Effects of Lifestyle on Hemostasis, Fibrinolysis, and Platelet Reactivity a Systematic Review

REVIEW ARTICLE Effects of Lifestyle on Hemostasis, Fibrinolysis, and Platelet Reactivity A Systematic Review

Kaeng W. Lee, MRCP; Gregory Y. H. Lip, MD, FRCP

he pathophysiology of atherothrombosis in cardiovascular disease is complex and mul- tifactorial. No doubt, lifestyle habits such as exercise, smoking, diet, and alcohol consumption may have significant influence on cardiovascular disease. As the hemostatic system is assuming an increasingly prominent role in the pathogenesis and progres- Tsion of atherovascular diseases, this review evaluates the effects of lifestyle habits (or lifestyle modifications) on blood coagulation, fibrinolysis, and platelet reactivity. Arch Intern Med. 2003;163:2368-2392

The process of initiation, progression, and tions and the effects of exercise or physical complication of atherothrombosis in car- activity, psychosocial stress, diet, and other diovascular disease is complex and can be lifestyle habits on plasma indicators of influenced by multiple factors. Ischemic thrombogenesis are less well established. coronary syndromes such as unstable an- Further evidence of the influence of gina, myocardial infarction (MI), and sud- lifestyle changes on cardiovascular risk fac- den ischemic death share common patho- tors and clinical outcomes is illustrated by physiologic processes characterized by data from salt restriction and blood pres- coronary plaque rupture with superim- sure reduction11 and improved mortality posed thrombus formation.1,2 Indeed, there by diets rich in oily fish. In the Gruppo is substantial experimental and clinical evi- Italiano per lo Studio della Sopravvive- dence that blood hypercoagulability or nza nell’Infarto Miocardico Prevenzione thrombogenicity promotes thrombus for- (GISSI Prevenzione)12 trial and the Diet mation in the circulation, systemically and and Reinfarction Trial (DART),13 there was locally at the exposed atherogenic sur- a significant reduction in mortality after face of the disrupted plaque.3 MI by increasing dietary n-3 polyunsatu- A wide range of factors has been iden- rated fatty acids (n-3 PUFA) and fish in- tified in prospective epidemiologic studies take, respectively, and the mortality re- to have a systemic effect on blood throm- duction has been partly attributed to a bogenicity. Certainly, there is increasing evi- reduction in sudden cardiac death. In- dence of a close relationship between the deed, the recent reanalysis of the course traditional cardiovascular risk factors such of appearance of the effects of n-3 PUFA as diabetes mellitus, hypertension or hy- has showed an early and highly signifi- perlipidemia, and the increased thrombo- cant reduction of sudden cardiac death.14 genicity, which is characterized by hyper- However, many instances of sudden death coagulability, hypofibrinolysis, or increased have a thrombotic basis,15 with evidence platelet reactivity.4-6 Conversely, improve- of thrombus in the left main coronary ar- ments of these cardiovascular risk factors tery, and sudden death is not simply an have been associated with a lower prothrom- arrhythmogenic phenomenon. The aim of bic tendency.7-10 However, the associa- this review is to evaluate the effects of lifestyle habits (or lifestyle modifications) on From the Haemostasis, Thrombosis, and Vascular Biology Unit, University Department the plasma indices of the 3 main systems of Medicine, City Hospital, Birmingham, England. The authors have no relevant of thrombosis: blood coagulation, fibri- financial interest in this article. nolysis, and platelet reactivity.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2368

Downloaded From: https://jamanetwork.com/ on 09/23/2021 SEARCH STRATEGY of fibrin within the thrombus, disin- antigen (a marker of cross-linked fi- tegrating clots and hence maintain- brin turnover), and tPA antigen, We performed a search using elec- ing vascular patency. Fibrin de- have been identified as indepen- tronic databases (MEDLINE, grades into soluble fibrin degradation dent predictors of subsequent car- EMBASE, and DARE), and addition- products, including D-dimers. The diovascular events in prospective ally, abstracts from national and in- primary inhibitor of the fibrinolytic studies in both healthy subjects18-22 ternational cardiovascular meet- process is the plasminogen activa- and those with cardiovascular risk ings were reviewed to identify tor inhibitor type 1 (PAI-1), which factors23,24 or established coronary unpublished studies. Relevant au- inhibits plasminogen activation by heart disease (CHD).25 In addition, thors of these studies were con- binding with tPA to form the PAI/ platelet hyperaggregation,26,27 plasma tacted to obtain further data. tPA complexes. Therefore, im- viscosity,28-30 and PAI-1 levels31-34 paired fibrinolytic function may be re- have also been associated with car- BASIC PATHOPHYSIOLOGY flected in high plasma levels of PAI-1 diovascular morbidity and mortal- AND CLINICAL PERSPECTIVES or tPA antigen (which evaluates ity in both men and women in pro- mainly the inactive PAI-1/tPA com- spective studies. Thus, the potential The process of hemostasis and plexes) and/or indicated by low modifications of these hemostatic or thrombus formation depends on plasma levels of tPA activity or acti- thrombogenic factors by simple life- the fine balance between the co- vation products such as D-dimer and style changes as both primary and agulation and fibrinolysis systems plasmin 2–antiplasmin complex. Re- secondary prevention have at- (Figure). The slower intrinsic clot- duced plasmin generation leads to tracted considerable interest from ting pathway depends on circulat- suppression of fibrinolytic activity, the public health perspective. ing coagulation factors, such as fac- thus favoring fibrin persistence and The recognition that the onset tors IXa and VIIIa. The more rapid thrombosis. of cardiovascular events is fre- extrinsic pathway is activated when Most of these variables in the quently triggered by physical exer- blood is exposed to an extravascu- coagulation and fibrinolytic sys- tion or mental stresses has lead to a lar factor such as tissue factor. Fac- tems can be readily assayed using the possible link between neurohor- tor VII (FVII) plays a key role in the enzyme-linked immunosorbent as- monal activation and coronary ath- initiation of this coagulation mecha- say (ELISA) technique. It is impor- nism when it forms complexes with tant to distinguish the difference be- tissue factor from a disrupted ath- tween the measurement of activity Coagulation System eromatous plaque. Activation of the and antigen levels of these mol- coagulation system induces the for- ecules. Although the antigen levels Intrinsic Pathway Extrinsic Pathway mation of thrombin from prothrom- refer to the total amount of the cir- X IXa VIIa VII bin. Thrombin converts fibrinogen culating proteins (both bound and VIIIa TF into (insoluble) fibrin and induces free), activity levels refer to the func- platelet activation. tionally active portions of the pro- Xa The binding of fibrinogen to teins. Thus, elevated antigen levels platelet glycoprotein IIb/IIIa recep- of a particular molecule do not nec- Prothrombin Thrombin TAT tor leads to platelet aggregation. Fi- essarily reflect an increase in its func- PTF 1+ 2 brinogen is also the major determi- tional activity. For example, el- Platelets Platelet nant of blood and plasma viscosity, evated tPA antigen levels are often Activation explaining 50% of the latter. Hence, a reflection of high levels of circu- increased tendency of hemostasis lating PAI-1, resulting in a large por- Fibrin Fibrinogen Platelet and thrombosis may be reflected in tion of tPA antigens being bound to Aggregation high levels of plasma fibrinogen, PAI-1 and thereby rendering it in- FPA + B Plasma Viscosity FVII, factor VIII (FVIII), thrombin active.16 generation, platelet reactivity, and Many of these systemic throm- Plasmin Plasminogen high plasma viscosity. Increased bogenic factors may be involved in PAI-1 PAP α2AP thrombin generation may be indi- the initiation of early atheroscle- cated by high activation markers, rotic lesions and contribute to the PAI/tPA Complex such as prothrombin activation frag- progression of coronary thrombo- FDPs, Including D-Dimers tPA ment 1+2 (F1+2) and thrombin- sis, plaque growth, and its clinical

antithrombin complex (TAT), asso- sequelae. For example, plasma fi- Fibrinolytic System ciated with a decrease in clotting brinogen has been shown to stimu- time. late vascular smooth muscle migra- On the other hand, activation tion and proliferation, promote Summary of the coagulation and fibrinolysis pathways. TF indicates tissue factor; TAT, of the fibrinolytic system induces the platelet aggregation, and contrib- thrombin-antithrombin complex; PTF, 17 conversion of plasminogen to plas- ute to blood viscosity and thrombi. prothrombin fragments; FPA+B, fibrinopeptides min by plasminogen activators. Tis- Many of these indices, including fi- A and B; PAP, plasmin-antiplasmin complex; ␣ ␣ sue plasminogen activator (tPA) is brinogen, FVII, von Willebrand fac- 2AP, 2-antiplasmin; PAI-1, plasminogen activator inhibitor 1; tPA, tissue plasminogen the main fibrinolytic stimulator. tor (vWf, a marker of endothelial activator; and FDPs, fibrinogen degradation Plasmin promotes the degradation damage or dysfunction), D-dimer products. Solid arrowhead denotes activation.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2369

Downloaded From: https://jamanetwork.com/ on 09/23/2021 erothrombogenesis. The increase in extent, with a large sample size and and the lack of standardization in the sympathoadrenal activation may not adequate statistical power. Further- analytical methods used for the as- only trigger plaque rupture but also more, differences in aptitudes to- sessment of various hemostatic fac- directly induce a hypercoagulable ward motivation and compliance be- tors, particularly in the assessment state, leading to a rapid propaga- tween studied subjects are difficult to of platelet reactivity. tion of occlusive coronary throm- control and may confound the data. bus and, hence, sudden death. Typi- These may account for some of the Health vs Disease cally, an intravenous infusion of conflicting results seen among stud- epinephrine is used to mimic sym- ies that investigate the effects of ex- The available evidence from the in- pathetic activation and to examine ercise or dietary changes on throm- tervention or randomized con- the adrenergic effects on thrombo- bogenic factors. trolled trials would suggest that ex- genic markers.35,36 The literature ercise or physical training evokes suggests a dose-dependent stimu- EXERCISE AND multiple effects on blood hemosta- lation of FVIII clotting activity, THROMBOGENESIS sis in healthy individuals and in pa- vWf antigen, tPA activity, and tients with atherovascular disease. platelets within a 15- to 40-minute Many long-term epidemiologic stud- For example, patients with athero- infusion of norepinephrine. Al- ies have demonstrated an unequivo- vascular disease have higher basal though in healthy individuals the cal and strong relationship of in- levels of PAI-1 and lack a similar de- increase in coagulability may be creased fitness, exercise, or physical gree of increase in tPA activity after counteracted by a rapid rise in fi- activity during leisure time with re- exercise when compared with brinolytic activity, such hemostatic duced cardiovascular risk. Regular healthy subjects.64 In addition, balance between coagulation and exercise is known to lower body higher thrombin generation has been fibrinolysis may be impaired in weight and blood pressure and im- found in patients with peripheral subjects with atherosclerotic dis- prove lipid profile (with a decrease vascular disease in response to ex- ease or risk factors, and hence, cat- in serum cholesterol levels and an ercise, whereas no such increase was echolamine surge may trigger a hy- increase in high-density lipoprotein- detected in healthy controls.59 percoagulable state and enhance cholesterol [HDL-C] levels). In ad- the odds of overt thrombosis.37-39 dition, regular exercise enhances Effects of Short-term vs Regular The precise mechanisms under- functional capacity and psychologi- Exercise on Coagulation lying the hemostatic changes with cal well-being, as well as quality of and Fibrinolysis sympathetic activation remain un- life. The underlying biological clear. The lack of inhibition by aspi- mechanisms through which these While bearing in mind the consider- rin in the increase of platelet aggrega- beneficial effects are mediated must able inconsistency of the results of bility and platelet secretory activity be interrelated. various exercise studies due to meth- during norepinephrine infusion40 or There are many reports on the odologic variations, there are impor- exercise37 suggests that platelets are effects of exercise on coagulation, fi- tant differences between the effects of not being stimulated through the cy- brinolysis, and platelet activation. moderate endurance physical train- clooxygenase-dependent pathway. These mainly consist of interven- ing and short-term strenuous exer- Furthermore, exercise and mental tion studies, prospective random- cise on both the hemostatic and fi- stress induced in platelet-depen- ized controlled trials (Table 1), and brinolytic variables (Table 2). By and dent thrombin generation is sup- numerous large, population-based, large, regular physical activities of pressed by ␤-blocker therapy but cross-sectional studies. moderate intensity in training pro- not by aspirin, further support of the Several cross-sectional studies grams enhance blood fibrinolytic important role of sympathoadrenal have consistently shown a positive capacity and possibly also reduce activation.41 effect toward an antithrombotic blood coagulation, although the lat- It should be emphasized, how- state, especially in lowering plasma ter remains disputable. Conversely, ever, that lifestyle modifications rarely levels of fibrinogen and improving short-term strenuous exercise seems involve a single component; for ex- fibrinolytic capacity by long-term to induce a hypercoagulable state ample, an increase in exercise activ- regular exercise.62,63 However, it simultaneously with an increase in ity may accompany concomitant im- should be noted that most interven- fibrinolytic capacity as evidenced by provement in diet, which may in turn tion or randomized controlled trials increased levels of fibrinogen, FVIII lead to weight loss and better psycho- lack a comprehensive evaluation of coagulant, and platelet activities, logical well-being. In addition, such both the hemostatic and fibrino- higher thrombin generation and efforts may also modify other known lytic variables and thus provide only hemoconcentration, markedly in- independent cardiovascular risk fac- fragmentary data on the potential creased tPA activity, and possibly tors, such as lipid levels or hyperten- changes in hemostasis attributable also decreased PAI-1 and tPA anti- sion. It is therefore difficult to sepa- to physical exercise. Indeed, some gen levels. The rise in tPA activity rate the effects of these various factors studies have yielded conflicting data, is most apparent and seems to be in clinical studies, especially in ob- and this may be due to variations in directly proportional to the level of servational studies, although coinci- exercise protocol or training pro- exercise intensity.52,66 However, dental confounding variables can be grams used, populations studied the increased level of fibrinolytic statistically controlled, at least to some (age, sex, CHD), seasonal factors, activity seems to fall sharply dur-

