ONLINE RESOURCE

Thrombelastography Does Not Predict Clinical Response to rtPA for Acute Ischemic Stroke

1Mark M McDonald, 1Jeremy Wetzel, 1Stuart Fraser, 1Andrea Elliott, 1Ritvij Bowry, 1Jorge F Kawano-Castillo, 2Chunyan Cai, 1Navdeep Sangha, 1Jessica Messier, 1Amanda Hassler, 1Joancy Archeval-Lao, 1Stephanie A Parker, 2Mohammad H Rahbar, 3Evan G Pivalizza, 1,4 Tiffany R Chang, 5James C Grotta, 1Department of Neurology, University of Texas Health Science Center at Houston, 2Department of Internal Medicine, University of Texas Health Science Center at Houston, 3Department of Anesthesiology, University of Texas Health Science Center at Houston, 4Department of Neurosurgery, University of Texas Health Science Center at Houston, 5Memorial Hermann Hospital, Houston, TX

The University of Texas Health Science Center at Houston. 7000 Fannin St, Houston, TX 77030

Memorial Hermann Hospital. 6411 Fannin St, Houston, TX 77030

Corresponding Author: James C Grotta

Phone: (713) 704-4000 Fax: (713) 704-9096

Key words: thrombolysis, stroke, haemocoagulation, thromboelastography, thrombus

SUPPLEMENTAL METHODS

Blood Sampling and Processing

Approximately 10 mL of whole blood was collected into a citrated tube by MHHED nurses prior to tPA administration. The blood was kept at room temperature and taken for TEG analysis within 2 hours of sample collection. One milliliter of whole blood was placed into a kaolin vial and gently mixed by inverting 5 times. 340 uL of this solution was then pipetted into the disposable cup in the machine well. Next, the citrate was reversed by the addition of 20uL of 0.2M calcium chloride into the cup. A pipette was then used to gently mix the blood in the cup. The test was run on a computerized TEG coagulation analyzer (Haemonetic Corp, Model 5000). The TEG analyzer was tested for quality assurance through daily quality control using normal and abnormal controls for operational and calibration verification. All study personnel performing the testing and quality controls were trained in the procedures.

TEG Parameters

The following TEG values were documented at the completion of the test: SP (minutes), R (minutes), K (minutes), Delta (minutes), Angle (degrees), MA (mm), G (dynes/cm2), and LY30 (percentage). Split point (SP) is the time from the start of sample analysis to the beginning of detectable fibrin formation. Reaction time (R) is the time from the start of the test until clot stiffness reaches an amplitude of 2 mm. Kinetics (K) is the time elapsed to achieve a clot firmness amplitude of 20 mm. Delta is the time from SP until R and reflects the time to reach maximum speed of clot formation. Angle (alpha) is a measure of the angle between the slope of the line starting at R running tangent to the TEG tracing with the horizontal axis. Angle reflects the rate of clot strengthening. Maximum amplitude (MA) is the maximum width of the TEG tracing and indicates maximum clot strength. G measures clot firmness and is derived from MA {G = (5,000*MA)/ (100-MA)}. LY30 measures the percentage decrease in MA over a 30 minute interval and is a measure of clot lysis.

Statistical Analysis

Continuous variables with normal distributions were summarized by mean and standard deviation while variables with skewed distributions were summarized by median and interquartile range. Categorical variables were summarized by frequency and percentage. Differences in baseline demographics, labs, and clinical data among outcomes groups {1. RCI (+) vs. (-); 2. HDMCA (+) vs. (-); 3. Hemorrhagic transformation (+) vs (-)} were compared using two sample t-test (or Wilcoxon rank sum test as appropriate) for continuous variables and Chi-square test (or Fisher’s exact test as appropriate) for categorical variables. To compare baseline TEG values among different groups, we first conducted simultaneous comparisons of all components of TEG values based on multivariate analysis (Hotelling-Lawley Trace test). Subsequently, we conducted post-hoc analysis for each individual TEG component. Multivariable linear regression models were conducted to compare the differences of baseline TEG components among groups after adjusting for potential confounding effects. For situations that linear regression was not appropriate, we used other suitable regression models. For example, since the distribution of post LY30 was highly skewed, we used univariable and multivariable quantile regression models to compare the median of post LY30 between different groups. In multivariable quantile regression models we controlled for the effect of confounders. The identification of confounders was based on both a priori and empirical considerations. First, variables shown previously to be correlated with TEG components (e.g., age, smoking status, platelet count, and antithrombotic medications) were included in the analysis. Second, through univariable analyses using p-value < 0.20, we identified the factors which both differed among groups and were associated with TEG components. Lastly, the covariates were considered to be confounders if the regression coefficient of the group indicator varies by >20% when the covariate is added to or deleted from the multivariable model.

