Hemostatic responses to exercise, dehydration, and simulated bleeding in heat-stressed humans

Abstract

Heat stress followed by an accompanying hemorrhagic challenge may influence hemostasis. We tested the hypothesis that hemostatic responses would be increased by passive heat stress, as well as exercise-induced heat stress, each with accompanying central hypovolemia to simulate a hemorrhagic insult. In aim 1, subjects were exposed to passive heating or normothermic time control, each followed by progressive lower-body negative pressure (LBNP) to presyncope. In aim 2 subjects exercised in hyperthermic environmental conditions, with and without accompanying dehydration, each also followed by progressive LBNP to presyncope. At baseline, pre-LBNP, and post-LBNP (<1, 30, and 60 min), hemostatic activity of venous blood was evaluated by plasma markers of hemostasis and thrombelastography. For aim 1, both hyperthermic and normothermic LBNP (H-LBNP and N-LBNP, respectively) resulted in higher levels of factor V, factor VIII, and von Willebrand factor antigen compared with the time control trial (all P < 0.05), but these responses were temperature independent. Hyperthermia increased fibrinolysis [clot lysis 30 min after the maximal amplitude reflecting clot strength (LY₃₀)] to 5.1% post-LBNP compared with 1.5% (time control) and 2.7% in N-LBNP (P = 0.05 for main effect). Hyperthermia also potentiated increased platelet counts post-LBNP as follows: 274 K/µl for H-LBNP, 246 K/µl for N-LBNP, and 196 K/µl for time control (P < 0.05 for the interaction). For aim 2, hydration status associated with exercise in the heat did not affect the hemostatic activity, but fibrinolysis (LY₃₀) was increased to 6–10% when subjects were dehydrated compared with an increase to 2–4% when hydrated (P = 0.05 for treatment). Central hypovolemia via LBNP is a primary driver of hemostasis compared with hyperthermia and dehydration effects. However, hyperthermia does induce significant thrombocytosis and by itself causes an increase in clot lysis. Dehydration associated with exercise-induced heat stress increases clot lysis but does not affect exercise-activated or subsequent hypovolemia-activated hemostasis in hyperthermic humans. Clinical implications of these findings are that quickly restoring a hemorrhaging hypovolemic trauma patient with cold noncoagulant fluids (crystalloids) can have serious deleterious effects on the body’s innate ability to form essential clots, and several factors can increase clot lysis, which should therefore be closely monitored.

INTRODUCTION

Individuals who experience trauma and bleeding are often subject to a variety of environmental conditions that may further alter their innate coagulation and fibrinolytic systems. For example, soldiers in a desert environment, police officers, or even construction workers may suffer a traumatic injury while working under hot environmental conditions. Furthermore, these individuals are predisposed to varying degrees and combinations of hyperthermia, dehydration, hypovolemia, and physical exertion that induce a catecholamine response, all of which by themselves may affect hemostatic function. Consistent with such scenarios, central hypovolemia activates hemostasis whether it is provoked by controlled bleeding (34, 45), simulated bleeding by lower-body negative pressure (LBNP) (4, 43, 44), or orthostatic stress (4). However, the interactive effects of heat stress combined with a hemorrhagic insult on hemostatic activity are less clear.

Meyer et al. investigated the combined effect of heat stress and central hypovolemia on markers of hemostatic and fibrinolytic activities (28). Although subjects were heat stressed during resting conditions (i.e., passively elevated skin and core temperatures), they were subsequently exposed to a simulated hemorrhagic challenge via progressive LBNP. It was observed that these combined conditions resulted in a hypercoagulable state. However, a limitation of their study design was the absence of a normothermic comparison group, and thus it was unclear whether heat stress contributed to that hypercoagulable state. Additionally, the effects of exercise-induced hyperthermia on hemostatic activity during a subsequent hemorrhagic insult are unknown. This latter question is particularly relevant in military settings where soldiers may be performing exercise (e.g., foot patrols) while in hyperthermic environmental conditions before suffering a hemorrhage.