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2370

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 1. Intervention or Controlled Randomized Clinical Trials of Exercise on Thrombogenic Factors

Source Subjects Exercise Main Results Summary Regular Exercise or Training Zanettini et al,42 14 Sedentary 12 Weeks of aerobic Exercise ↓ resting and 24-h BP, LVM (echo), and Fg; Regular exercise ↓ BP, LVM, and Fg, 1997 subjects with exercise; 8 resumed FVII unchanged; 2 mo after detraining, all indices improves coronary risk mild HT sedentary lifestyle and returned to baseline except LVM were reexamined 2 mo later El Sayed et al,43 25 Young subjects 12 Weeks of exercise (30 No significant differences in the ↑ of APTT, TCT, Maximum exercise transiently ↑ 1995 randomized: min, 3 times per week at FVIII antigen and activity, and tPA activity after coagulation and fibrinolysis. active exercise or 70% [6 wk] and 80% [6 exercise in both groups Physical conditioning appears not control wk] of maximum HR) to influence coagulation and fibrinolysis at rest or at max exercise El-Sayed,44 18 healthy subjects; 12 Weeks of No difference in resting tPA antigen and activity or High intensity exercise conditioning 1996 2 groups: high- preconditioned exercise: PAI antigen and activity; no significant change in significantly ↓ resting PAI activity. vs low-intensity exercise on BE for 20 tPA antigen and activity and PAI antigen after This may be a favorable effect of exercise min, 3 times per week at training; PAI activity ↓ significantly with exercise conditioning

80% or 30% V˙ O2max high-intensity exercise Van den Burg et 20 Young sedentary 12 Weeks of submaximal Posttraining, FVIII activity ↑, with APPT ↓ during Training promotes both coagulation al,45 1997 men vs 19 training maximum exercise. PAI antigen and activity and and fibrinolysis during exercise nontraining basal and exercise-induced tPA antigen were ↓, and may reverse unfavorable controls and tPA activity/antigen ratio ↑ in the training seasonal effects on fibrinolysis group. Controls had ↑ basal PAI antigen and activity, with ↑ basal and exercise-induced tPA antigen; basal and exercise-induced tPA activity were unchanged, but tPA activity/antigen ratio ↓ 46 Stratton et al, 10 Young (aged 6 Months of intensive Posttraining, V˙ O2max ↑ by 18% in the young group Intensive training ↑ resting tPA 1991 24-30 y) and 13 endurance exercise and by 22% in the older group. The older group activity and ↓ Fg and PAI activity older men (aged training had ↑ tPA activity by 39%, tPA in active form ↑ by in older men 60-82 y) 141%, PAI activity ↓ by 58%, and Fg ↓ by 13% posttraining. The young group had no significant changes in any of the measured indices Schuit et al,47 Elderly (aged 60-80 6 Months of intensive ↑ tPA activity and Fg in exercise but not in controls. Training ↑ fibrinolysis, but may 1997 y); active training program The ↓ in PAI antigen was significantly associated cause chronically ↑ plasma levels exercise or with ↓ in TGs and insulin. ↑ Fg coincided with ↑ of acute phase proteins in the controls CRP elderly Ponjee et al,48 20 Sedentary males 9 Months of intensity No significant change in FVIII activity, vWf, and TAT Exercise induces physical stress, 1993 and 15 sedentary training, 3-4 times per during training. In both groups, no change in Fg which has significant effects on females week; all ran a 15- and after 24 wk but ↑ before the 21-km race and still Fg, even at rest. In contrast to 21-km race after 24 and ↑ significantly 5 d later acute postexercise effects, 36 wk, respectively. regular exercise does not induce Blood drawn before a long-term activation of the training and 5 d before coagulation system and 5 d after both races Vaisanen et al,49 132 Males (aged 3 Years of regular Aerobic threshold ↑ by 8.8% in exercise but ↓ by PAI activity unchanged by regular 1999 52-62 y) low-to-moderate intensity 1.1% in controls. PAI activity unchanged in either exercise in the whole group over randomized to exercise group but 4G allele homozygotes in exercise group the 3 y but may ↑ in 4G allele exercise or had a 36% ↓ in PAI homozygotes control group Acute Exercise 50 Hegde et al, 10 Healthy men Ran at 70-75% V˙ O2max or Exercise ↑ FVIII activity by 66% and ↓ APPT and Exercise ↑ coagulation and 2001 walked at 1.2 mph for 30 remained the same on recovery. tPA activity and fibrinolysis but coagulation min. Blood drawn at rest, D-dimers ↑ after run. D-dimers remained ↑ but tPA sustained during a time when after exercise, and every ↓↓ at 1-h recovery fibrinolysis ↓, thus ↑ prothrombic 20 min for 1-h recovery risk in the first hour of recovery time Lin et al,51 1999 11 Moderately active Blood drawn at rest, Exercise ↑ FVIII activity, ↓ APPT. Exercise ↑ tPA ↑ Fibrinolysis during exercise seems young men immediately after, and 2, antigen and activity, ↑ total fibrin/Fg DPs, but ↓ PAI to counterbalance the ↑ in 6, and 24 h after BE activity. FVIII activity ↑ persisted 2 and 6 h into coagulation but this hemostatic recovery while fibrinolytic activity ↓ sharply balance is not maintained during recovery Rankinen et al,52 9 Healthy men (aged Maximum and 2 Baseline Fg, tPA, and PAI activity similar in each Acute exercise ↑ fibrinolysis, which 1995 23-37 y) randomized exercise. tPA activity ↑ after each exercise. PAI ↓ in normalized 24 h later. Fg submaximum (30 min at maximum and anaerobic exercise but not aerobic unchanged 50% [aerobic threshold] exercise. All 24 h postexercise activity was similar and 78% [anaerobic to baseline levels. Fg unchanged in any exercise

threshold] V˙ O2max)BE separated by 7 d

(continued)

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2371

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 1. Intervention or Controlled Randomized Clinical Trials of Exercise on Thrombogenic Factors (cont)

Source Subjects Exercise Main Results Summary Acute Exercise (cont) Prisco et al,53 12 Male marathon Marathon; blood drawn Immediately after the race, Fg ↓, but ↑ in PTF Persistence of coagulation and 1998 runners (aged day before, F1 + 2, TAT, and ECLT, and ↑ tPA and PAI antigen, fibrinolysis activation up to 24 h 35±7y) immediately after, D-dimer, and ↑↑ Fg DPs. All indices unchanged at after the end of the race and 24 and 48 h after 24 h, but returned to baseline at 48 h run Cerneca et al,54 7 Rowers, 12 Before and after Significant ↓ basal protein C in rowers that ↓ further Physical activity benefits the 1999 marathon near-maximum after exercise. Significantly ↑ basal ATIII, protein coagulation system, particularly runners, 7 exercise: rowing C, and protein S activity in runners vs rowers. A fibrinolysis, but certain subjects weightlifters, and machine, treadmill, high percentage of weightlifters had ↓ in tPA and may be at risk of thrombosis 7 healthy own exercise ↑ PAI after exercise. Controls showed ↑ controls equipment and BE, fibrinolytic activity and all anticoagulants after respectively exercise Watts et al,55 100 Athletes vs 25 10- To 26.2-mile race. No difference in baseline Fg, FVII, FVIII, or vWf A hypocoagulable rather than a 1991 nonexercise Blood drawn before between the 2 groups but athletes showed ↑ hypercoagulable state during controls and after races fibrinolytic activity and ↓ adrenaline-PAggr. running in athletes Immediately postrace, ↑ platelet count and FVIII clotting but no evidence of consumption or thrombin modification of FVIII clotting. Adrenaline-PAggr ↓ and fibrinolysis ↑ after the race 56 Wang et al, 23 Healthy men BE at 60% V˙ O2max for 30 PAggr and adhesiveness ↑ by short-term strenuous PAggr and adhesiveness may be ↓ 1995 (mean age, 21 y) min/d, 5 d/wk for 8 exercise in both groups, but ↓ after exercise in the by training but reversed to the randomized to wk, then trained subjects. Deconditioning reversed the pretraining state after control or deconditioned for 12 resting and postexercise effects to the pretraining deconditioning training group wk state Li et al,57 1999 15 Healthy men Exercise to exhaustion Exercise ↑ sP-sel expression and levels, ↑ platelet Exercise ↑ platelet and leukocyte with and without by BE, initial and leukocyte interaction, platelet-platelet and activation and Aggr in vivo and↑ 1 week of workload of 30 W platelet-leukocyte Aggr, and PTF F1 + 2. Aspirin responsiveness in vitro. Aspirin pretreatment with and increments of 10 had no effects on all these indices did not attenuate the prothrombic aspirin W/min effects of exercise Andreotti et al,58 27 CHD patients on Blood drawn at rest, Exercise induced myocardial ischemia in 14 patients. Mild exercise transiently ↑ PAggr in 2001 aspirin vs 12 immediately after, PAggr ↑ in all CHD patients at peak exercise but patients with CHD. The effect is healthy controls and 0.5 and 3 h after return to baseline at 3 hr. No changes in controls. independent of myocardial mild exercise vWf was similar in both groups ischemia, occurs despite aspirin (Յstage III modified Bruce) Mustonen et 15 PVD patients vs Blood drawn before and Exercise ↑ tPA antigen and activity, D-dimer, PAP, Sudden catecholamine release and al,59 1998 15 healthy after submaximum and catecholamines in both groups but higher local ischemia during exercise controls treadmill test levels in patients, both at rest and after exercise. may ↑ the preexisting TAT ↑ in patients but not in controls after prothrombic potential of the exercise. PAI antigen unchanged atherosclerotic vessel wall Li-Saw-Hee et 20 Patients with Blood drawn before, Basal vWf and Fg ↑ in AF. Significant ↑ in Fg and ↓ Acute exercise induced a al,60 2001 chronic AF vs 2 after, and at 20 min in PAI after exercise but no changes in vWf or hypercoagulable state in chronic groups of post−standard Bruce sP-sel in AF patients. All indices unchanged in AF, and possible ↑ in fibrinolysis. matched controls exercise controls, despite longer duration and greater No significant effect on ET dysfn in SR workload of exercise or platelet activation Gibbs et al,61 20 Patients with Blood drawn before, Basal vWf and sP-sel ↑ in CHF. Significant ↑ in Acute exercise induced a 2001 stable CHF vs 2 after, and at 20 min plasma viscosity, Fg, and hematocrit after hypercoagulable state in CHF. groups of post−standard Bruce exercise and positive correlation between exercise Moderate exercise should be matched controls exercise workload and plasma viscosity encouraged in CHF patients but vigorous exercise should be avoided