SUPPLEMENTAL TABLES

Table 1. Comparison of demographics and baseline lab variables between patients with and patients without RCI

VARIABLES / RCI No
(N=102) / RCI Yes
(N=32)
Age (years), mean±SD / 67.4±14.9 / 63.5±15.1
Male, n(%) / 66 (64.7) / 11 (34.4)
Race
African American
Caucasian
Hispanic
Other / 24 (23.5)
47 (46.1)
24 (23.5)
7 (6.9) / 12 (37.5)
11 (34.4)
7 (21.9)
2 (6.3)
Hypertension, n(%) / 82 (80.4) / 25 (78.1)
Hyperlipidemia, n(%) / 33 (32.4) / 7 (21.9)
Diabetes Mellitus, n(%) / 44 (43.1) / 4 (12.5)
Coronary artery disease, n(%) / 15 (14.7) / 3 (9.4)
Smoking, n(%) / 19 (18.8) * / 7 (21.9)
Aspirin, n(%) / 42 (41.2) / 10 (31.3)
Clopidogrel, n(%) / 14 (13.7) / 4 (12.5)
Warfarin, n(%) / 7 (6.9) / 1 (3.1)
Glucose (mg/L), median (Q1, Q3) / 123.5 (108, 170) / 109.5 (100.5, 135.5)
Hemoglobin (mEq/L), mean±SD / 13.8±1.8 / 13.2±1.6
Platelet count (mEq/L), mean±SD / 210.6±63.2 / 239.4±82.3
PT, mean±SD / 13.5±1.2* / 13.4±1.0
Partial thromboplastin time, median (Q1, Q3) / 29.2 (26.9, 31.6) * / 30.7 (27.8, 32.7) †
International normalized ratio, mean±SD / 1.0±0.2‡ / 1.0±0.1§
National institutes of health stroke scale score, median (Q1, Q3) / 8 (6, 14) / 6 (4, 11.5)
Time of 36 hour NIHSS measurement from baseline blood draw (hours), mean±SD / 29.7±9.4‡ / 32.3±10.4
Baseline systolic blood pressure (mm Hg), mean±SD / 172.6±33.4 / 164.7±35.1
Baseline diastolic blood pressure(mm Hg), mean±SD / 90.2±20.9 / 88.3±23.1

*N=101; †N=31; ‡N=91; §N=27; Abbreviations: SD, standard deviation; Q1, 1st quartile; Q3, 3rd quartile

Table 2: Comparison of demographics and baseline lab variables between patients with and patients without hemorrhagic transformation

