Thromboelastography-Based Evaluation of Gender-Associated Hypercoagulability

Abstract

Background

Age, race, and gender differences in coagulation status of healthy volunteers have been reported in previous case series; however, rigorous multivariate analysis adjusting for these factors is lacking. We aimed to investigate the effects of age, race, and gender on baseline coagulation status in healthy volunteers.

Methods

Thirty healthy volunteer controls with no history of bleeding or thrombotic events and no previous anticoagulant or antiplatelet use were recruited. Citrated and heparinized blood samples were drawn, and kaolin and platelet-mapping thromboelastography (TEG) assays performed.

Results

Thirty participants had a mean age of 37, mean body mass index of 29 kg/m², and were 47% African-American and 70% female. Women were significantly older than men (40 ± 11 y vs 28 ± 7 y, P = .002); there were no significant differences in demographics by race. Multivariate analysis of variance for the effect of age, race, and gender across TEG parameters yielded evidence for gender differences in hypercoagulability (Pillai’s trace P = .02), which appear to be driven by differences in K-time, alpha angle, maximal amplitude, and G parameter. Women were hypercoagulable compared to men, as manifested by shorter K-time, steeper alpha angle, higher maximal amplitude, and larger G parameter.

Discussion

Women at baseline have relatively hypercoagulable fibrin deposition kinetics, platelet contributions to clot formation, and overall clot strength compared to men, even when adjusted for age and race. Additional research is needed to specifically detail the key patient-level factors, clinical implications, and opportunities for tailored therapy related to gender-associated hypercoagulability.

Introduction

Males and females are known to have variations in life expectancy, comorbidities, and manifestations of acute and chronic diseases.¹ Recent studies have also found that female sex may confer a protective advantage in certain clinical settings;^2-5 however, controversy exists regarding the effects of sexual dimorphism on outcomes after traumatic injury.^8-15 Some studies have shown improved survival among females,^6,9-15 while others have reported no benefit.^7,8,16 While the effect on survival after traumatic injury remains unclear, the effect of sex on the coagulation profile has been clearly demonstrated. Several studies have reported that females have a more hypercoagulable profile and present as less coagulopathic than males after severe injury.^17,18 Based on these findings the question has arisen whether this difference in coagulation is present at baseline, or represents a more exaggerated response to injury in female patients.

Investigators have found that females are more hypercoagulable than males at baseline based on analysis of thromboelastography (TEG) parameters^19-24; however, no previous study has performed multivariate analysis to determine if these differences are present when confounders such as age and race are accounted for, which have both been shown to affect coagulation.^24-26 Therefore, we sought to evaluate the simultaneous roles of gender, age, and race on TEG parameters in healthy volunteers. We hypothesized that healthy females have a hypercoagulable TEG profile compared to males at baseline, even when accounting for age and race.

Methods

Study Design and Population

After review and approval by the University of Mississippi Medical Center Institutional Review Board, thirty healthy volunteer men and women were recruited by an institutional intranet add and informational flyers, from 7/12/18 – 9/18/18. Volunteers were excluded if they screened positive for known pregnancy, known bleeding or clotting disorders, and/or current anticoagulant or antiplatelet use by phone interview. After informed consent was obtained, patients filled out a short health questionnaire collecting self-reported information on height, weight, age, gender, race, history of heart disease, diabetes, high blood pressure, smoking, and tobacco use, and current medications—specifically non-steroidal anti-inflammatory medications (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), and fish oil.

Thromboelastography

Citrate- (3.2%) and heparin- (68 U/4 mL) anticoagulated whole blood was drawn in parallel and processed within 2 hour. After incubation with gentle rocking at room temperature for 30 minute, samples were run on a TEG5000 Thromboelastograph Hemostasis Analyzer (Haemonetics Corp: Braintree, MA) according to manufacturer’s instructions within 30 minutes of venipuncture. Citrated kaolin-activated TEG parameters recorded included R time, coagulation time (K), alpha angle, maximal amplitude (MA), G value, and lysis percentage 30 minutes after MA (LY30). To obtain platelet-mapping TEG parameters, heparinized blood was activated with 10 μL of manufacturer-provided reptilase/factor XIIa/phospholipid activator solution to obtain MA_Fibrin, activator +1 mM arachidonic acid to obtain MA_AA, and activator +2 μM adenosine diphosphate (ADP) to obtain MA_ADP.