Abbreviations: ACA, arachidonic acid; ADP, adenosine diphosphate; Aggr, aggregation; ALA, ␣-linolenic acid; ␣2AP, ␣2-antiplasmin; APTT, activated prothrombin time; AT, antithrombin; ␤-Tbg, ␤-thromboglobulin; BE, bicycle ergometer; BMI, body mass index; BP, blood pressure; BT, bleeding time; CABG, coronary artery bypass grafting; CECs, circulating endothelial cells; CHD, coronary heart disease; CRP, C-reactive protein; DBR, double-blind randomized; DDAVP, 1-desamino-8-D-arginine vasopressin; DHA, docosahexaenoic acid; DPs, degradation products; ECLT, euglobulin lysis time; EPA, eicosapentaenoic acid; ET dysfn, endothelial dysfunction; F1 + 2, fragments1+2;Fg,fibrinogen; FMD, flow-mediated vasodilation; FPA, fibrinopeptide A; FVII, factor VII; FVIII, factor VIII; HDL, high-density lipoprotein; HR, heart rate; HT, hypertension; HUVECs, human umbilical vein endothelial cells; IGT, impaired glucose tolerance; IMT, carotid intimal media thickness; IVUS, intravascular ultrasound; LAD, left anterior descending coronary artery; LDL, low-density lipoprotein; LV, left ventricle; LVM, LV mass; MI, myocardial infarction; n-3 and n-6, omega-3 and

omega-6 polyunsaturated fatty acid; PAggr, platelet aggregation; PAI, plasminogen activator inhibitor; PAP, plasmin ␣2-antiplasmin; PF, platelet factor; PHC, platelet hemostasis capacity; PL, phospholipid; PTF, prothrombin fragment; RBCs, red blood cells; rehab, rehabilitation; SNP, sodium nitroprusside; sP-sel, soluble P-selectin; suppl, supplementation; TAT, thrombin–antithrombin complex; Tbx, thromboxane; TCT, thrombin clotting time; TGs, triglycerides; tPA, tissue plasminogen activator; vWf, von Willebrand factor; WBCLT, whole blood clot lysis time; WHR, waist-hip ratio; ↓, decrease; ↑, increase.

ing the recovery period, whereas This phenomenon has been den cardiac death in susceptible sed- activation of the coagulation cas- thought to possibly precipitate acute entary individuals or patients with cade is persistent.50,51 coronary thrombosis, leading to sud- preexisting atherovascular disease

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2372

Downloaded From: https://jamanetwork.com/ on 09/23/2021 who may not sustain their fibrino- by a decrease in PAI-1 levels and also lytic capacity (perhaps due to endo- a rise in tPA activity. Indeed, Szy- Table 2. Summary of the Effects of thelial dysfunction) when they are manski et al73 demonstrated that per- Exercise on Thrombogenic Factors* exposed to unaccustomed, short- sons who are habitually active have Physical term, strenuous physical exertion. the lowest basal PAI-1 activity but Exercise However, recent studies have sug- the highest increase of tPA activity Thrombogenic gested that functional fibrinolytic ac- in response to exercise when com- Markers Regular Acute tivity was similar in physically ac- pared with inactive subjects. This Fibrinogen ↓↑ tive men with and without a history effect has been repeatedly demon- Factor VII ↓↔ of MI67 and in older men with hy- strated using various exercise inten- Plasma viscosity ↓↑ 46,47,72,74-76 tPA ↑↑† pertension when compared with sity and duration. One ↓↓ 68 77 PAI-1 normotensive subjects. study reported lower PAI-1 levels Platelet activation ↓↑ Although the results of the in those participating in regular Fibrinopeptide A Unknown ↑ studies that investigated the effects sporting activities than the respec- Thrombin generation‡ Unknown ↑ of short-term exercise on fibrino- tive age-matched sedentary indi- lytic markers are more consistent, viduals or elderly athletes and Abbreviations: PAI-1, plasminogen activator studies that investigated plasma post-MI patients, but tPA activities 1; tPA, tissue plasminogen activator. *Adapted from Imhof and Koenig.65 fibrinogen concentration have pro- were significantly higher after exer- †In healthy subjects. duced conflicting data. Several cise in those with lower pretest PAI-1 ‡A rise in thrombin generation is indicated by studies have reported a significant level. elevated levels of thrombin−antithrombin III complex and prothrombin fragments1+2. increase in plasma fibrinogen after strenuous exercise,54,69,70 but others Endurance Exercise or Physical using different protocols have Training and Coagulation shown either no significant gen levels in both males and fe- effects52,71,72 or even a reduction in Most exercise studies43,50,51,69,78-80 of males after 24 weeks of training. plasma fibrinogen level after short- varying degrees of intensity and du- Only a few reports80,83,86-88 on term, intense physical exercise.53 It ration have been found to induce a the effects of exercise on FVII are is possible that the changes in significant increase in FVIII coagu- available and, again, with mixed re- coagulation markers depend on the lant activity. However, others have sults. Moderate exercise has no sig- type of physical exercise to which demonstrated that regular training nificant influence on FVII or at the subjects are subjected. Cerneca et exercise does not seem to induce sig- most the effect is only relatively al54 demonstrated that this might nificant effect on resting or postex- short-lived. Studies89-93 on the ef- be the case, since rowers, mara- ercise levels of FVIII activity and fects of exercise on fibrinopeptide A thon runners, and healthy controls antigen in normal healthy and sed- (a marker of thrombin activity and revealed a significant increase in entary subjects,45,48,55,81 although 4 fibrin formation) have again pro- plasma fibrinogen levels after near- weeks of physical training has been duced conflicting results, although maximum exercise tests, whereas shown to lower resting levels of raised plasma markers of thrombin weightlifters showed no significant FVIII activity and antigen in post-MI generation (TAT and F 1+2) with change. patients.82 Although high levels of short-term exercise had been re- Interestingly, genetic factors plasma fibrinogen are usually found ported. The overall mechanisms un- might also explain the different ef- in patients with CHD or cardiovas- derlying these changes in coagula- fects of exercise on hemostatic or fi- cular risk factors, randomized con- tion or fibrinolysis, in response to brinolytic factors. For example, Mont- trolled trials have found mixed re- short-term or long-term endurance gomery et al70 investigated the effects sults in fibrinogen levels in response exercise, are poorly understood and of long-term physical training and to regular physical activity.46,76,83-85 still remain speculative, but inter- short-term, intensive exercise on For example, 12 weeks of aerobic ex- actions involving neurohormonal plasma fibrinogen levels in 156 men ercise training in sedentary hyper- pathways are very likely.59,94-96 in the British Army and found that tensive subjects lowered blood pres- subjects carrying the A allele of the sure, left ventricular mass, and Exercise Effects G453A polymorphism in the ␤-fi- plasma fibrinogen levels, but changes on Platelet Reactivity brinogen gene showed a higher in- returned to baseline values (except crease in plasma fibrinogen than men left ventricular mass) at 2 months af- The effects of exercise on platelet ag- with the GG genotype. ter detraining.42 Similarly, 6 months gregation and activation have been of intensive endurance training re- extensively studied, but the results Endurance Exercise or Physical duced plasma fibrinogen in elderly are highly variable.97,98 It is notewor- Training and Fibrinolysis men but not in young men.46 On the thy that measurements of platelet re- other hand, Schuit et al47 reported activity either in vitro or ex vivo ag- Various studies have consistently re- a significant increase in plasma fi- gregability assays or in vivo platelet ported a significant improvement in brinogen levels in elderly males af- secretory products (mainly ␤-throm- fibrinolytic capacity following regu- ter a same duration of intensive boglobulin and platelet factor 4) are lar exercise or physical training. The training, whereas Ponjee et al48 re- associated with considerable meth- increase in fibrinolysis is indicated ported no change in plasma fibrino- odologic difficulties and thus may

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2373

Downloaded From: https://jamanetwork.com/ on 09/23/2021 account for the discrepancies of re- tion are associated with a prothrom- reduces PAI-1 and tPA antigen lev- sults reported in the literature. bic or hypercoagulable state. We els,33,112 suggesting a causal rela- By and large, short-term, demonstrated that short-term exertion (Table 3). strenuous exercise induces a tran- cise to exhaustion significantly in- Recent evidence has shown that sient increase in agonist-induced creased plasma fibrinogen and lower elevated plasma PAI-1 activity seen platelet aggregation both in vitro and PAI-1 levels but had no influence on in obese individuals may be caused ex vivo and an increase in platelet vWf or soluble P-selectin levels in by increased PAI-1 release from vis- counts, adhesiveness, and in vivo patients with atrial fibrillation when ceral adipose tissue.122,123 However, platelet secretory activity. Overall, compared with age- and sex- a liposuction procedure that re- these effects seem to be more pro- matched patients in sinus rhythm.60 moves visceral adipose tissue and nounced in sedentary than healthy In another similar exercise study in achieves a weight reduction of 5% subjects.99,100 In contrast, long- patients with stable congestive heart after 3 months without change in term endurance physical training failure, we also found that plasma lifestyle does not seem to signifi- (preconditioned) in men and women viscosity, fibrinogen, and hemato- cantly reduce plasma levels of vWf, at moderate intensity (50%-55% crit levels were significantly in- fibrinogen, or PAI-1.117 In contrast, of peak oxygen consumption) seems creased, both immediately after ex- surgical removal of adipose tissue in to suppress platelet adhesiveness ercise and at 20 minutes into the 19 patients with morbid obesity with and aggregation both at rest and af- recovery period.61 a mean body weight reduction of 50 ter short-term, strenuous exercise. kg at 6 months and 64 kg at 12 However, the effects reversed back WEIGHT REDUCTION months has led to a significant re- to the pretraining state after a pe- duction in FVII, fibrinogen, and riod of deconditioning and hence the Overweight and obesity, assessed ei- PAI-1 activity and a slight increase importance of regular moderate ex- ther by body mass index (BMI), a in tPA activity. Therefore, it seems ercise to maintain such potential measure of weight in kilograms di- that a large amount of adipose tis- benefits.56,101 vided by the square of height in me- sue may need to be removed artifi- Thus, as with fibrinolytic re- ters, or waist-to-hip circumference cially before an improvement of he- sponse, platelet reactivity in response ratio (WHR), are associated with in- mostatic and fibrinolytic profiles to exercise also seems to be both creased cardiovascular morbidity could be detected, or it might be that duration and intensity dependent. and mortality.104,105 Indeed, there is a change in lifestyle, including in- However,theunderlyingmechanisms increasing evidence that moderate creased exercise and dietary con- remain unclear. Increases in catechol- weight loss could result in regres- trol leading to weight reduction, is amineconcentrationsandshearstress sion of coronary arterial lesions and a prerequisite for improvement in are probably important.102,103 Perhaps significantly reduces cardiac events coagulation and fibrinolysis. The lat- the short-term response may be re- and total mortality.106 ter seems more likely the case, since lated to the release of tPA from en- Most of the present data on there is evidence that limited weight dothelial cells associated with higher thrombogenic profile in over- loss (Ͻ7 kg) by lifestyle modifica- catecholamine release during exer- weight or obese persons relate to tions alone could lead to a reduc- cise. Interestingly, the fact that aspi- PAI-1 and tPA antigens. For ex- tion of hemostatic factors, FVII, and rin treatment had no significant in- ample, both BMI and WHR corre- PAI-1 levels, while increasing the fluence on platelet activation induced late strongly and positively with he- tPA activity.106 by heavy exercise in patients with mostatic factors but negatively with Indeed, obesity and syndrome of stable angina pectoris and matched fibrinolytic activity.107 It has been insulin resistance are inextricably healthy controls suggests that the re- shown that women with high WHR linked with hypertriglyceridemia, hy- sponse may not be cyclooxygenase- have significantly higher fibrino- perinsulinemia, hypo-HDL-choles- pathway dependent.37,38 This im- gen and PAI-1 levels compared with terolemia, glucose intolerance, and plies that aspirin may have a limited obese women with a low WHR or hypertension. It is known that both antithrombotic effect during physi- with lean women.108 Similarly, af- triglyceride and insulin resistance cor- cal exercise and probably also in ter adjusting for other lifestyle vari- relate strongly and positively with other situations with increased cat- ables, obese men (BMI Ͼ30) had PAI-1.124-126 In fact, dietary interven- echolamine levels such as during 50% higher PAI-1 activity and 30% tion with a low-saturated-fat diet127 or acute psychological stress. The higher tPA antigen when com- gemfibrozil treatment128 that lowers chronic platelet response, however, pared with men of “ideal” BMI serum triglyceride levels has been may be related to nitric oxide release (Ͻ25).107 In addition, high fibrino- accompanied by improvement and as a consequence of regular low-to- gen and plasma viscosity have also even normalization of the fibrino- moderate exercise training.101 been found to be associated with in- lytic activity. Thus, it is plausible that creasing BMI,107,109,110 although, overweight reduction improves fibrino- Effects of Exercise in Specific all, there is only little evidence that lytic capacity via modifications in Patient Groups on Coagulation weight reduction reduces plasma fi- both lipids and insulin resistance and Fibrinolysis brinogen or viscosity levels.106,111 On profiles. However, data on the in- the other hand, there is plenty of evi- teractions between physical activity There is increasing evidence that pa- dence that weight reduction by regu- and diet and hemostasis are scarce, tients with chronic atrial fibrilla- lar exercise and dietary changes and it is likely that moderation in both