VARIABLES / Hemorrhagic Transformation No
(N=123) / Hemorrhagic Transformation Yes
(N=17)
Age (years), mean±SD / 65.4±15.0 / 72.6±11.2
Male, n(%) / 69 (56.1) / 10 (58.8)
Race
African American
Caucasian
Hispanic
Other / 37 (30.1)
49 (39.8)
29 (23.6)
8 (6.5) / 3 (17.7)
9 (52.9)
3 (17.7)
2 (11.8)
Hypertension, n(%) / 96 (78.1) / 15 (88.2)
Hyperlipidemia, n(%) / 37 (30.1) / 5 (29.4)
Diabetes Mellitus, n(%) / 46 (37.4) / 5 (29.4)
Coronary artery disease, n(%) / 15 (12.2) / 3 (17.7)
Smoking, n(%) / 24 (19.5) / 2 (11.8)
Aspirin, n(%) / 45 (36.6) / 8 (47.1)
Clopidogrel, n(%) / 16 (13.0) / 3 (17.7)
Warfarin, n(%) / 5 (4.1) / 3 (17.7)
Glucose (mg/L), median (Q1, Q3) / 119 (106, 158) / 135 (107, 169)
Hemoglobin (mEq/L), mean±SD / 13.6±1.8 / 13.8±1.8
Platelet count (mEq/L), mean±SD / 217.3±71.2 / 213.4±45.0
PT, mean±SD / 13.4±1.0* / 13.8±2.0
Partial thromboplastin time, median (Q1, Q3) / 29.2 (26.9, 32.0) † / 28.2 (26.2, 31.5)
International normalized ratio, mean±SD / 1.0±0.1* / 1.1±0.3
National institutes of health stroke scale score, median (Q1, Q3) / 7 (4, 12) / 13 (8, 19)
Time of 36 hour NIHSS measurement from baseline blood draw (hours), mean±SD / 30.4±9.6‡ / 30.7±10.1
Baseline systolic blood pressure (mm Hg), mean±SD / 169.3±34.5 / 177.8±29.4
Baseline diastolic blood pressure(mm Hg), mean±SD / 90.0±22.6 / 91.1±19.1

* N=122; † N=121; ‡ N=107; Abbreviations: SD, standard deviation; Q1, 1st quartile; Q3, 3rd quartile

Table 3: Comparison of demographics and baseline lab variables between patients with and patients without HDMCA

VARIABLES / HDMCA No
(N=111) / HDMCA Yes
(N=30)
Age (years), mean±SD / 66.0±14.4 / 67.9±16.3
Male, n(%) / 64 (57.7) / 16 (53.3)
Race
African American
Caucasian
Hispanic
Other / 36 (32.4)
45 (40.5)
22 (19.8)
8 (7.2) / 4 (13.3)
14 (46.7)
10 (33.3)
2 (6.7)
Hypertension, n(%) / 90 (81.1) / 22 (73.3)
Hyperlipidemia, n(%) / 31 (27.9) / 12 (40.0)
Diabetes Mellitus, n(%) / 43 (38.7) / 8 (26.7)
Coronary artery disease, n(%) / 12 (10.8) / 7 (23.3)
Smoking, n(%) / 22 (20.0) * / 4 (13.3)
Aspirin, n(%) / 44 (39.6) / 10 (33.3)
Clopidogrel, n(%) / 14 (12.6) / 5 (16.7)
Warfarin, n(%) / 6 (5.4) / 2 (6.7)
Glucose (mg/L), median (Q1, Q3) / 119 (105, 169) / 122 (109, 145)
Hemoglobin (mEq/L), mean±SD / 13.7±1.7 / 13.7±1.9
Platelet count (mEq/L), mean±SD / 217.6±73.0 / 210.2±51.0
PT, mean±SD / 13.4±0.9 / 13.8±1.6†
Partial thromboplastin time, median (Q1, Q3) / 28.9 (26.8, 32.0) * / 30.3 (26.9, 32.9) †
International normalized ratio, mean±SD / 1.0±0.1 / 1.1±0.2†
National institutes of health stroke scale score, median (Q1, Q3) / 7 (4, 11) / 12.5 (6, 17)
Time of 36 hour NIHSS measurement from baseline blood draw (hours), mean±SD / 30.1±9.6‡ / 31.7±9.6§
Baseline systolic blood pressure (mm Hg), mean±SD / 173.5±34.1 / 158.7±30.6
Baseline diastolic blood pressure(mm Hg), mean±SD / 89.8±21.0 / 91.3±25.9

*N=110; †N=29; ‡N=100;§N=24; Abbreviations: SD, standard deviation; Q1, 1st quartile; Q3, 3rd quartile;