The purpose of this study was to evaluate the combined effects of hyperthermia, hydration status during exercise, and subsequent central hypovolemia on the coagulation and fibrinolytic systems. LBNP was used to induce central hypovolemia to simulate a hemorrhagic insult. In aim 1, hemostasis was evaluated pre- to post-LBNP-induced hypovolemia while subjects were normothermic and again while hyperthermic. In aim 2, subjects exercised in hyperthermic environmental conditions while hydrated, while dehydrated, and, finally, while dehydrated until their core body temperature increased to the same level as the hydrated trial; each of the three conditions was then followed by the LBNP-induced hypovolemia protocol. In aim 1, we hypothesized that, compared with the normothermic trial, hyperthermia would potentiate the hypercoagulable response to LBNP-induced hypovolemia. In aim 2, we hypothesized that dehydration would further potentiate the hypercoagulable response compared with when subjects were hydrated and then exposed to the LBNP challenge.

METHODS

Subjects

The study protocols and informed consents were reviewed and approved by the Institutional Review Boards of the University of Texas Southwestern Medical Center and Texas Health Presbyterian Hospital of Dallas, and the study was conducted in accordance with the Declaration of Helsinki. The data presented herein were obtained simultaneously with the data presented in the published manuscripts of Schlader et al. (36) and Gagnon et al. (10), each addressing hypotheses unique to those addressed in the present work.

For both aims, male subjects from the Dallas-Fort Worth area of Texas were recruited and provided written informed consent before study-related activities were performed. Females were not included given the small sample size and potential for known coagulation differences based on circulating sex hormones (12). In the 12 subjects addressing aim 1, their mean (SD) characteristics were as follows: 33 (4) yr (range 24–41), height 183 (8) cm (range 170–192), and weight 85 (12) kg (range 69–108). Eleven subjects participated in aim 2, having the following characteristics: 32 (7) yr (range 26–44), height 183 (10) cm (range 170–204), and weight 83 (11) kg (range 72–100). Before the study, subjects received both written and verbal information outlining all procedures and risks associated with the experiment. Subjects underwent medical screening, including their medical history. The subjects were normotensive, nonsmoking, and had no known cardiovascular, metabolic, neurological, or hemostatic diseases. To reduce the risk of atypical physiological responses, subjects were instructed to maintain their sleeping habits and, for 24 h before the study, to refrain from alcohol, intensive exercise, and autonomic stimulants such as prescription (e.g., antihistamines and decongestants) or nonprescription (e.g., caffeine) drugs.

Instrumentation

Core body temperature was measured from intestinal temperature via a telemetric temperature pill (HQ, Palmetto, FL) that was swallowed by each subject a minimum of ~2 h before data collection. For aim 1, subjects were dressed in a two-piece tube-lined suit (Med-Eng, Ottawa, ON, Canada) that permitted the control of skin temperature by changing the temperature of the water perfusing the suit. To induce heat stress, the tube-lined suit was perfused with 49°C water until core body temperature increased by ~1.2°C, at which point the temperature of the water perfusing the suit was slightly reduced to attenuate further increases in core body temperature. For aim 2, subjects were dressed only in shorts and running shoes.

Each subject was instrumented with a 21-gauge catheter in an antecubital vein for blood sampling and with noninvasive devices [electrocardiogram, and finger photoplethysmography (Finometer; Finapres Medical Systems, Amsterdam, The Netherlands)] for continuous cardiovascular monitoring (heart rate, arterial blood pressures, and stroke volume) used to identify LBNP termination. After instrumentation, subjects rested supine for a minimum of 30 min before baseline measurements were obtained.

Aim 1: passive hyperthermia.

The experimental model for both aims is depicted in Fig. 1. Before the aforementioned rest period, subjects were positioned in the LBNP chamber. This aim required three unique laboratory visits, with each visit separated by ~60 days. The visits are as follows: 1) subjects remained normothermic throughout the trial and were not exposed to LBNP (time control). Rather, subjects remained supine for a similar duration relative to the other trials in this aim; 2) subjects remained normothermic during this session followed by LBNP until presyncope (N-LBNP); and 3) subjects were passively heated and then exposed to LBNP to presyncope (H-LBNP). For all three visits, data were obtained for 60 min following the end of LBNP, or a similar period of time for the time control trial. For the heat stress trials, skin cooling began shortly (2–3 min) after termination of LBNP by perfusing cool water through the suit. The order of exposure to these three trials was randomized.

Fig. 1.

Description of the interventions for the two study aims. Blood samples were obtained at: baseline, prebody negative pressure (LBNP), and post-LBNP (<1, 30, and 60 min). Each session was separated by ~60 days. Cooling of hyperthermic subjects started shortly after ceasing LBNP. “Isodehydrated” denotes subjects exercised until a change in core body temperature (ΔT_cb) similar to the change during hyperthermic hydrated exercise for 90 min.