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2374

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 3. Weight Reduction and Thrombogenic Factors

Source Subjects Design Main Results Conclusion Rissanen et 51 Obese women DBR, 12-mo weight ↓ trial of Weight ↓ similar in orlistat and placebo groups. Maintenance of modest al,113 2001 (mean age, 44 y; placebo and orlistat, plus Orlistat did not influence coagulation factors weight ↓ is associated BMI, 36 kg/m2) hypoenergetic diet adjusted beyond its effect on weight ↓. PAI and FVII ↓ in with ↓ in PAI and FVII in for actual body weight at 6 first 3 mo and correlated with ↓ weight. Between obese women. Change of mo. Measurements at 3- to 6 and 12 mo, changes of PAI and FVII paralleled serum insulin is 6-mo intervals changes of weight; activities ↑ with weight associated with changes rebound but remained below the 6-mo values if of PAI. Fg is not affected weight ↓ continued. Serum insulin correlated with by modest weight ↓ PAI and FVII at 6 mo and with PAI at 12 mo. No changes in Fg at any time Marckmann et 36 Obese patients in Randomized: very-low-energy Mean weight ↓ of 13.6 kg. After 24-wk weight Major weight ↓ is associated al,114 1998 a 2-stranded diet (8 wk) or low-energy diet maintenance, PAI antigen ↓ by 34%; TGs, 30%; with sustained and marked randomized (17 wk), both given anorectic FVIIc, 12%; cholesterol, 9%; Fg, 6%; but HDL-C ↑ improvements in lipids intervention study compound, then by 5%. All changes highly significant. No and coagulation profile, rerandomized to 24-wk differences between slimming or maintenance irrespective of the tested maintenance diets regimens slimming and maintenance regimens Folsom et al,115 90 Men and 88 Randomized to 1 of 4 weight Treatment ↓ weight 9.4 kg in men and 7.4 kg in Weight ↓ can improve 1993 women; loss treatment groups or women and significantly ↓ PAI, tPA antigen, and abnormalities in moderately controls. Measurements at FVII. ↓ In these variables correlated with the ↓ in hemostatic factors overweight baseline and 6 mo weight and TGs. No change in D-dimer, Fg, or associated with obesity protein C with weight ↓ Primrose et 19 Patients Blood drawn before and at 6 Surgery ↓ mean weight of 50 kg at 6 mo and 64 kg Surgical treatment of morbid al,116 1992 underwent and 12 mo after operation at 12 mo. At 12 mo, significant ↓ in TC, FVII, Fg, obesity may have a surgery for and PAI. tPA activity ↑ slightly. No changes in long-term beneficial effect obesity APPT, FVIII, vWf, ␣2AP, TAT, protein C, ␤-Tbg, PF on mortality from 4, FPA, or platelet count atherovascular disease Berntorp et 53 Patients with Blood drawn preoperatively, 2 Weight ↓ of 4 kg sustained during follow-up. Slight Surgical removal of adipose al,117 1998 Dercum disease and 4 wk, and 3 mo ↑ in coagulation factors 2 and 4 wk tissue, without change in postoperatively postoperatively. At 3 mo the values ↓ to lifestyle, does not seem to preoperative levels except for PAI, which still improve coagulation and slightly ↑ fibrinolysis Sudi et al,118 20 Obese boys and Blood drawn before and after 3 Estimates of adiposity, insulin, and TGs correlated Fibrinolytic indices 2001 40 obese girls wk of low-caloric diet and with PAI and tPA antigen. WHR correlated with associated with body (mean age, 12 y) exercise. Body composition fibrinolytic indices only in girls. Insulin and tPA mass ↓ but can occur assessed by bioelectrical antigen contributed to PAI, whereas percent fat independently of a impedance mass and TGs contributed to tPA antigen. Weight concomitant ↓ in fatness. loss ↓ adiposity, abdominal adiposity, fibrinolytic, Initial PAI and tPA antigen and metabolic indices. Initial PAI and changes in predict changes of body mass contributed to ↓ in PAI. Initial tPA fibrinolytic indices antigen contributed to changes in tPA antigen Lindahl et al,76 186 Obese subjects Randomized to active program Active group ↓ weight (5.4 kg vs 0.5 kg), ↑ oxygen Intense lifestyle program has 1999 with IGT 1-y intensified dietary and consumption by 10% (↓ by 1% in controls), ↓ sustained beneficial exercise or usual-care PAI (31% vs 12%), and ↓ tPA antigen (14% vs effects on fibrinolysis controls 6%) Mavri et al,119 52 Healthy, Weight ↓ program with a At end of program, PAI antigen and activity ↓. Leptin Weight loss with hypocaloric 1999 premenopausal, hypocaloric diet. Blood drawn ↓ but no change in adipsin. PAI associated with diet ↓ PAI, which is more obese women (33 at entry, 1 wk, end of BMI, body fat, leptin, and insulin. At 5 mo, PAI closely related to changes completed the program, and 5 mo follow-up remained ↓ in 14 women who maintained weight in adipose tissue than to study) ↓ but ↑ in 16 women who regained weight. ↑ In changes in metabolic PAI correlated with ↑ in body fat and leptin. BMI variables, suggesting a was the major determinant of PAI level significant role for adipose tissue in regulating plasma PAI Charles et al,120 324 Nondiabetic DBR 1 y: placebo or metformin PAI activity and antigen ↓ significantly but similarly Weight ↓ associated with ↓ 1998 patients with groups, in addition to diet in both groups. This ↓ mainly in subjects who lost in PAI. Metformin may central obesity and exercise weight. tPA antigen and vWf ↓ significantly more have effect on production (aged 35-65 y; in the metformin group or metabolism of tPA mean BMI, 32.5 antigen and vWf kg/m2) Calles- 19 Elderly obese 11 Weeks of energy-restricted Initially elevated PAI ↓ by 50%, with a ↓ in tPA/PAI Energy restriction induces Escandon et subjects (aged diet by 3700-4600 kJ/d deficit complexes but no change in tPA. PAP complex ↑ moderate weight ↓ and al,121 1996 60-70 y; BMI Ͼ32 by 20%. The ↓ in PAI and the ↑ in PAP complexes leads to diminution of ↑ kg/m2) correlated with weight and fat mass ↓s. No plasma PAI and relief of correlation between fibrinolytic variables and inhibition of fibrinolysis in baseline substrates or insulin, but change in PAI elderly, obese subjects correlated with change in plasma triacylglycerols

Abbreviations: For an explanation of abbreviations, see footnote to Table 1.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2375

