibrin clot strength measured by TEG [22]. Fibrinogen has been previously shown to be elevated in subjects with diabetes mellitus [11,23]. In our study, fibrinogen levels were also higher in subjects with DM, and fibrinogen remained a significant covariate in multivariable linear regression analysis. In contrast, PAI-1 and factor XIIIa were not significantly different between diabetics and nondiabetics.
We did not find any significant correlation between daily insulin dose and plasma TEG-MA among the insulindependent subpopulation, while elevated plasma MA was associated with worse glycemic control. Possible mechanisms linking hyperglycemia to a prothrombotic state include effects on gene transcription of coagulation factors, loss of the endothelial layer through glycosylation with a deficiency of nitric oxide, direct glycosylation of coagulation factors altering their activity [24,25], as well as increased platelet activation [4,[26][27][28][29]. Combined hyperinsulinemia and hyperglycemia has been demonstrated to acutely increase circulating tissue factor procoagulant activity (TF-PCA) ninefold in previously healthy subjects [30].
Current recommendations in primary prevention of cardiovascular disease among patients with DM do not yet incorporate use of therapies targeted towards reduction of thrombin formation and fibrin stabilization. However, recent results from the COMPASS study demonstrate a significant reduction in cardiovascular death/myocardial infarction or stroke in patients with stable cardiovascular disease who were treated with rivaroxaban 2.5 mg twice daily in addition to aspirin as compared to patients who were treated with aspirin alone (HR 0.76; 95% CI: 0.66-0.86) [31]. The reduction in risk with rivaroxaban was more pronounced in the diabetic subgroup (HR 0.74; 95% CI: 0.61-0.9) [31]. Future studies will need to assess whether TEG parameters in diabetic patients could be used to risk stratify and develop treatment strategies in diabetic patients with known or suspected cardiovascular disease.
Limitations to our study include that HbA1C values were collected retrospectively and were not standardized to a collection at the time of TEG blood sample collection; as values were only recorded if collected within a three-month window, glycemic control at time of analysis was assumed to be equivalent. We were not able to assess thrombolysis characteristics of patients with DM, as spontaneous thrombolysis generally does not occur with TEG after activation with kaolin and CaCl2.
In conclusion, our data show that diabetes mellitus is associated with increased fibrin clot strength (TEG-MA) as measured in platelet-poor plasma, a marker of thrombotic risk in patients with coronary artery disease. Worse glycemic control as defined by an elevated HbA1C is moderately correlated with higher plasma TEG-MA, whereas there was no statistically significant correlation with use or dosage of insulin. Further studies are warranted to examine the mechanisms associated with increased plasma TEG-MA and to examine the utility of plasma TEG-MA in personalized approaches to guide antithrombotic therapies in patients with diabetes mellitus.

Disclosure
The contents of this manuscript have in part been presented as a poster at the Arteriosclerosis, Thrombosis, and Vascular Biology (ATVB) scientific meeting in 2015, and the corresponding abstract has been published in Arteriosclerosis, Thrombosis, and Vascular Biology, May 2015, Volume 35, Issue Suppl 1: A597.

Conflicts of Interest
Rolf P. Kreutz has served as scientific consultant for Roche Diagnostics, Coramed, and Haemonetics Corporation.

Acknowledgments
The study was supported in part by the Indiana University School of Medicine-IU Health Strategic Research Initiative, the IU Health Values Grant, and the CTSI Specimen Storage Facility (SSF), subsidized by the Indiana Clinical and Translational Sciences Institute which is supported,