Aim 2: exercise-induced hyperthermia and dehydration.

This aim also required three unique laboratory visits, with each visit separated by ~60 days. The subject exercised on a treadmill in a heated chamber (40°C and 30% relative humidity) before being exposed to LBNP. Subjects were exposed to the following three sessions: 1) subjects exercised for 90 min at a moderate workload (~55% of maximal aerobic capacity) while being provided approximately body temperature drinking water (38.6 ± 1.0°C) every 15 min to offset sweat loss (hydrated), with sweat loss determined by weighing the subjects at 15-min increments; 2) subjects exercised at the same workload for 90 min, without being provided water (dehydrated-90); and 3) subjects exercised at the same workload without fluid ingestion until their core body temperature increased similar to the core body temperature achieved during the hydrated trial (isodehydrated). Upon completing each of these exercise trials, and while remaining in the heated chamber, subjects were rapidly transferred to the LBNP chamber, and the LBNP ramp ensued. This study was originally designed to evaluate differences between the hydrated and isodehydrated trials. However, given differences in exercise duration between these trials, the dehydrated-90 trial was later added. Thus, the order of these visits was not randomized.

LBNP Procedure

For both aims, subjects were supine with the lower body in the LBNP chamber, which was sealed at the iliac crest, and instructed to avoid leg movements. The LBNP trial followed a stepwise intensification of chamber decompression starting at 20 mmHg, with the level of LBNP increasing by 10 mmHg every 3 min. Termination of LBNP was determined by one or more of the following criteria: 1) continuous decline of systolic blood pressure below 80 mmHg, 2) sudden drop in heart rate, and/or 3) voluntary termination by the subject because of presyncopal symptoms such as dizziness, impaired vision, nausea, sweating, or malaise.

Blood Samples

At baseline, pre-LBNP, and <1, 30, and 60 min post-LBNP, blood was sampled in 3.2% sodium citrate (BD; BD Vacutainer Citrate Tubes) for plasma markers and thrombelastography (TEG) and in disodium EDTA (BD; BD Vacutainer K2EDTA Tubes) for cell counts. Blood samples for plasma analyses were centrifuged (15 min at 2,000 g and room temperature), whereafter plasma was immediately isolated and stored at −80°C until analysis.

Plasma Markers of Hemostatic Activity

Coagulation function was assessed with prothrombin time (PT), activated partial thromboplastin time (aPTT), fibrinogen, coagulation factor V (FV), coagulation factor VIII, D-dimer, von Willebrand factor (vWF), and antithrombin III (ATIII) tests (STA-R Evolution; Diagnostica Stago, Parsippany, NJ). Quantitative determination of tissue plasminogen activator (tPA) was performed by enzyme-linked immunosorbent assay using the ASSERACHROM tPA kit (Diagnostica Stago, S.A.S. France). Analyses for all samples were processed per manufacturer’s directions at 37°C.

Thrombelastography

Citrated blood was analyzed for coagulation competence (TEG model 5000; Hemoscope). Within 30 min of collection, 1 ml of blood was activated by kaolin, and 340 µl blood were added to a TEG cup containing 20 µl 0.2 M CaCl₂ and analyzed at 37°C for reaction time until initial fibrin formation, rate of clot formation, maximal amplitude reflecting clot strength (MA), overall clot function and effectiveness, and clot lysis 30 min after MA (LY₃₀).

Hematology

Within 2 h of collection, EDTA-stabilized blood was analyzed for red blood cell counts, hematocrit, white blood cell count, and platelet count (CELL-DYN 3700; Abbott Diagnostics). Percentage change in plasma volumes was calculated using the methods described by Dill and Costill (8).

Statistical Analysis

Factors are presented as means ± SE. For both aims, two-way ANOVA repeated-measures fixed-effect tests were used to compare each factor across the different treatment groups over time. Factors with a significant interaction effect were further analyzed to determine treatment effects at each time point. P values were therefore calculated to determine differences between interventions (“treatment P value”), over the time course of each visit from baseline to 60 min post-LBNP (“time P value”), and a combination of both (“treatment × time P value”). Data with high variance were log transformed to facilitate normal distribution and variance (version 9.2; SAS Institute, Cary, NC). The probabilities of observing chance effects on the dependent variables of interest are presented as exact P values.

RESULTS

Aim 1: Passive Hyperthermia

Four subjects did not participate in the non-LBNP time control session while all data from one subject were excluded because the subject did not elicit cardiovascular compensation (i.e., did not exhibit an appropriate increase in heart rate) to progressive hypovolemia during the normothermic trial, resulting in uncommonly low LBNP tolerance.