Downloaded From: https://jamanetwork.com/ on 09/23/2021 efforts would yield a more powerful lipid constituents in these diets Intervention Studies impact on coagulation and fibrinoly- contribute to coronary risk is un- sis systems than either lifestyle modi- known. The rapidity of onset of the There are a large number of inter- fication alone. Clearly, more studies beneficial effects seen in these stud- vention studies on the effect of n-3 are needed to dissect such complex ies suggests that the diet might have PUFA in the form of fish oil cap- interactions. anti-inflammatory, antithrom- sules, fishmeals, or its precursor botic, and even membrane stabiliz- ␣-linolenic acid on various hemo- DIETARY LIPIDS ON ing and hence antiarrhythmic ef- static factors (Table 4). We dis- THROMBOGENESIS fects besides lowering the rate of cuss its effects on coagulation, fibri- progression of atherosclerosis. In- nolysis, and platelet reactivity. Outcome Studies With n-3 PUFA deed, the effects of dietary manipu- lations or supplementation with Coagulation and Rheology The role of dietary changes in modi- individual or complex dietary lip- fying CHD risk has been well estab- ids on thrombogenic variables have The effect of n-3 PUFA on fibrino- lished.129 The recent Lyon Diet Heart attracted considerable interest with gen and blood rheology has been ex- Study130 reported a 50% to 70% a particular emphasis on n-3 PUFA, tensively studied. Reductions in fi- lower risk of recurrent heart dis- although the interplay between brinogen (the largest contributor to ease as measured by different com- these lipid constituents and the co- plasma viscosity) and increased binations of outcome measures, in- agulation system remains largely erythrocyte flexibility (a major com- cluding cardiac death and nonfatal unclear. ponent of whole-blood viscosity) myocardial infarctions in survivors would be desirable for vascular ben- of first-MI patients who received a Epidemiologic Studies efit. By and large, however, many Mediterranean diet (with more fish, studies have shown no or little im- more fiber, but less fat) supple- Thus far, only indirect evidence links provement in fibrinogen levels after mented with the precursor of n-3 dietary saturated fatty acids with en- giving n-3 PUFA supplements to dif- PUFA, ␣-linolenic acid (18:3n-3, de- hanced thrombogenesis in hu- ferent types of patients.140,147,157,158,161- 132,133 164 rived mainly from vegetable or seed mans. When unsaturated fatty 163 One study has reported an in- oil) when compared with controls acids of the n-9, n-6, or n-3 fami- crease in erythrocyte flexibility but an who received usual care. The study lies replace saturated fatty acids in unaltered fibrinogen level. Indeed, is in parallel with the results of other the diet of experimental animals, the n-3 PUFA has been linked to im- secondary prevention dietary trials, development of atherothrombosis provement in erythrocyte flexibil- namely, the DART and GISSI Pre- was inhibited, but the doses supple- ity with lower whole-blood viscos- venzione trials,12,13 which similarly mented tended to be much higher ity in few studies,156,165,166 although used a diet with low intake of total than in human clinical stud- others have yielded little effects.167- 134,135 and saturated fats and/or increased ies. Data from the Coronary Ar- 169 Simply measuring erythrocyte in- intake of oily fish rich in n-3 PUFA. tery Risk Development in Young filtration might be an inadequate Indeed, at least 2 servings of fish per Adults136 study showed that usual in- method for detecting small but sig- week, especially fatty fish, equiva- take of fish or dietary supplemen- nificant differences in erythrocyte lent to an intake of n-3 PUFA ap- tation with ␣-linolenic acid, EPA, flexibility, and this may account for proaching 1 g/d have been recom- and DHA was not associated with the inconsistency of results among mended by the American Heart levels of FVIII, fibrinogen, or vWf studies. Association.131 and, hence, suggests that usual cus- There is also little agreement on The experimental group in the tomary intakes of fish and n-3 PUFA the effects of n-3 PUFA on clotting Lyon Diet Heart Study had higher in populations that generally do not factors, such as FVII or FVIII. One plasma levels of oleic acid, ␣-linole- consume large amounts of these food author170 has reported reduced lev- nic acid, and eicosapentaenoic acid items are not associated with these els of FVIII with n-3 supplements, but (EPA, 20:5n-3). In the GISSI Preven- hemostatic factors. Similar results most authors145,171-176 have found ei- zion trial, patients received daily were also found in the PRIME (Pro- ther no influence on or an increase doses of n-3 PUFA as 1 gelatin cap- spective Epidemiological Study of in FVII or FVIII by fish oil ingestion sule containing EPA and docosa- Myocardial Infarction) sub- or n-3 PUFA diet. Studies on the in- hexaenoic acid (DHA, 22:6n-3) as study,137 which showed no relation- take of n-3 PUFA and vWf concen- ethyl esters. In the DART study, at ships between fatty acids and fibrino- tration have also been similarly con- least 3 servings of fatty fish or ap- gen, vWf, PAI-1, or FVII levels. By flicting. Most studies136,171,172,174,177,178 proximately 15 fish oil capsules per contrast, results from another cross- have not been able to show an effect week led to a significant 29% reduc- sectional study138 suggest that in- with these fatty acids. In addition, tion in both cardiac and total mor- creases in dietary n-3 PUFA intake few studies have even suggested tality within 4 months. The low in- from fish is negatively associated that n-3 PUFA, including ␣-linole- cidence of cardiovascular disease with fibrinogen, FVIII, and vWf and nic acid, may have antithrombotic among Greenland Eskimos and coast positively associated with protein C effects by enhancing protein C ac- land Japanese has been related to high levels. Such differences have been at- tivity,171,174 increasing tissue factor intake of the marine n-3 PUFA: EPA tributed to higher EPA and DHA in- pathway inhibitor,175 or reducing and DHA. However, how particular takes in the latter study. the expression of procoagulant tis-

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2376

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 4. Intervention and Cross-sectional Studies of n-3-Polyunsaturated Fatty Acids on Thrombogenic Markers

Source Subjects Design Summary Shahar et al,138 15 000 Subjects of the Cross-sectional analysis adjusted for sex, Dietary intake of n-3 negatively associated with Fg, FVIII, and 1993 ARIC Study race, age, BMI, smoking, alcohol, vWF (black and white patients) and positively associated diabetes, and field center. Usual dietary with protein C (white patients only). ↑ In n-3 intake from intake assessed by a food frequency fish may modify several coagulation factors questionnaire Scarabin et 283 Subjects of the Cross-sectional analysis adjusted for age, Only tPA antigen inversely associated with marine n-3. FVII, al,137 2001 PRIME study center, and BMI Fg, PAI, D-dimer, and vWf were not associated with n-3. D-dimer was positively associated with ACA and eicosamonoenoic acid. Marine n-3 may favorably influence tPA antigen Iacoviello et 6 Healthy volunteers DBR, cross-over study: 5.3 g EPA and DHA Aspirin plus n-3 ↓ tPA antigen and the fibrinolytic response al,139 1992 (aged 24-37 y) or n-6 (control) daily for 29 d. Aspirin (40 to venous occlusion in all subjects. PAI activity before mg/d) then added for another 14 d; 2-mo stasis ↑ by n-3 suppl, but not affected by aspirin washout before cross-over Oosthuizen et 20 Healthy young DBR, cross-over study: 6 g fish oil or olive Fish oil independently ↓ TGs, FVc, and FVIIc. Both groups al,140 1994 volunteers oil (placebo) daily suppl for 6 wk significantly ↑ PAI and ↓ FXc and Fg in women, who had higher initial levels than men Boberg et al,141 14 Diabetic patients vs DBR, cross-over study: 10 g maximum EPA PAI activity ↑ in diabetic patients compared with controls. 1992 healthy controls (3 g n-3) or olive oil (placebo) daily for 2 Despite a ↓ in TG and unchanged insulin levels, there was consecutive 8-wk periods a significant ↑ in PAI activity after maximum EPA. In diabetic patients given n-3 suppl, PAI activity ↑ though TG ↓ Radack et al,142 10 Patients with DBR, cross-over study: n-3 fish oil vs n-6 Fg ↓ in both groups. No significant changes in tPA activity, 1990 hyperlipoproteinemia corn oil suppl PAI, protein C antigen, ATIII activity, BT, and platelet type IIb or IV counts Prisco et al,143 20 Normolipemic DBR study: n-3 (4-g capsules) or placebo No significant changes in plasma Fg, PAI antigen or activity, 1994 healthy men (aged (4-g olive oil capsules) daily for 4 mo PT F1 + 2, lipids, and Lp(a) in both groups though there 27-41 y) was trend of ↓ in Fg, TG, and Lp(a) with n-3 during treatment and wash-out Myrup et al,144 29 Insulin-dependent DBR study: fish oil (4.6 g n-3) or placebo No change in n-3 platelet lipids, transcapillary escape rate of 2001 diabetic patients with (olive oil) daily suppl for 1 y albumin, PT F1 + 2, TAT, fibrinolytic indices, Fg, FVII nephropathy antigen and activity, thrombomodulin, vWf, PF 4, and ␤-Tbg after1yoffish oil suppl compared with olive oil Toft et al,145 78 Untreated DBR study:4gEPAandDHAorcorn oil PAI activity changed similarly in fish oil and corn oil groups, 1997 hypertensive patients (placebo) daily suppl for 16 wk as did tPA, FVIIc, and platelet count. Fg levels ↑ with fish oil and corn oil suppl Hansen et al,146 224 Healthy men (aged DBR study: EPA or DHA, or corn oil The ↑ in PAI activity was not different between groups. No 2000 36-56 y) (placebo) daily suppl for 7 wk correlation between change in TGs or PL n-3 and PAI activity. PAI associated with BMI, apoB100, TGs, and n-6 but not with n-3. Only 21% of the variation in PAI activity is attributable to these variables Nilsen et al,147 20 Accepted for CABG DBR study: n-3 (3.15 g EPA and 1.89 g No significant changes in BT, collagen-PAggr, and TxB2 1991 DHA) or corn oil (controls) daily for 5-6 production. Indices of extrinsic coagulation, including mo. Surgery performed mid-intervention phospholipase C–sensitive FVII and extrinsic pathway inhibitor, unchanged in both groups. TG with n-3 ↓ and Fg ↓ in controls Hellsten et al,148 40 Healthy subjects DBR, parallel trial: 6 g cod fish liver oil (2 g PAI unchanged with fish oil but ↓ with corn oil. tPA activity 1993 n-3) or 6 g corn oil daily suppl for 5 mo and mass unchanged in both groups Rogers et al,149 60 Male volunteers Randomized: maximum EPA fish oil or olive Fish oil ↓ TG by 54% and ↓ diastolic BP by 7%. BT ↑ by 1987 oil daily suppl for 3-6 wk 12%, but not significant. Heparin thrombin clotting time ↑ by 14% but thrombin time, Fg, or PF 4 unchanged. Fish oil also ↓ RBC pore transit time by 23%, but not significantly. No differences between the 2 groups in TC, HDL, blood counts, or PAggr

(continued)

sue factor activity on monocyte ex nolysis, others had found no change (EPA, docosapentaenoic acid, and vivo.179 in PAI-1 level or in tPA antigen after DHA) but not with n-3 PUFA from dietary intervention,145,148,180 and few vegetable origin (␣-linolenic acid).137 Fibrinolysis even reported significant increased in This is in agreement with a large in- PAI-1 activity.146,160,181,182 Moreover, tervention study184 in diabetic sub- The data available on the effects of n-3 data available on tPA activity also jects that showed a decrease in tPA an- PUFA on fibrinolytic activity are also seem contradictory.145,148,176,183 How- tigen after fish supplementation but inconsistent. Since the initial study by ever, a study has reported that tPA an- no effect on PAI-1 activity with high Barcelli et al,153 which suggested that tigen level was inversely related to n-3 ␣-linolenic acid diet intake in a n-3 PUFA may enhance plasma fibri- PUFA derived mostly from fish oil double-blind intervention trial.185

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2377

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 4. Intervention and Cross-sectional Studies of n-3-Polyunsaturated Fatty Acids on Thrombogenic Markers (cont)