Statistical Analysis

No parameter had more than 5% missing data. Confidence intervals were constructed by bootstrapping the differences in sex means (mean for males subtracted from the mean for females) for TEG parameters and continuous demographic variables 50 000 times using the bias corrected and accelerated method;^27,28 confidence intervals for sex differences in proportions (proportion for males subtracted from the proportion for females) for binary demographic variables were calculated according to Newcombe’s method.²⁹ Finally, multivariate analysis of covariance was used to determine which of sex, race, and age were significant predictors of the 9 TEG parameters based on the value of the Pillai Trace test statistic. The data were analyzed using R version 3.6.2³⁰ with the boot³¹ package and using PROC FREQ in SAS version 9.4 (SAS Institute, Cary, NC).

Results

Twenty-one females and nine males were recruited for the study. Demographics of the healthy volunteers are shown in Table 1. Age was statistically different between female and male participants with females being an average of 11.75 years (y) older (99% CI, 1.99 to 19.57 y older). All other demographic variables were similar between the 2 groups. With regards to TEG parameters, females had statistically shorter K times, steeper alpha angles, higher maximal amplitude (MA), and higher G value when compared to males (Table 2). Specifically, the estimated differences were −.78 (99% CI, −2.36 to −.27) for K time, 7.82 (99% CI, 2.77 to 19.23) for alpha angle, 7.50 (99% CI, 3.31 to 13.41) for MA, and 2926 Kdynes/cm² (99% CI, 1322 to 4799 Kdynes/cm²) for G. In contrast, R time, % LY30, and platelet-mapping parameters were not statistically distinct between females and males, as all the confidence intervals include zero (Table 2). When multivariate analysis of covariance was performed, race and age were determined to be jointly insignificant (P = .74) for the aggregate TEG parameters, and thus sex alone among the 3 predictor variables was considered sufficient (P = .02) to explain differences among TEG parameters in the multivariate model (Table 3). Graphical representation of the differences in TEG parameters by gender is shown in Figure 1.

Table 1.

Demographics in Females Versus Males.

Variable	Female N = 21	Male N = 9	Difference of Means/Proportions (99% CI)
Age (years), mean (min, max)	40 (23, 60)	28.25 (22, 42)	11.75 (1.99, 19.57)
Race (% AA)	11 (52)	3 (33)	19% (−27, 5%)
Body mass index (kg/m2), mean (min, max)	29.75 (19.57, 50.02)	27.34 (24.37, 34.43)	2.41 (−2.14, 7.16)
Current NSAID use, N (%)	8 (38)	4 (44)	6% (−34, 47)
Current SSRI use, N (%)	3 (14)	2 (22)	8% (−25%, 51)
Current fish oil use, N (%)	0 (0)	1 (11)	11% (−15%, 54)
Current tobacco use, N (%)	0 (0)	1 (11)	11% (−15%, 54)

Abbreviations: AA, African American; CI, confidence interval; kg, kilogram; m, meter; max, maximum; min, minimum; N, number; NSAID, nonsteroidal anti-inflammatory; SSRI, selective serotonin reuptake inhibitor.

Biased corrected confidence intervals for sex differences in means for continuous variables were calculated from 50 000 bootstrap samples. Confidence intervals for sex differences in proportions were calculated according to Newcombe’s method.

Table 2.

Thromboelastography (TEG) Parameters in Females Versus Males.