Core body temperature and plasma volume changed with hyperthermia and subsequent LBNP interventions (Fig. 2) (P < 0.001, treatment × time). During H-LBNP, calculated plasma volume decreased by 10.5% at pre-LBNP, 17.8% 1 min post-LBNP, and 8.1% at 30 and 8.3% at 60 min post-LBNP. These reductions in plasma volume were greater relative to the respective values throughout the time control trial, where the decrease in plasma volume ranged between 3.2 and 3.5%. During N-LBNP, plasma volume decreased by 16.1% 1 min post-LBNP but was similar at all other time points compared with the control values. Compared with the time control trial, there were notable treatment and time effects (main-effects model) for hematocrit and calcium that were greater during both N-LBNP and H-LBNP (Fig. 2) (P < 0.01). However, these changes were consistent with the concentrating effect of the calculated plasma volume decreases.

Fig. 2.

Physiological changes in aim 1: passive hyperthermia trial. Hyper LBNP, hyperthermic lower-body negative pressure; Norm LBNP, normothermic lower-body negative pressure; Plt, platelet. P < 0.05, treatment effect (*), time effect (†), and treatment × time effect (‡). Note: for temperature, values were not obtained at 60 min or for the control group.

Subjects in the H-LBNP trial had a higher platelet count at all time points than during both the N-LBNP and time control trials (Fig. 2) (P < 0.001), most markedly immediately after LBNP when platelet count was 40% greater than the time control trial. Changes in coagulation activity over time for variables that were statistically significant (treatment × time fixed-effects model, P < 0.05) are noted in Fig. 3. The most pronounced effects were in factor V (P < 0.002), factor VIII (P = 0.06), and vWF antigen (P < 0.001), which remained elevated compared with the time control trial at the final blood draw (equivalent to 60 min post-LBNP), although there were no differences between N-LBNP and H-LBNP. Factor V values were higher at all time points for both N-LBNP and H-LBNP, peaking ~20% higher than the time control trial at 1 min post-LBNP, which is ~6% higher than what would be expected for concentration effects of hypovolemia. Factor VIII concentrations were significantly increased (20–58%) post-LBNP during both the hyperthermic and normothermic trial, compared with the time control trial, well above a simple concentration effect. vWF antigen exhibited a similar pattern of increase (30–47%) post-LBNP, again, well above a concentration effect (Fig. 3). Both N-LBNP and H-LBNP had shortened aPTT at 60 min post-LBNP (treatment, P < 0.0001; treatment × time, P = 0.0747). Most TEG values were similar between trials (Table 1). Hyperthemia alone increased fibrinolysis based on LY₃₀ immediately before and after LBNP, although this returned to baseline values at 30 and 60 min post-LBNP (and cooling) (Fig. 4) (P = 0.05, treatment; P = 0.001, time; P = 0.13, treatment × time). Whereas there was an increased concentration of fibrinogen with LBNP, LBNP induced a significant spike in tPA concentration (P = 0.0002, treatment × time) and increase in clot lysis (LY₃₀) more so in the H-LBNP trial compared with N-LBNP, although values later returned to baseline (Fig. 4). There were no differences in other TEG values or D-dimer levels between any trials (Table 1).

Fig. 3.

Coagulation activity of aim 1: passive hyperthermia trial. Control, normothermic, no lower-body negative pressure (LBNP) (n = 7 subjects); Hyper LBNP, hyperthermic LBNP (n = 11 subjects); Norm LBNP, normothermic LBNP (n = 11 subjects); PT, prothrombin time; aPTT, activated partial thromboplastin time; AT3, antithrombin 3; vWF, von Willebrand factor. P < 0.05, time effect (†) and treatment × time effect (‡) (repeated-measures ANOVA fixed-effect test).

Table 1.