Source Subjects Design Summary Li et al,150 1990 8 Healthy subjects Randomized: fish oil (6 g EPA) or vegetable In fish oil group, ACA, ADP, and collagen-PAggr ↑ but oil daily suppl for 25 d platelet adhesion to Fg and collagen I at low shear rates ↓ by 60%-65%. No changes in platelet adhesiveness in 5 subjects who had vegetable oil suppl Andrioli et al,151 60 Healthy volunteers Randomized into 3 groups: 20 mL fish oil Fish oil significantly ↓ stimulated adhesion with ADP and 1999 (0.3 g n-6; 3.6 g n-3; n-6/n-3 ratio 0.1) or thrombin. Soy lecithin ↑ platelet adhesion in all test 25 g soy lecithin (1.5 g n-6; 0.5 g n-3; conditions with ADP and thrombin. No changes in n-6/n-3 ratio 3) daily suppl or usual diet controls. Platelet adhesion correlated to changes in the (control) for 15 d platelet n-6/n-3 ratio caused by the different suppl. Fish oil rich in n-3 inhibit stimulated platelet adhesiveness and soy lecithin rich in n-6 ↑ adhesion in all test conditions. n-6/n-3 Ratio is a determinant of platelet adhesion Freese et al,152 29 Female and 17 male Randomized: fish oil plus sunflower oil (EPA No differences between the 2 groups in collagen-PAggr and 1997 healthy subjects plus DHA: 5.2 g/d) or linseed oil (ALA: 5.9 Tbx production, Aggr to the TbxA2 mimic I-BOP, urinary (aged 20-44 y) g/d) daily for 4 wk excretion of 11-dehydro-TbxB2, and ␤-Tbg, BT, Fg, ATIII activity, FVIIc activity, or PAI activity. Suppl of ALA from vegetable oil and EPA plus DHA from a marine source have similar effects on hemostatic factors Barcelli et al,153 9 Healthy subjects Maximum EPA fish oil (5 g n-3) daily suppl Vascular plasminogen activator ↑. Inhibitors of vascular 1985 for2wk plasminogen activator, of plasmin, and of PAP ↓.No significant changes in TCl, TGs, HDL, or LDL levels Mehta et al,154 8 Patients with CHD and Maximum EPA fish oil daily suppl for 4 wk In both groups, TG and PAI ↓ significantly but tPA antigen 1988 4 healthy subjects was unaltered. The magnitude of ↓ in TGs was dependent on baseline TG. PAI ↓ correlated with ↓ TGs (r = 0.79) Schmidt et al,155 10 Healthy males Dose ranging: 1.3 g, 4 g, or9gofn-3daily BT, PAI, and HDL ↑, while Fg and TGs ↓ in a dose-dependent 1990 suppl for 6 wk fashion. Highest daily dose (9 g) ↓ TGs, Fg, and vWf, while BT, tPA antigen, PAI, and HDL:TC ratio ↑ Miller et al,156 5 Insulin-dependent Maximum EPA fish oil given without other ACA and collagen-PAggr ↓ equally in both groups, but TG, 1987 diabetics and 5 diet modification for 8 wk lipids, glucose, HbA1c, platelet count, and osmotic healthy volunteers fragility of RBC were unchanged. Whole blood viscosity ↑ in diabetic patients at baseline, but ↓ in both groups at 8 wk. Diabetic patients ↑ vWF at baseline but ↓ in both groups at 6 wk. Addition of EPA or crude fish oil to HUVEC cultures did not change vWF in the supernatant Schmidt et al,157 24 Healthy volunteers 4 g n-3 Daily suppl for 9 mo BT and Fg ↑, while vWf and fibrinolysis ↓. TGs ↓ and a trend 1992 of ↑ HDL, while no changes in TC, LDL, and apolipoproteins A1 and B after 9 mo. Systolic and diastolic BP ↓ with n-3 suppl Smith et al,158 40 Patients with 1 g Fish oil capsules (3.4 g EPA and DHA) TG ↑ by 25%, TC ↑ by 5%, and HDL unaltered. Ivy BT also 1989 previous MI daily suppl for 4 wk; 22 of 40 subjects ↑. Blood glucose and PAI and FVII-PL complex showed a had concomitant long-term oral ↓ trend. Fg ↑ significantly, but clotting time in the anticoagulants combined PT test ↓ in patients who had oral anticoagulants Li et al,159 1999 17 Male vegetarians All had a low-ALA diet for 14 d; then EPA, DHA, total n-3, and n-3:n-6 ratio significantly ↑, randomized to either a moderate- or a whereas ACA:EPA ratio ↓ in platelet PLs, plasma PLs, and high-ALA diet for 28 d triacylglycerols after moderate-ALA or high-ALA diet compared with the low-ALA diet. No differences in thrombotic markers. ALA from vegetable oils (canola and linseed) has a beneficial effect on n-3 concentrations of platelet PLs and plasma lipids in vegetarian males Emeis et al,160 76 Healthy males Fish (mackerel) or meat (control) paste with No changes in plasminogen, PAP, tPA antigen, and 1989 daily main meal for 6 wk euglobulin tPA activity in both groups. In the fish group, total PAI activity ↑ by 45% due to a 71% ↑ in PAI-1. This ↑ could not be ascribed to a diet-induced acute-phase reaction and changes in serum TGs or insulin. No change of PAI activity in controls

Abbreviations: For an explanation of abbreviations, see footnote to Table 1.

Many studies have reported a However, dietary intervention with with n-3 PUFAs. In fact, by pooling positive correlation between serum n-3 PUFA, which is well known for data from all these studies, Hansen triglycerides and PAI-1 activity. Di- its ability to lower triglyceride lev- et al146 were able to calculate that ap- etary interventions, such as a low- els, has not been shown to be par- proximately a 17% increase in PAI-1 saturated-fat diet127 or gemfibrozil allel with a decrease in PAI activ- activity during intervention could be treatment128 to lower serum triglyc- ity,146 indicating that a causal attributed to the fish oil supple- eride levels, have been accompa- relationship is unlikely between lev- ment. Notably, there are few data nied by improvement and even nor- els of triglycerides and PAI-1 activ- on the effect of n-3 PUFA on D- malization of the fibrinolytic activity. ity during dietary supplementation dimer.184,186

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2378

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Platelet Reactivity Thus far, it seems that there is let count and activity.107,207 How- little evidence to support the hy- ever, heavy or binge alcohol intake The effects of n-3 PUFA on platelet pothesis that changes in coagula- is associated with lower fibrino- reactivity have been extensively tion, platelet reactivity, or fibrinolytic capacity with relatively greater investigated. Studies of Greenland lysis systems could account for, at increase in PAI-1 and tPA antigen Eskimos have shown that very high least, some of the beneficial effects than tPA activity.205,208 In addition, intake of marine n-3 PUFA mark- afforded by a Mediterranean diet heavy alcohol consumption also edly inhibited platelet reactivity, with n-3 PUFA supplementation in seems to shift the pendulum to- lowered platelet count, prolonged the secondary prevention trials men- ward a more procoagulant state, with bleeding time, decreased the ratio tioned herein. The major effect of n-3 a rise in the plasma levels of FVII, of proaggregatory thromboxanes to PUFA may be antiarrhythmic rather fibrinogen, and viscosity. Indeed, antiaggregatory prostacyclins, and than antithrombotic.199,200 It re- this may sufficiently predispose in- caused favorable changes in lipid mains unsettled whether the di- dividuals to thrombosis, and in the and lipoprotein profiles. These verse effects of n-3 PUFA supple- presence of an impaired fibrino- findings are of importance for their mentation on thrombogenic indices lytic state, this may contribute to the low incidence of CHD.187 Although are due to different time of supple- increased incidence of ischemic few studies144,174,188 have shown no mentation, patient type, or sepa- stroke seen in heavier or binge drink- significant influence on platelet rate effects of EPA and DHA in ers.20,209 Thus, the results from these reactivity with n-3 PUFA supple- a mixture of fish oils. It is likely studies seem to partly explain the mentation, most other intervention that fish oil and n-3 PUFA have complex relationship between level studies have demonstrated signifi- multifaceted actions in the second- of alcohol consumption and cardio- cant inhibition in platelet reactivity ary prevention of cardiovascular vascular risk seen in large epidemio- of one sort or another but with disease.201 logic outcome studies on alcohol. conflicting combinations of effects with different agonists in vitro or ALCOHOL Experimental Studies: ex vivo.189-195 It seems that platelet Coagulation aggregation induced by low-dose Light-to-moderate alcohol consump- collagen was the most commonly tion (Ͻ30 g/d, ie, 1 to 2 drinks per The results from experimental stud- reported index to be influenced. day) is associated with 10% to 40% ies on the effects of moderate alco- One study152 has demonstrated no lower risks of MI and cardiovascu- hol consumption in both healthy significant difference between lar death compared with absti- subjects and subjects with CHD are supplemented ␣-linolenic acid nence. However, heavy alcohol con- contradictory (Table 5). In par- from vegetable oil and n-3 PUFA sumption or binge drinking ticular, the alcohol effects on fibrino- from a marine source (EPA and increases such cardiovascular risks, gen are variable. For example, Pel- DHA) in their effects on collagen- including stroke.202 The reduction in legrini et al212 found a decrease in induced platelet aggregation and cardiovascular risks with moderate fibrinogen level after consumption thromboxane production, aggrega- alcohol intake has mainly been at- of 30 g of alcohol that consisted of tion to the thromboxane A2 mimic, tributed to an increase of HDL-C lev- red wine and alcohol diluted in fruit urinary excretion of 11-dehydro- els, but this only accounts for 50% juice for 4 weeks but found no thromboxane B2 and ␤-thrombo- of the protective effect.203,204 Increas- change in fibrinogen level after con- globulin, bleeding time, plasma ing evidence has indicated that sumption of dealcoholized red wine. fibrinogen concentration, anti- thrombogenic factors may play an Because alcohol diluted in fruit juice thrombin III activity, FVII coagu- important role in mediating such a had an effect similar to that of red lant activity, or PAI-1 activity. complex association independent of wine, it seems that alcohol is the ef- However, another study has shown HDL-C levels.205,206 fective mediator in alcoholic drinks. that a high ␣-linolenic acid diet has On the other hand, Gorinstein et al210 no significant effect on throm- Epidemiologic Studies: studied the effect of beer (20 g/d of boxane production and platelet Coagulation and Fibrinolysis alcohol) over 30 days and found no aggregation with collagen.185 Nota- change in fibrinogen levels. It could bly, data from studies195 on other The mass of the previously pub- be argued that there may be other fatty acids (mainly n-6 PUFA), lished data on alcohol consump- substances in beer that inhibit the such as linoleic acid on platelet tion and hemostasis comes from epi- beneficial effect of alcohol on fi- reactivity, were highly variable, demiologic studies, with only few brinogen. However, this may also be especially in the in vitro assessment experimental data reported. Gener- due to differences in study design, of platelet aggregations. Similarly, ally, most cross-sectional studies timing of blood samplings, quan- there was also lack of agreement on have shown that light-to-moderate tity or regularity of intake (besides the effect of n-3 PUFA on platelet alcohol intake is associated with a type of beverage used), and intra- adhesiveness.150,151,196-198 Novel, more favorable coagulation and fi- patient and interpatient variability well-validated methods for measur- brinolytic profiles as indicated by in alcohol metabolisms. A recent ing platelet aggregation are desper- lower levels of fibrinogen, white meta-analysis223 of all experimental ately needed to solve current con- blood cell count, plasma viscosity, studies that assessed the effects of troversies.195 FVII, and vWf, as well as lower plate- moderate alcohol intake on lipid lev-

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2379

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 5. Experimental Studies: Effects of Alcohol on Thrombogenic Markers

Source Study Design Alcohol Main Results Gorinstein et al,210 28 Patients with CHD. 22 Had alcohol 330 mL/d of beer (20 g alcohol) Alcohol ↓ FVII activity and antigen and ↓ PAI. Fg 1997 for 4 wk, the other 6 as controls and PT factor unchanged Dimmitt et al,211 55 Men (aged 21-65 y); 4 wk cross-over Predominantly beer; 13 to 58 mL/d Alcohol ↓ Fg and platelet count, but ↑ FVII, tPA 1998 antigen, and PAI Pellegrini et al,212 11 Men (aged 20-45 y); 4 wk Alternating 30 g/d red wine, 30 g/d clear Alcohol in either red wine or alcoholic juice 1996 randomized cross-over with 4 wk fruit juice, and dealcoholized red wine similarly ↓ Fg and collagen-PAggr. No effects washout on ADP-PAggr, tPA antigen, vWF, or plasminogen levels Elmer et al,213 7 Men and 3 women; blood drawn 1 and 2 mL/kg weight of 40% alcohol (average No effects on factors V, VII VIII, XII, 1984 2 h after alcohol intake 53.5 g for 70 kg of weight) plasminogen, or Fg. BT and PAggr ↓ at 1- and 2-h after alcohol intake Pikaar et al,214 12 Men (aged 21-29 y); 5-wk 4 Different amounts of red wine: 0, 23, Dose-response ↓ in tPA activity and 1987 randomized cross-over; blood drawn and 46 g/d of alcohol and in “binge collagen-PAggr. Slight ↑ in plasminogen but after overnight fast drinking” no effects on Fg and ADP-PAggr El-Sayed et al,215 11 Men (mean age, 22.8 y). 0.5 g/kg alcohol or nonalcoholic drink No difference in tPA activity, Fg, fibrin, or Fg DPs 2000 Randomized cross-over, 1-wk washout; blood drawn 45 min after alcohol intake van de Wiel et 50 Men (mean age, 26 y). 2 All consumed red wine 12.5% volume. 2 Glasses had no significant disturbance on the al,216 2001 Experiments: (1) low vs high alcohol Controls had mineral water. Low, 2 circadian rhythm, whereas intake Ն4 glasses intake vs controls; (2) moderate vs glasses (250 mL, 20 g ethanol); inhibited fibrinolysis significantly with a binge vs controls moderate, 4 glasses; high, 6 glasses; dramatic ↑ PAI antigen and PAI activity; tPA binge, 8 glasses antigen also ↑, but tPA activity and PAP complexes ↓. In binge drinkers, the WBCLT ↑ the following morning indicated continued inhibition of fibrinolysis Johansen et al,217 9 Healthy students (aged 23-28 y). Low-dose, loading dose of red wine to Red wine impaired fibrinolysis shown by ↑ 1999 Randomized cross-over; 1 of 3 obtain 0.5/mL (0.40 g/kg ethanol); WBCLT (3.6%, 20.7%, and 55.7% for control, regimens: control, low-dose wine, high-dose, 1/mL (0.80 g/kg ethanol) low-, and high-dose wine, respectively) due to and high-dose wine within 1 h. Followed by 0.13 g/kg ↑ in PAI antigen (−0.8, 4.8, and 11 ng/mL, ethanol per hour for 3 h respectively). No effect of red wine on fibrinolytic system the next morning. Strong correlation observed between WBCLT and PAI antigen Hendriks et al,218 8 Men (aged 45-55 y); 4 treatments 40 g (Red wine, beer, or spirits) or PAI ↑ at 1, 3, 5, and 9 h; tPA antigen ↑ at 3, 5, 1994 randomized controlled order on 4 d mineral water and 9 h; tPA activity ↓ at 1, 3, and 5 h but ↑ at over a period of 11 d 13 h after meal. Similar changes for each beverage type except ↑ in PAI and in 13-h tPA activity were slightly stronger for spirits than for other beverages Numminen et 20 Healthy men. Randomized Ethanol in fruit juice or fruit juice alone Acute ingestion of a large but tolerable dose of al,219 2000 cross-over; 1-wk washout period alcohol transiently ↑ Tbx-mediated platelet activation. A 7-fold ↑ in PAI activity after both morning and evening intakes of alcohol McConnell et 11 Men and 9 women (aged 23-51 y) Beer, 13.5 g/d (low-dose alcohol) No significant ↑ in tPA antigen and activity or al,220 1997 PAI antigen and activity or PT factor and TAT. vWf ↓ but not significant Veenstra et al,221 2 Age groups, 20-30 y and 45-55 y; 8 30 g In red port and wine ADP-PAggr ↑ postprandially but ↓ in overnight

