Moderate hypoxia suppresses exercise-induced procoagulant changes


Thomas DeLoughery, Hematology L586, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239-3098, USA.


Hypoxia has been implicated as a stimulant of coagulation. As exertion is known to affect haemostasis, we sought to control for this by using a standardized protocol. Subjects were exercised both at room air and at 12% oxygen. Exercise produced an increase in procoagulant factors, which was reduced with hypoxic exercise. Room air exercise increased fibrinolytic markers. Hypoxic exercise did not affect the increase in tissue plasminogen activator, but decreased the increase in plasminogen activator inhibitor-1 expression. Thus, it appears that hypoxia may exert an antithrombotic effect by both damping exercise-induced procoagulant changes and stimulating fibrinolysis.

One of the risks of mountaineering is thrombosis (Cucinell & Pitts, 1987). Both venous and arterial thrombosis have been reported in climbers (LeRoux et al, 1992) but the aetiology of this risk is unknown. It has been hypothesized that hypoxia may induce a procoagulant state. The evidence for this is based on basic science and animal studies. It has been harder to prove hypoxia is a procoagulant stimulus in human beings either in altitude chambers or field studies (Grover & Bartsch, 2002). In addition, few of these studies have controlled for exercise. As exercise is a potent initiator of coagulation, this is an important variable to control (Andrew et al, 1986; el-Sayed, 1996). In addition, exercise may synergize with underlying hypercoagulable states to increase coagulation activation (Weiss et al, 2003). We tested this idea using a standard protocol that hypoxia plus exercise is a more potent stimulant of coagulation than isolated hypoxia or exercise. Surprisingly, we found that hypoxia actually decreased exercise-induced coagulation activation.

Materials and methods

The protocol was approved by the Institutional Review Board. Eight subjects underwent baseline exercise testing on an exercise treadmill (Medigraphic, St Paul, MN, USA) to determine the maximum exercise capacity (VO2 max). At least 24 h later the exercise test was repeated. The subjects exercised to 50% VO2 max and maintained this level for 10 min, then increased the exercise level to 75% VO2 max for a further 10 min. The exercise level was then increased to 100% VO2 max. At the next testing session subjects were made hypoxic by breathing oxygen admixed with nitrogen to achieve an inhaled oxygen percentage of 12% (equivalent to an altitude of 4600 m). After 30 min of hypoxia the exercise test was repeated using the same protocol but under hypoxic conditions. Blood was drawn at baseline, after 30 min of hypoxia and at 50%, 75%, and 100% VO2 max. As preliminary data indicated no significant changes at 50% VO2 max, the blood draw from this time point was eliminated.

Laboratory assays

Laboratory assays were performed either in the General Research Center Core laboratory or the Oregon Health and Science University Special Hemostasis Laboratory. The following assays were performed: factor VIII (FVIII) levels [Stago-activated partial thromboplastin time (APTT) automate reagent, Stago VIII DP), von Willebrand factor antigen (VWF; Stago VW Liatest), DDimer [Asserachrom DDi enzyme-linked immunosorbent assay (ELISA)] (all Diagnostica Stago, Parsippany, NJ, USA), prothrombin fragment 1.2 (F1.2; DadeBehring F1.2 kit; Dade-Behring, Deerfield, IL, USA) tissue plasminogen activator (tPA) and plasminogen activator inhibitor (PAI; TintElize ELISA, Biopool International, Leiden, the Netherlands), thrombin activator fibrinolysis inhibitor (TAFI; American Diagnostica, Stamford, CT, USA). Statistics were analysed using StatView (SAS, Cary, NC, USA). P < 0·05 was considered significant.



All subjects tolerated the hypoxic exercise. VO2 max was 44 ± 1 ml/kg/min and fell to 33 ± 0·9 ml/kg/min with hypoxic exercise. The 30-min exposure to hypoxia produced no significant changes in any coagulation parameter (Table I).