Results of aim 1: passive hyperthermia trial

Factor	Treatment	Baseline	Pre-LBNP	Time Post-LBNP, min			P Value
				1	30	60	Treatment	Time	Treatment × Time
d-Dimer, µg/ml	Control	0.3	0.3	0.4	0.4	0.4	0.37	0.04	0.50
	Hyper LBNP	0.2	0.2	0.4	0.6	0.6
	Norm LBNP	0.2	0.2	0.3	0.3	0.3
TEG angle, deg	Control	60.6	62.2	64.5	65.0	64.0	0.55	<0.0001	0.79
	Hyper LBNP	61.3	63.3	64.6	66.6	66.6
	Norm LBNP	61.6	62.7	63.6	65.8	65.1
TEG G, dyn/s	Control	6,992.9	7,066.3	7,674.2	7,518.5	7,398.7	0.97	<0.0001	0.88
	Hyper LBNP	6,850.0	7,048.2	7,840.8	7,891.0	7,496.7
	Norm LBNP	7,005.9	7,256.8	7,915.4	7,694.3	7,493.2
TEG MA, mm	Control	58.0	58.2	60.3	59.8	59.4	0.98	<0.0001	0.85
	Hyper LBNP	57.4	58.0	60.8	60.9	59.7
	Norm LBNP	58.1	58.9	60.9	60.4	59.6
TEG R, min	Control	7.2	6.1	5.2	5.1	5.5	0.56	<0.0001	0.29
	Hyper LBNP	6.8	6.0	5.4	4.4	4.5
	Norm LBNP	6.8	6.2	5.7	4.8	5.1

Control, time control trial in which subjects remained supine and normothermic without lower-body negative pressure (LBNP) (n = 7 subjects); Hyper LBNP, hyperthermic LBNP (n = 11 subjects); Norm LBNP, normothermic LBNP (n = 11 subjects); TEG, thromboelastography; MA, maximum amplitude P values based on repeated-measures ANOVA fixed-effect tests; “treatment,” comparison between interventions; “time,” comparison between time points.

Fig. 4.

Aim 1 fibrinogen and fibrinolytic activity. Control, normothermic, no lower-body negative pressure (LBNP) (n = 7 subjects); Hyper LBNP, hyperthermic LBNP (n = 11 subjects); Norm LBNP, normothermic LBNP (n = 11 subjects); LY₃₀, thrombelastography (TEG) lysis value at 30 min; tPA, tissue plasminogen activator. P < 0.05, time effect (†) and treatment × time effect (‡) (repeated-measures ANOVA fixed-effect test).

Aim 2: Exercise-Induced Hyperthermia and Dehydration

Three subjects were unable to return for the dehydrated-90 trial. Blood composition and hemostatic variables at baseline and during the interventions are reported in Table 2. Core body temperatures remained higher and plasma volumes were ~15% lower in the dehydrated-90 trial (Fig. 5) (P < 0.001). Calcium trended highest for the dehydrated-90 session throughout the trial (Fig. 5) (P = 0.0003, treatment × time). Between trials, there were no statistical treatment effects on coagulation activity (e.g., PT, aPTT, ATIII, vWF, and FV) (Fig. 6). The hydrated subjects had decreased lysis at all time points in the trial (2–4%) compared with both the dehydrated-90 and isodehydrated sessions, based on LY₃₀ TEG values, which ranged between 6 and 10% (Fig. 7) (P = 0.0002, treatment; P = not significant for time and treatment × time). There were no differences detected with other TEG values or D-dimer levels (Table 2). In the dehydrated-90 trial, tPA concentration trended higher throughout the trial (Fig. 7) (P = 0.034, treatment × time effect).

Table 2.

Results of aim 2: Exercised-induced hyperthermia and dehydration trial

Factor	Treatment	Baseline	Pre-LBNP	Time Post-LBNP, min			P Value
				1	30	60	Treatment	Time	Treatment × Time
d-Dimer, µg/ml	Dehydrated-90	0.2	0.3	0.2	0.2	0.1	0.18	0.25	0.56
	Isodehydrated	0.2	0.2	0.2	0.2	0.2
	Hydrated	0.2	0.4	0.3	0.4	0.2
TEG angle, deg	Dehydrated-90	63.3	65.2	64.4	66.6	67.5	0.51	<0.01	0.84
	Isodehydrated	61.8	62.8	64.6	63.8	64.3
	Hydrated	62.3	64.1	64.1	65.4	65.4
TEG G, dyn/s	Dehydrated-90	6,021.1	7,665.5	6,900.7	6,671.3	7,526.8	0.54	<0.0001	0.54
	Isodehydrated	6,556.3	7,608.7	7,408.0	6,981.8	7,169.0
	Hydrated	7,008.2	8,037.4	7,587.6	7,504.2	7,871.0
TEG MA, mm	Dehydrated-90	53.9	59.7	56.5	57.6	59.1	0.29	<0.0001	0.34
	Isodehydrated	55.8	59.9	58.7	56.8	57.9
	Hydrated	58.3	61.8	60.3	60.1	61.5
TEG R, min	Dehydrated-90	6.8	6.0	5.4	4.9	4.8	0.79	<0.0001	1.00
	Isodehydrated	7.0	6.5	5.4	5.2	5.3
	Hydrated	6.9	6.2	5.6	5.1	5.3