1990 men each fast. No effects on Tbx B2. tPA activity ↓ and PAI ↑ only postprandially and predominantly affect older age group Lacoste et al,222 6 Men and 6 women (mean age, 31 y). 2 oz of 40% alcohol (cognac, 24 g of Compared with baseline, platelet thrombus 2001 At 20 min and 6 h after alcohol alcohol) formation at both the low and high shear rate consumption, blood drawn was flow was significantly ↓ at 20 min and 6 h. infused into a validated ex vivo Men and women showed equal benefit. Badimon superfusion system to Moderate alcohol intake significantly inhibited examine platelet-thrombus formation platelet thrombus deposition under low and on arterial media strips under arterial high shear rates of arterial flow conditions flow conditions simulating vessel stenosis

Abbreviations: For an explanation of abbreviations, see footnote to Table 1.

els and hemostatic factors has con- through favorable changes in lipids mechanism(s) by which moderate cluded that moderate alcohol in- (higher HDL-C level) and hemo- alcohol intake decreases fibrinogen take of 30 g/d is causally related to static profile (lower plasma fibrino- and increases HDL-C levels is not an overall 24.7% lower risk of CHD gen levels). However, the precise known.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2380

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Experimental Studies: inhibit platelet aggregation to most ute to the apparent protection alco- Fibrinolysis specific agonists (adenosine diphos- hol confers against ischemic phate [ADP], thrombin, collagen, coronary and cerebral events. On the Most reports seem to indicate that epinephrine) in platelet-rich plasma. other hand, consistent evidence sug- short-term alcohol ingestion leads to This platelet inhibitory effect seems gests that the relatively greater in- inhibition of the fibrinolytic sys- to persist for several hours after al- crease in PAI-1 and tPA antigen than tem through a rise in circulating cohol intake.224 However, such ben- tPA activity and the rebound phe- PAI-1 levels.211,212,216,217,220,221 Nota- eficial effect is not seen in binge nomenon of platelet hyperaggrega- bly, a recent study217 has shown an drinkers or in individuals with al- bility with short-term alcohol acute, dose-dependent rise in PAI-1 coholism after alcohol withdrawal; (binge) intake may attenuate this antigen level with a parallel prolon- instead, a rebound phenomenon of benefit, resulting in a net antifibri- gation of whole blood clot lysis time platelet hyperaggregability (espe- nolytic effect of ethanol consump- after intake of the high dose of red cially toward thrombin agonist in tion,205 predisposing individuals to wine. In addition, there was also a vitro)225 and loss of the normal cir- coronary thrombosis and contrib- tendency for tPA antigen to in- cadian periodicity of the hemo- uting to the increased incidence of crease dose dependently, although static system is observed.219 This may ischemic stroke. this was only significant for the high- explain the increased ischemic dose wine group. In the case of binge strokes or sudden deaths that are SMOKING drinking, this inhibition in fibrino- known to occur after episodes of lysis effect persists into the morn- binge or heavy drinking.226 Intrigu- The direct effects of smoking on ath- ing following the evening of alco- ingly, such a rebound phenom- erothrombogenesis are still un- hol consumption,216 and when it enon is not observed after moder- clear. This may be mediated by its coincides with the physiologic morn- ate red wine consumption in many adverse effects on endothe- ing dip in fibrinolytic activity, this humans and, in fact, this protec- lial function, vascular tone, hemo- may predispose susceptible indi- tion afforded by red wine has been stasis, lipid profile, and inflamma- viduals to sudden cardiac death. duplicated in rats by alcohol with tory cells. Much of the data regarding However, the exact reason for de- grade tannins added, which con- the effects of smoking on thrombo- creased fibrinolysis after short- tain the polyphenolic compounds genic factors have been derived from term alcohol intake is still unre- with which red wines are richly en- epidemiologic and cross-sectional solved. Indeed, several studies have dowed. However, it is still unclear studies. However, intervention stud- discussed whether it is the ethanol how red wine or wine phenolics in ies are accumulating and have re- component itself or other media- particular could significantly in- ported an increase in blood coagu- tors in red wine (or beer) that inhibit platelet aggregation. lability but impaired fibrinolysis in duce acute changes in tPA or PAI-1. In an interesting study by habitual smokers when compared Veenstra et al221 and Hendriks et al218 Lacoste et al,222 rather than evaluat- with nonsmoking controls have pointed out that the red wine ing platelet function and platelet in- (Table 6). It seems that higher effect on t-PA and PAI-1 antigen lev- hibition, the authors assessed the plasma levels of fibrinogen and vis- els is probably caused by the effect effect of alcohol directly on platelet- cosity are the main contributors to of ethanol but not the effects of other dependent thrombosis in 12 healthy higher coagulability found in smok- mediators, such as the phenolic com- subjects in an ex vivo model that ers, whereas the lower fibrinolytic pounds in red wine or port wine, al- simulates a deep arterial wall in- potential is mainly attributed to an though these constituents might jury exposed to shear forces typical increase in PAI-1 activity and pos- contribute to the effects on platelet of flow at sites of stenosed arteries, sibly also a decrease in tPA activity function. However, it is difficult to reflecting the in vivo situation of and lower plasminogen lev- distinguish between the effects of red coronary thrombosis. The study els.107,228,229,244 wine and alcohol per se. Notably, in demonstrated for the first time that a study211 with longer-term con- moderate alcohol consumption (24 Coagulation sumption of beer (4 weeks), the tPA g of alcohol) in humans had a po- antigen and PAI-1 levels were in- tent extracorporeal antithrombotic In cross-sectional epidemiological creased markedly. effect both at the time of peak alco- studies, lifetime duration of smok- hol concentration and 6 hours after ing is a strong determinant of ini- Experimental Studies: Platelet alcohol ingestion when blood alco- tial plasma fibrinogen levels. The ef- Reactivity hol level has returned to baseline. fects of smoking on the hemostatic Overall, the balance of antico- system remain for many years be- Alcohol has also been thought to re- agulant, procoagulant, and fibrino- fore an exsmoker reverts to a plasma duce CHD risk by decreasing plate- lytic effects in any individual in re- level similar to that of a lifetime non- let reactivity. Indeed, several stud- sponse to alcohol intake may vary, smoker,229,245 although a decrease in ies in humans and animals have depending on quantity and type of fibrinogen levels follows quickly af- demonstrated that the immediate alcoholic beverage ingested and ter cessation of smoking.246 In pro- effect of light-to-moderate alcohol, other variables.204,227 The lower level spective data, smoking cessation and either added in vitro to platelets or of plasma fibrinogen with moder- the adoption or resumption of smok- 10 to 20 minutes after ingestion, can ate alcohol intake may well contrib- ing are associated with a decrease or

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2381

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 6. Effect of Smoking on Thrombogenic Factors

Source Methods Main Results Conclusion Eliasson et al,228 Cross-sectional study; 604 men and 662 Male smokers had 0.34 g/L ↑ Fg than nonsmokers. Cigarette smoking is associated with ↑ 1995 women (aged 25-64 y): smokers, Numbers of cigarettes smoked correlated with Fg. Fg, unaltered fibrinolysis, and normal ex-smokers, snuff dippers, and Fg not affected by snuff dipping. Tobacco use had glucose tolerance and insulin levels. nonsmokers no relationship with tPA or PAI activity and no Moist oral snuff does not seem to influence on glucose and postload insulin levels affect these cardiovascular risk factors Meade et al,229 Cross-sectional; 2023 white men Duration of smoking is a determinant of initial Fg level, A substantial part of the relation between 1987 which ↓ soon after cessation but returned to smoking and IHD seems to be normal levels only after 5 years. Smoking cessation mediated through the Fg concentration and the adoption or resumption of smoking associated with a ↓ or an ↑, respectively, of ~0.15 g/L in plasma Fg equivalent to a ↓ or an ↑ in the risk of IHD by ~20% Simpson et al,230 Cross-sectional study; 54 healthy PAI antigen ↑ in smokers vs nonsmokers with Smokers associated with ↓ fibrinolysis as 1997 individuals; current, ex-smokers, or intermediate levels in ex-smokers. tPA and platelet reflected by ↑ PAI. Smoking cessation nonsmokers (aged 25-40 y) pool PAI were not different in the 3 groups. tPA:PAI seems to normalize fibrinolysis. ratio similar in ex-smokers and nonsmokers but ↓ Platelet pool of PAI not quantitatively in smokers. PAI antigen and activity correlated with affected by smoking. Effect of chronic pack-years of cigarettes smoked. PAI correlated smoking on PAI may be mediated by strongly with tPA and TGs. tPA correlated strongly TGs and insulin resistance with TGs Enderle et al,231 30 Healthy smokers (mean age, 40.6 y) FMD significantly ↓ and IMT tended to be ↑ in Peripheral ET dysfn is common in 2000 vs nonsmokers. FMD as marker of ET smokers. TAT, Fg, PAP, tPA, and PAI activity did not smokers even without major dysfn differ between smokers and controls alterations in markers of coagulation and fibrinolysis Newby et al,232 12 Smokers (5-20 cigarettes per day) and Substance P caused a dose-dependent ↑ in blood flow Smoking caused marked inhibition of 1999 12 nonsmokers (aged 25-55 y). and local release of tPA antigen and activity in substance P−induced tPA release in Forearm blood flow assessed by smokers and nonsmokers, but the ↑ was vivo. ET dysfn may ↑ the risk of venous occlusion plethysmography significantly lower in smokers that the release of atherothrombosis through a ↓ in the following infusions of substance P tPA antigen and activity ↓ by 51% and 53%, acute fibrinolytic capacity respectively. PAI did not change Newby et al,233 15 Ex-smokers or current smokers vs 10 Substance P and SNP ↑ LAD blood flow, but ↑ Coronary plaque burden and smoking are 2001 nonsmokers (mean age, 56 y). Saline, coronary sinus tPA antigen and activity only with associated with ↓ acute local substance P, or SNP infused into LAD; substance P. Release of active tPA strongly fibrinolytic capacity. There may be a blood flow measured by IVUS and with inversely correlated with LAD plaque burden. direct link between endogenous arterial and coronary sinus sampling. Smoking associated with ↓ coronary release of fibrinolysis, ET dysfn, and active tPA atherothrombosis in the coronary circulation Allen et al,234 30 Healthy men (aged 30-40 y): smokers Smokers had ↓ baseline fibrinolytic activity as Smokers have ↓ fibrinolytic capacity both 1985 (Ն20 cigarettes per day) vs indicated by dilute blood clot lysis, euglobulin-fibrin at rest and in response to DDAVP nonsmokers. Fibrinolysis was studied plate assay, and tPA activity. No differences between compared with nonsmokers at rest and after infusion of DDAVP the groups in various fibrinolytic inhibitors or in the intrinsic fibrinolytic activation pathways. The ↑ levels of tPA activity and FVIII R antigen in response to DDAVP also ↓ in smokers Belch et al,235 10 Habitual smokers vs nonsmokers. Smokers had ↑ plasma Fg and viscosity but ↓ ↑ Hypercoagulability in smokers 1984 Blood drawn before and 10 min after plasminogen and tPA. After smoking 3 cigarettes, compared with controls, which smoking 3 cigarettes over 30 min smokers had an ↑ in ADP-PAggr and ↑ in $2M and becomes more pronounced FVIII antigen but plasma viscosity and red cell immediately after smoking 3 cigarettes deformability ↓