Table I.  Haematology and coagulation response after 30 min of 12% fractions of inspired O2 (FIO2).
  1. WBC, white blood cell count; Hct, haematocrit; Plt, platelet count; Retic, reticulocytes; MPV, mean platelet volume; VWF, von Willebrand factor; F1.2, prothrombin fragment 1 + 2; DD, D-dimer; EGCLT, euglobulin clot lysis time; PAI-1, plasminogen activator inhibitor-1; tPA, tissue plasminogen activator; TAFI, thrombin activator fibrinolysis inhibitor.

  2. Results expressed as mean ± standard error.

WBC (×109/l)4·883 ± 0·3025·783 ± 0·761180·22
Hct (%)40·1 ± 1·341·5 ± 1·430·27
Plt (×109/l)216 ± 9·1224 ± 7·230·04
Retic (%)100155 ± 26550·09
MPV (fl)8·6 ± 0·38·9 ± 0·2630·20
Factor VIII (U/dl)140 ± 12147 ± 1250·25
VWF (U/dl)110 ± 10125 ± 13120·23
F1.2 (nmol/l)0·56 ± 0·111·19 ± 0·66330·25
DD (nmol/l)151·5 ± 4153·5 ± 3910·75
EGCLT (%)10075250·14
PAI-1 (ng/ml)26·9 ± 7·924·7 ± 6−80·90
TPA (ng/ml)4·7 ± 0·774·3 ± 0·8−80·42
TAFI (ng/ml)18·8 ± 1·918·5 ± 1·8−10·6

With normoxic exercise to VO2 max, the haematocrit and white blood cell count increased significantly. With hypoxic exercise, the haematocrit increase was not as high (Table II). Under normoxic conditions, the platelet count rose at 75% VO2 max but fell to baseline at VO2 max. In the hypoxic tests the platelet counts rose by 15% at VO2 max (P = 0·04). The mean platelet volume increased in the normoxic tests but not under hypoxia.

Table II.  Haematology and coagulation response with normoxic and hypoxic exercise.
TestBaseline75% VO2 max100% VO2 max%P*P
  1. -C, control; -H, hypoxic; WBC, white blood cell count (×109/l); Hct, haematocrit (%); Plt, platelet count (×109/l); Retic, reticulocytes (%); MPV, mean platelet volume (fl); VWF, von Willebrand factor (U/dl); F1.2, prothrombin fragment 1 + 2 (nmol/l); DD, D-dimer (nmol/l); EGCLT, euglobulin clot lysis time (%); PAI-1, plasminogen activator inhibitor-1 (ng/ml); tPA, tissue plasminogen activator (ng/ml); TAFI, thrombin activator fibrinolysis inhibitor (ng/ml).

  2. Results expressed as mean ± standard error.

  3. *Baseline vs. 100% VO2 max.

  4. †Control 100% VO2 max vs. hypoxic 100% VO2 max.

WBC-C4·783 ± 0·1497·267 ± 0·58·65 ± 0·693810·0140·19
WBC-H5·783 ± 0·7617·666 ± 0·3958·216 ± 0·693420·05
Hct-C41·1 ± 1·144·1 ± 1·3544·53 ± 1·080·0010·02
Hct-H41·5 ± 1·442·9 ± 1·343·6 ± 1·050·023
Plt-C215 ± 15225 ± 8·7212 ± 1710·930·03
Plt-H224 ± 7·2252 ± 9·4258 ± 14150·002
Retic-C100199 ± 55272 ± 1401720·280·93
Retic-H100102 ± 38187 ± 291870·007
MPV-C8·8 ± 329·6 ± 0·289·6 ± 0·34190·010·01
MPV-H8·85 ± 0·269·1 ± 0·239·1 ± 3130·14
Factor VIII-C133 ± 16227 ± 40500 ± 572760·010·02
Factor VIII-H147 ± 12194 ± 32274 ± 50870·04
VWF-C110 ± 10138 ± 10198 ± 21800·020·60
VWF-H125 ± 13140 ± 16171 ± 16370·04
F1.2-C0·84 ± 0·361·39 ± 0·372·21 ± 0·521630·010·07
F1.2-H1.2 ± 0·241·19 ± 0·620·93 ± 0·25200·55
DD-C147 ± 29n/a6000 ± 360068350·030·04
DD-H139 ± 39n/a200 ± 38440·04
EGCLT-C100%55 ± 2526 ± 12·2740·020·70
EGCLT-H74·9 ± 1432 ± 1015·5 ± 4·8790·01
PAI-C20·6 ± 1·537·9 ± 7·637·8 ± 6·0830·020·16
PAI-H24·7 ± 6·029·8 ± 8·530·2 ± 8·4220·11
tPA-C4·7 ± 0·911·8 ± 8·126·5 ± 6·64600·010·53
tPA-H4·3 ± 2·112·1 ± 7·925·9 ± 6·15000·01
TAFI-C19·9 ± 2·420·0 ± 2·216·6 ± 2·5−160·120·25
TAFI-H18·5 ± 1·918·2 ± 1·918·6 ± 2·1−0·50·92