Dehydrated-90, exercise for 90 min in the heat without hydration (n = 9 subjects); isodehydrated, exercise in the heat without hydration until the same core body temperature achieved in the hydrated trial (n = 12 subject); hydrated, exercise in the heat for 90 min (n = 12 subjects); TEG, thromboelastography; MA, maximum amplitude. P values are based on repeated-measures ANOVA fixed-effect tests; “treatment,” comparison between interventions; “time,” comparison between time points.

Fig. 5.

Physiological changes in aim 2: exercised-induced hyperthermia and dehydration trial. dehydrated-90, Exercise for 90 min in the heat without hydration (n = 9 subjects); isodehydrated, exercise in the heat without hydration until the same core body temperature achieved in the hydrated trial (n = 12 subjects); hydrated, exercise in the heat for 90 min (n = 12 subjects); Plt, platelet. P < 0.05, treatment effect (*), time effect (†), and treatment × time effect (‡) (repeated-measures ANOVA fixed-effect test).

Fig. 6.

Coagulation activity of aim 2: exercise-induced hyperthermia and dehydration. dehydrated-90, Exercise for 90 min in the heat without hydration (n = 9 subjects); isodehydrated, exercise in the heat without hydration until the same core body temperature achieved in the hydrated trial (n = 12 subjects); hydrated, exercise in the heat for 90 min (n = 12 subjects); PT, prothrombin time; aPTT, activated partial thromboplastin time; AT3, antithrombin 3; vWF, von Willebrand factor. P < 0.05, time effect (†) and treatment × time effect (‡) (repeated-measures ANOVA fixed-effect test).

Fig. 7.

Aim 2 fibrinogen and fibrinolytic activity. dehydrated-90, Exercise for 90 min in the heat without hydration (n = 9 subjects); isodehydrated, exercise in the heat without hydration until the same core body temperature achieved in the hydrated trial (n = 12 subjects); hydrated, exercise in the heat for 90 min (n = 12 subjects); LY₃₀, thrombelastography (TEG) lysis value at 30 min; tPA, tissue plasminogen activator. P < 0.05, treatment effect (*), time effect (†), and treatment × time effect (‡) (repeated-measures ANOVA fixed-effect test).

DISCUSSION

This is the first study to report the hemostatic profile of humans exposed to a combination of dehydration, hyperthermia, and exercise followed by progressive central hypovolemia. In this study of healthy subjects that undertook LBNP-induced central hypovolemia, we found that central hypovolemia was the primary driver of hemostasis, beyond the additional effects of hyperthermia, dehydration, and exercise (Fig. 8). Fibrinolysis, however, is primarily activated by hyperthermia and dehydration, with little effect from LBNP. The one exception was that hyperthermia did enhance hemostasis by way of thrombocytosis and likely platelet activation. Most notably, we found that the increased concentrations of factors V and VIII, vWF antigen, and fibrinogen are likely part of the mechanisms of hypovolemia-activated hemostasis, although we did not prove causality. In the hyperthermic condition, dehydration did not affect exercise-activated or subsequent hypovolemia-activated hemostasis, disproving our hypothesis.

Fig. 8.

Conceptual model of the effects of dehydration, lower-body negative pressure (LBNP)-induced central hypovolemia, and hyperthermia on clot formation and fibrinolysis (clot lysis). Dehydration and hyperthermia both increased tissue plasminogen activator (tPA) concentrations and were found to increase clot lysis, based on increased LY₃₀ results. Hyperthermia also increased platelet count, which contributes to clot formation. LBNP-induced central hypovolemia also causes increased platelet count and increases factors V and VIII, von Willebrand factor (vWF), and fibrinogen, leading to improved clot formation and hemostasis.