TEG Parameters, Mean (min, max)	Female N = 21	Male N = 9	Difference of means (99% CI)
R time	6.66 (4.70, 8.90)	6.53 (5.80, 7.60)	.12 (−.72, .88)
K time	1.60 (1.30, 2.20)	2.39 (1.60, 4.60)	−.78 (−2.36, −.27)
Alpha angle	67.00 (59.60, 71.50)	59.18 (41.30, 67.20)	7.82 (2.77, 19.23)
Maximal amplitude (MA)	68.17 (62.80, 74.80)	60.67 (50.20, 66.40)	7.50 (3.31, 13.41)
G value	10 848 (8400, 14 800)	7922 (5000, 9900)	2926 (1322, 4799)
% LY-30	.95 (.00, 6.20)	.52 (.00, 1.30)	.43 (−.21, 1.74)
Fibrin MA	8.80 (2.10, 34.80)	4.20 (2.00, 15.50)	4.60 (−1.96, 10.81)
% AA inhibition	65.73 (18.15, 100.00)	67.50 (48.53, 98.34)	−1.77 (−23.32, 14.59)
% ADP inhibition	57.35 (15.91, 100.00)	63.61 (45.81, 98.76)	−6.26 (−29.80, 8.68)

Abbreviations: AA, arachidonic acid; ADP, adenosine diphosphate; CI, confidence interval; MA, maximal amplitude; max, maximum; min, minimum; N, number; TEG, thromboelastography.

Biased corrected confidence intervals for sex differences in means for TEG parameters were calculated from 50 000 bootstrap samples.

Table 3.

Multivariable Analysis of (1) Sex as a Single Predictor and of (2) Race and Age as Simultaneous Predictors of Aggregate TEG Parameters.

Covariate	Pillai trace	P-value
Sex	.650	.02
Race and Age	.575	.74

Figure 1.

TEG parameters in females versus males. Abbreviations: MA, maximal amplitude; AA, arachidonic acid; ADP, adenosine diphosphate.

Discussion

Numerous studies have investigated the presence of gender dimorphisms in trauma, sepsis, and critically ill patients. Female septic patients have been shown to have an improved prognosis compared to their male counterparts, and although they have a similar clinical course, one study showed mortality for women is 26% versus 70% in men.^4,9 These findings may be explained by studies showing that males have an increased susceptibility towards infectious complications and a slightly greater severity of organ dysfunction and longer organ recovery time than females.^7,14,32 While some studies in trauma have found a survival advantage for younger premenopausal women, which is when women are likely still under the influence of the hormones of their menstrual cycle,^9-12,14,15 others have found no marked survival advantage relative to gender after traumatic injury.^7,8 Further confounding the picture, other studies have shown that postmenopausal status in females confers a survival benefit rather than premenopausal status.¹¹ It is likely these differences are mediated by other mechanisms in addition to sex hormones and require further investigation.¹³

This study assessed the effect of gender on baseline coagulation status in healthy volunteers using TEG, a whole blood viscoelastic assay that provides a comprehensive analysis of hemostasis including specific measurements of clot initiation, propagation, strength, and fibrinolysis. TEG is considered more sensitive than routine assays in detecting hypercoagulability.^18,33 We observed significant differences in coagulation between males and females with women demonstrating a hypercoagulable TEG profile compared to men. Females tended towards shorter K-times, steeper alpha angles, higher maximal amplitudes, and larger G parameters than their male counterparts. These results indicate that at baseline, females have faster fibrin deposition kinetics, greater overall clot strength, and increased platelet contribution to clot formation than males. These findings are consistent with previous studies showing a gender dimorphism in baseline coagulation of healthy volunteers, noting that women were hypercoagulable compared to men based on K time, alpha angle, MA, and fibrin and thrombin MA, respectively. ^19,21,22 Additional work has demonstrated that hypercoagulability was even more pronounced in pregnant females, and was not impacted by oral contraceptive status.²³

Two previous studies also evaluated the effect of age on the coagulation profile of healthy volunteers: while Scarpelini et al did not defect an effect of age on coagulation, Roeloffzen et al found significant differences in TEG parameters between subjects <50 years of age and >50 years of age, as well as a trend toward increasing hypercoagulability with increasing age by decade.^21,24 One additional study of age effects on TEG parameters demonstrated that age was significantly correlated with R time, K time, alpha angle, and MA in healthy volunteer males, but that only R time was significantly correlated among different ages in females.²⁵ Race also has been shown to have an effect on platelet function and venous thromboembolism (VTE), with African Americans having the highest risk for VTE followed by Caucasians, Hispanics, and those of Asian descent.^26,34 There was no significant difference between the percentage of African American volunteers in each group in this study. Overall, our multivariate results indicate that female gender remains a significant predictor of hypercoagulability at baseline, irrespective of age and race.