In the normoxic tests, all the markers of coagulation studied were increased. FVIII and von Willebrand protein levels rose by 276% and 80% respectively. Levels of the prothrombin fragment F1.2, a marker of thrombin activation, increased by 163% at VO2 max. Levels of the fibrin degradation product, D-dimer, were markedly increased, by almost 8000%, with normoxic exercise.

In contrast, with hypoxic exercise the coagulation response was low. FVIII and von Willebrand protein rose by only 80% and 37%, respectively, both significantly lower than the normoxic study. F1.2 levels were not significantly elevated and the D-dimer level rose by only 44%.

The euglobulin clot lysis time was shortened in both groups. tPA levels at VO2 max in both conditions rose by c. 500% with no difference between normoxia and hypoxia. However, under normoxic conditions, PAI-1, the inhibitor of tPA, was significantly elevated by 80% at VO2 max but was unchanged under hypoxic conditions. TAFI was unchanged in both conditions.


In this study, we failed to show that hypoxia-amplified exercise-induced procoagulant changes. In fact, these procoagulant changes were not as prominent under hypoxic conditions.

The data indicating that moderate hypoxia alone could activate coagulation is controversial. O'Brodovich et al (1984) found an increase in FVIII but no change in fragment E with decompression (O'Brodovich et al, 1984). In a widely publicized study, Bendz et al (2000) found an increase in F1.2 within 2 h of mild hypoxic exposure with no change in D-dimer. However, in a more recent study this finding of increase F1.2 with hypoxia could not be replicated (Crosby et al, 2003).

However, these studies are at variance with other studies. Andrew et al (1987) found no changes in coagulation factors with exposure to simulated altitude up to 8488 m (280 Torr). An extensive series of high altitude studies found no increase in fibrinopeptide  A or F1.2 at altitudes ranges from 3457 to 4559 m (Bartsch et al, 1982, 2001). The bulk of experimental data does not support activation of coagulation by modest hypoxia (Grover & Bartsch, 2002).

Two previous studies have examined hypoxic exercise and coagulation. One reported a reduced increase in FVIII with hypoxic exercise at 380 (18 000 ft) and 282 Torr (25 000 ft) compared with baseline (Andrew et al, 1987). Bartsch et al (1982) studied exercise at 3457 m and found a 100% rise in FVIII when compared with a non-exercising high-altitude control group, but no increase in fibrinopeptide  A.

The lower exercise-induced activation coagulation with hypoxia could simply be due to the decreased intensity of exercise. Andrew et al (1986) showed that a rise in FVIII was dependent on exercise intensity, with FVIII elevations seen only with a minimum exercise intensity of 80% VO2 max and over. This may be one explanation, as the VO2 max dropped by 25% with hypoxia in our study.

Of note, tPA elevation and the shortening of the euglobulin clot lysis time was not effected by hypoxia. However, the exercise-stimulated PAI-1 rise was lowered by hypoxia. This suggests that the fibrinolytic response to exercise may not decrease with, or even be augmented by, hypoxia.

In summary, we have shown that moderate hypoxia lowers exercise-induced procoagulant changes but does not alter the fibrinolytic response. These results suggest that initial exposure to moderate altitudes does not lead to a hypercoagulable state. However, we cannot rule-out that higher altitudes or more prolonged exposure may induce procoagulant changes.


This study was funded by National Institute of Health Clinical Research Center Grant M01 RR000334 and Mazama Research Grant (Portland, OR, USA).