Aim 1: Passive Heat Stress

Our findings confirm previous human and in vitro work correlating LBNP with increased hemostasis. Previously, in a noncontrolled study of 11 healthy subjects, Meyer and colleagues found that hyperthermia and subsequent central hypovolemia induced a hypercoagulable state (28). Given that we found no notable differences in hemostasis between the N-LBNP and H-LBNP groups, it seems that the hypercoagulable state from Meyer’s trial was more likely the result of the LBNP intervention, as opposed to the hyperthermia. Whereas H-LBNP and N-LBNP did not demonstrate notable differences on most assessments, hyperthermia did seem to induce a relative thrombocytosis. Others have reported hyperthermia-induced increases in plasma β-thromboglobulin and platelet factor 4, suggesting that hyperthermia activates platelets (38). More recently, Lawrence and colleagues investigated a newer measure of viscoelastic clot strength (clot microstructure), including the fractal dimension of clot (21, 22). With whole blood samples from 136 volunteers, there seemed to be little effect on clot microstructure at blood temperatures above 37°C, despite a correlation with decreased temperature and poor clot strength (21). Our findings confirm a prior report where hyperthermia activated fibrinolysis (9), since we found that, in aim 1, hyperthermia increased lysis (LY₃₀). Notably, in vitro tests of elevated blood temperature should be distinguished from a core body temperature measured from subjects in vivo, since the latter reflects the effects of hyperthermia on endothelium and, indeed, the entire organism. In this study, the increased lysis and tPA concentrations noted in the aim 1 hyperthermic limb returned to baseline by 30 min post-LBNP. This may reflect the balance of the fibrinolytic system, confirming work that others have noted, i.e., hyperthermia can induce plasminogen activator inhibitor-1 antigen release from platelets and endothelial cells and cause tPA release. This has previously been shown in several studies both in vivo (20, 40) and in vitro (13, 17, 33). Another explanation is that, as subjects’ temperatures decreased after heating, there was simply less hyperthermia-induced fibrinolysis.

Activation of the coagulation system by LBNP models of central hypovolemia has previously been identified (4, 26, 41, 43, 44). Several mechanisms have been discussed for this response. In separate studies of healthy volunteers, Zaar and colleagues identified platelet activation through activation of glycoprotein IIb/IIIa, acceleration of clot kinetics demonstrated on TEG, increased thrombin generation, and increased plasma protein C during progressive LBNP (43, 44). In a study of controlled blood loss of healthy patients, Zaar also found activation of the coagulation system with up to 900 ml blood removed (45). A comparison of controlled LBNP and blood withdrawal models found similarly accelerated coagulation via TEG (41). It is also possible that endogenous catecholamine release during hypovolemia may contribute to the observed coagulation changes. Consistent with that hypothesis, previous work has demonstrated activation of platelets and enhanced hemostasis with administration of epinephrine (42).

Aim 2: Exercise Heat Stress With and Without Accompanying Dehydration

The effects of exercise itself on coagulation are well studied (5, 14, 27, 32). Whereas the mechanisms are still being elucidated, increased platelet and procoagulant activity has been observed (32). These effects are more notable with higher-intensity exercises. However, in those studies, body temperature was not assessed, environmental conditions varied, and hydration status was not controlled. There are less data on exercise and fibrinolysis, although many studies have shown increased tPA levels with high-intensity exercise (20, 32). Other variables that must be considered here are the subjects’ baseline level of exercise fitness and overall health, which can alter their response to high-intensity exercise. That said, in the present study, subjects exercised at the same relative workload (e.g., ~55% of each individual’s maximal aerobic capacity). Another study revealed activation of clotting and increased clot strength, measured on viscoelastic testing and standard coagulation assays, after a marathon run (39). Results from aim 2 confirmed these results, while showing that modest dehydration had little additional impact.

There are few data on the isolated effects of dehydration and hemoconcentration on hemostasis; however, the theoretical implications of dehydration leading to increased viscosity and hypercoagulability based on Virchow’s triad are often acknowledged (24). A clinical example of this may be seen in infants experiencing hypernatremic dehydration who are at a high risk of central sinus venous thrombosis (7). Hydration has been studied in the setting of LBNP and hyperthermia previously, where it was found that hydrated subjects better tolerated the presyncopal effects of LBNP, although coagulation profiles were not reported (25, 36). The present data suggest that hydration decreased clot lysis in hyperthermic subjects. This may be in part explained by the trend of higher tPA concentration noted in the dehydrated-90 subjects noted at all time points (Fig. 7).