The clinical implications of female-specific hypercoagulability remain unclear. Several studies have shown that the gender dimorphism in coagulation at baseline persists after traumatic injury.^17,18 This finding is thought to be protective and confer a survival advantage. In a prospective cohort study, Schreiber et al¹⁸ found that greater than 80% of women were hypercoagulable by day 1 and the mean time to onset of clotting by TEG was 1 minute less than in men during this time period. However, the prevalence of hypercoagulability on day 2 in women was significantly less than on day 1 and was equivalent to men for the remaining days of the study. This suggests that women have a greater propensity to develop a hypercoagulable state early after injury, which may explain the reported survival benefit in females, since it would result in reduced hemorrhage early after traumatic injury. The rapid resolution may also be significant because it might explain why studies have shown that men have higher odds of thromboembolic complications when age-matched with premenopausal females even after controlling for injury severity score and mechanism of injury.

Gender dimorphism in coagulation may also have important clinical implications in the setting of thromboembolic disease. Venous thromboembolism is a life-threatening complication of post-injury recovery. Kaufmann et al³⁵ demonstrated that 65% of trauma patients are hypercoagulable in the emergency department based on their TEG profile and untreated trauma patients undergoing routine venography have been found to have a 58% incidence of thromboembolic disease.³⁵ Responses to heparin and anti-platelet therapy have been shown to differ by gender. In a TEG based analysis of blood samples, Monte et al demonstrated female-specific resistance to heparin.¹⁹ Another study showed that young healthy females demonstrate reduced response to clopidogrel when compared to their male counterparts.³⁶ Additionally, research has shown that women have consistently more reactive platelets, to multiple agonists, compared to men, which may be related to differential response to stimuli based on sex, as Coleman et al have shown that estradiol treatment “feminizes” male platelet responses;³⁷ this effect has been noted to persist even after administration of aspirin.³⁸ These results indicate the need for further evaluation of gender-specific dosing strategies for deep venous thrombosis prophylaxis.

Additional research is also needed to fully elucidate the mechanistic drivers of female-specific hypercoagulability. Both sex hormones and intrinsic platelet function likely play a role in distinct gender coagulation profiles. Estrogen and progestins have been implicated in promoting coagulation and are higher in pre-menopausal females.²³ Experimental murine models have demonstrated that administration of estradiol leads abrogation of hemorrhagic shock.³⁸ Furthermore, sexual dimorphisms have been shown to exist in platelet function—specifically activation and aggregation.^37,38 Additional work is needed to elucidate the exact mechanisms of sexual dimorphism in coagulation and possible treatment strategies based on these mechanisms.

There are several limitations in this study. First, we did not specifically assess oral hormonal therapy use, menopausal status, previous hysterectomy and/or oophorectomy, or estrus cycle timing of female study participants. Future work could be enhanced by documentation of oral hormonal therapy, evaluation of menopausal status, and biochemical measurement of hormonal levels at the time of blood draw. Second, we had a larger number of female participants compared to male participants, likely due to the nature of the institutional advertising. Third, we did not measure hemoglobin or hematocrit in our samples; work by Roeloffzen et al²⁴ has shown a correlation between hemoglobin levels and speed to reach a certain clot strength (K time), even when levels fell within the “normal” range.²⁴

In summary, females demonstrate hypercoagulability in TEG K, alpha angle, MA, and G parameters, indicating that women at baseline have relatively hypercoagulable fibrin deposition kinetics, platelet contribution to clot formation, and overall clot strength when compared to men. Our study demonstrates that female gender remains predictive of hypercoagulability irrespective of age and race. Additional research is needed to fully elucidate the mechanisms of female-specific hypercoagulability and the clinical implications of this gender dimorphism in baseline coagulation status.

Key Takeaways.

• Gender-associated differences in coagulation have been observed, but are not well-documented using viscoelastic clot formation assays or when adjusted for age and race

• Women at baseline have hypercoagulable fibrin deposition kinetics, platelet contributions to clot formation, and overall clot strength compared to men, even when adjusted for age and race

• Gender-associated differences in coagulation may need to be considered along with other factors when devising optimal venous thromboembolism prophylaxis regimens