Perspectives and Significance

Given that dehydration, hyperthermia, and exercise themselves represent either absolute (dehydration) or relative (hyperthermia, exercise) forms of hypovolemia, it is not surprising that central hypovolemia proved to be a primary stimulant of hemostasis in this study, with the exception of hyperthermia stimulating thrombocytosis and clot lysis (35). A significant clinical implication is that quickly restoring a hemorrhaging hypovolemic trauma patient with relatively cold noncoagulant fluids (crystalloids) can have serious deleterious effects on the body’s innate ability to form essential clots. This further gives credence to the body of literature behind the concepts of damage control resuscitation (15, 18), limiting crystalloid fluid administration (19), avoiding hypothermia (23, 31), hypotensive resuscitation (1, 30), and restoring circulating blood volume with blood products that do not exacerbate a coagulopathy (16, 37). As noted earlier, hydration decreased clot lysis in hyperthermic subjects. Hydration status may therefore have potential in preventing hyperfibrinolysis in trauma patients, or those at risk of trauma, similar to how antifibrinolytics are now administered in the setting of bleeding trauma patients (36a). Furthermore, this emphasizes the importance of monitoring for hyperfibrinolysis in trauma patients, which has been recently shown to improve outcomes (11, 29).

Limitations

The major limitation of this study was not collecting and measuring the multiple other factors that are involved in the complex environment of the coagulation and fibrinolytic systems. In retrospect, measuring platelet activation by way of platelet mapping would have helped elucidate the role of platelets in this process. As discussed above, measuring plasminogen activator inhibitor-1 antigen would better indicate its role in decreasing fibrinolysis. Additionally, a measure of thrombin-generating potential with a thrombogram and measuring thrombin-antithrombin complexes would give a better explanation of the mechanisms underlying the effects of hypovolemia and hyperthermia on coagulation.

In general, blood samples were obtained after a relatively short duration of LBNP and hyperthermic exposure. Furthermore, LBNP cannot fully simulate the complicated human response to trauma, massive tissue injury, and trauma-induced coagulopathy. Our results may therefore not be translatable to a common clinical situation with prolonged central hypovolemia, shock, and/or hyperthermia, which may have progressive effects on the coagulation system. This has been described recently as “blood failure,” since progressive oxygen debt can lead to an endotheliopathy, which causes subsequent coagulation and fibrinolytic derangements (2).

At the conclusion of the LBNP challenge for both aims 1 and 2, the thermal conditions were returned to preheating baseline, i.e., cool followed by thermoneutral water was perfused through the water-perfused suit for aim 1, whereas the environmental chamber temperature was reduced to normothermic conditions in aim 2. Thus, the environmental conditions during the 60 min of data collection after the end of the simulated hemorrhagic challenges were different relative to the peak of the heating periods. We recognize and accept the limitations in doing so with respect to simulating a “field” condition, when a hemorrhaging soldier is frequently warmed (3).

Finally, the small sample sizes and subject attrition in this study may have resulted in an underpowered study to reveal clinically important differences between treatment groups.

In conclusion, humans, hypovolemia is a primary driver of hemostasis compared with the influence of hyperthermia and dehydration. However, hyperthermia does induce a significant thrombocytosis and by itself causes an increase in clot lysis. Dehydration increases clot lysis but does not seem to affect exercise-activated or subsequent hypovolemia-activated hemostasis in hyperthermic humans.

GRANTS

This research was supported in part by an appointment to the Postgraduate Research Participation Program at the US Army Institute of Surgical Research administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the US Department of Energy, the US Army Medical Research and Materiel Command (W81XWH-12-1-0152 to C. G. Crandall), and the National Heart, Lung, and Blood Institute (HL-61388 to C. G. Crandall).

DISCLOSURES

The view(s) expressed herein are those of the author(s) and do not reflect the official policy or position of Brooke Army Medical Center, the US Army Medical Department, the US Army Office of the Surgeon General, the Department of the Air Force, the Department of the Army or the Department of Defense or the US Government. No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

V.A.C. and C.G.C. conceived and designed research; M.A.B., M.Z., J.K.A., and C.G.C. analyzed data; M.A.B., M.Z., J.K.A., Z.J.S., D.G., V.A.C., A.P.C., and C.G.C. interpreted results of experiments; M.A.B. prepared figures; M.A.B. and C.G.C. drafted manuscript; M.A.B., M.Z., J.K.A., Z.J.S., D.G., E.R., J.K., V.A.C., A.P.C., and C.G.C. edited and revised manuscript; M.A.B., M.Z., J.K.A., Z.J.S., D.G., E.R., J.K., V.A.C., A.P.C., and C.G.C. approved final version of manuscript; Z.J.S., D.G., E.R., J.K., and C.G.C. performed experiments.