(continued)

an increase, respectively, of approxi- ers or light smokers. However, other arterial beds, which suggests a pos- mately 0.15 g/L in plasma fibrino- groups have reported no difference sible direct link among impaired en- gen. There is evidence to suggest that in markers of fibrinolysis and co- dogenous fibrinolysis, endothelial these changes are not due to con- agulation, although endothelial func- dysfunction, and arterial athero- current changes in other lifestyle tion has been found to be signifi- thrombosis in smokers.232,233,249 How- variables.107 cantly impaired in healthy smokers ever, although the basal level of compared with nonsmoking con- PAI-1 activity is higher in long- Fibrinolysis trols.231 Interestingly, smokers may term smokers, rapid smoking of 2 have markedly impaired acute sub- cigarettes in these patients neither Impaired fibrinolytic potential has stance P–induced endothelial re- stimulated fibrinolysis nor changed been found in smokers with CHD247 lease of active tPA in vivo from coro- levels of tPA or PAI-1 activities.236 and peripheral vascular disease,248 nary and brachial arteries and was Plasma PAI-1 antigen seems to cor- with higher levels of PAI-1 activity closely related to impaired endothe- relate with cumulative smoking in and tPA antigen than in nonsmok- lial function in the correspondence pack-years,230,236 and on the other

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2382

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 6. Effect of Smoking on Thrombogenic Factors (cont)

Source Methods Main Results Conclusion Haire et al,236 5 Healthy male smokers (aged 35-45 y) TPA antigen release in response to venous occlusion Smoking did not acutely alter 1989 was intact at both 8 am and 3 pm. Rapid smoking fibrinolysis in chronic smokers, but of 2 cigarettes neither ↑ fibrinolysis nor changed they had abnormal fibrinolysis with ↑ tPA or PAI. Functional PAI and ECLT ↑ in smokers PAI activity. Abnormal fibrinolysis vs matched controls. Plasma and platelet PAI may contribute to the thrombotic were similar in both groups diathesis of smokers Brockmann et 54 Healthy subjects into 4 groups: Smokers had ↑ Fg vs nonsmokers. No differences In habitual smokers, PHC and the al,237 2001 nonsmoking and smoking males and between smokers and nonsmokers in PHC, for agonists-induced PAggr assays are females. Effect of smoking on PHC neither the collagen/epinephrine nor the not significantly influenced or ↑ with an in vitro analyser was collagen/ADP cartridges. PAggr assays performed compared with healthy nonsmokers. examined within4hofblood sampled in parallel also showed no differences An immediate effect of smoking cannot be excluded Modesti et al,238 8 Male healthy smokers (Ͼ20 cigarettes Smokers had a significantly ↑ number of TbxA2 Changes could contribute to the ↑ 1989 per day) vs 9 healthy controls (aged platelet receptors vs nonsmokers. No differences responsiveness of platelet from 30-55 y). Platelet TxA2 receptors in the receptor affinity between the 2 groups smokers to external aggregating assessed by a radioligand binding stimuli method Pernerstofer et 20 Healthy smokers (Ͼ20 cigarettes per Smokers had ↑ sP-sel expression on platelets than P-sel expression on platelets ↑ in al,239 1998 day) vs 20 healthy nonsmokers nonsmokers. sP-sel expression on platelets and smokers but low-dose aspirin does (mean age, 30 y). DBR; aspirin (100 circulating sP-sel unchanged by aspirin not ↓ platelet activation in smokers mg/d) or placebo; cross-over; 2 wk wash-out period Blache et al,240 12 Fasting smokers. Blood drawn before Before aspirin: PAggr to thrombin and ADP, plasma Data indicate an interrelationship 1992 and at 10 min after inhaled smoke 1 nicotine, ␤-Tbg, and CECs all ↑ significantly after between platelet hyperactivity and ET cigarette. Experiment repeated 1 wk smoking, but PAggr ratio ↓. After aspirin: except injury. Aspirin inhibits later except that subjects had aspirin for plasma nicotine, all smoking-induced changes smoking-induced changes. Aspirin (650 mg) 10-14 h before blood were abolished by ingestion of aspirin may offset several of the deleterious sampling. acute effects of smoking. Long-term effects of both smoking and aspirin treatment remain unclear Blann et al,241 20 Smokers vs 10 nonsmoker controls Smokers had a transient ↑ in leukocyte count and Rapid smoking of 2 cigarettes in 1998 (aged 21-55 y). Blood drawn before, neutrophil activation, but vWf ↑ steadily at each succession activates leukocytes and immediately after, and at 10 and 30 time point. No changes in neutrophil elastase, causes ET cell damage but will not min after rapid smoking of 2 sICAM-1, Fg, platelet count, or sP-sel immediately influence platelet activity cigarettes in sequence Gleerup et al,242 10 Smokers and 11 nonsmokers, all ADP-PAggr ↑ in smokers at rest vs nonsmokers and Smoking adds a further element of ↑ 1996 with mild HT (diastolic BP, 90-110 persisted both in upright posture and after platelet activity to that inherently mm Hg). Successive measurements exercise. ␤-Tbg also ↑ in smokers. No difference present in HT after erect posture for 10 min and of ECLT and PAI between the 2 groups after a 5-min exercise test Davis et al,243 20 Healthy nonsmokers. Smoked 2 Mean ET cell counts before and after tobacco Much greater effects of tobacco smoking 1985 tobacco cigarettes in 20 min and at cigarettes were 2.3 and 4.8; before and after on ET cell counts and PAggr ratios cross-over 1 wk later smoked 2 nontobacco cigarettes were 2.5 and 3.0. The suggest that nontobacco smoking may cigarettes made from wheat, cocoa, corresponding mean PAggr ratios were 0.80 and be less harmful to the cardiovascular and citrus plants 0.65, 0.81, and 0.78, respectively system than tobacco smoking.

Abbreviations: For an explanation of abbreviations, see footnote to Table 1.

hand, other studies75 have sug- vitro compared with nonsmokers.237 changed P-selectin expression on gested that smoking cessation of at Certainly, no immediate influence in platelets and circulating P-selectin least 6 months was associated with platelet activity (as indicated by plate- plasma levels. This could indicate that a decrease in plasma PAI-1 activity. let count and soluble P-selectin) oc- enhanced thromboxane A2 produc- curs after rapid smoking of 2 ciga- tion may not be the primary mecha- Platelet Reactivity rettes in sequence compared with nism for increased P-selectin expres- nonsmoking controls, despite evi- sion in smokers.239 Similarly, Smoking is also known to be associ- dence of endothelial damage (indi- dipyridamole alone or in combina- ated with increased platelet throm- cated by elevated vWf level).241 On the tion with aspirin did not have any sig- bus formation,250 but studies on the other hand, increased P-selectin ex- nificant effect on plasma concentra- effects of smoking on platelet reac- pression on platelets has been dem- tions of ␤-thromboglobulin, platelet tivity have produced conflicting data. onstrated in young, healthy, ha- factor 4, the circulating endothelial For example, rapid smoking of 3 ciga- bitual smokers compared with cell count (indicates endothelial dam- rettes in habitual smokers increased nonsmokers, underlining the in- age), and the platelet aggregate ratio ADP-induced platelet aggregation in creased platelet activation in nicotine- in habitual, male smokers with vitro in one study,235 but others have abusing subjects.239 Importantly, 100 CHD.251 Smoking cessation for 6 shown no differences in various ago- mg/d of aspirin did not reduce plate- weeks, however, has resulted in a 29% nist-induced platelet aggregations in let activation as measured by un- reduction of circulating P-selectin

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2383

Downloaded From: https://jamanetwork.com/ on 09/23/2021 plasma levels in healthy smokers252 increase in PAI-1 and a decrease in platelet activation, which suggests and a decrease in platelet count 2 tPA activities.262,263 In addition, such that platelets are not being stimu- weeks after cessation.253 Further- long-term mental stress is also inde- lated through the cyclooxygenase- more, aspirin abolished the major pendently related to an increase in dependent pathway.37,38,41,268,269 How- cigarette smoke–induced endothe- prothrombic tendency, with in- ever, the mechanism(s) responsible lial damage and platelet hyperactiv- creased levels of fibrinogen, FVII an- for the increased of prothrombic ten- ity in the presence of high plasma tigen, and activity.264 Changes in he- dency secondary to psychosocial nicotine levels.240 Notably, passive ex- mostatic variables in response to stress may be related to the sympa- posure to tobacco smoke also seems psychosocial job stress are particu- thoadrenal pathways, but clearly this to raise the endothelial cell count and larly interesting. Significant eleva- needs further exploration. platelet aggregate ratio in a manner tion in coagulation FVII and FVIII lev- similar to that previously observed els, fibrinogen level, thrombocyte COFFEE, TEA, OR CAFFEINE with active smoking.254 count, thrombin level, and ADP- CONSUMPTION ON induced platelet aggregation has been THROMBOGENESIS PSYCHOSOCIAL STRESS reported during a period of increased workload compared with a Coffee The increase in mental stress, de- calm work period.265 High job demands at work, anger, or low socio- mands have also been significantly re- The possible link between coffee or economic strain has been associ- lated to decreases in tPA activity (ie, caffeine consumption and the risk ated with an increased risk of lower fibrinolytic capacity, indepen- of CHD is far from settled, but its ef- atherovascular disease.255-257 The in- dent of other traditional cardiovas- fects on various thrombogenic fac- creased cardiovascular risk may be cular risk factors)266 and hence in- tors might be relevant. However, secondary to excessive cardiovascu- crease the likelihood of fibrin limited data from intervention study lar reactivity to stress258 but may also deposition. The mechanism(s) un- are currently available. involve activation of the coagula- derlying these changes is unknown, tion and fibrinolysis systems. but impaired fibrinolysis in people Coagulation with long-term psychosocial stress has Coagulation and Fibrinolysis been linked to insulin resistance, obe- Two well-described randomized sity, and triglyceride levels.262 controlled trials270 have reported that Indeed, an association between psy- brewed or boiled coffee, caffeine- chological factors and several of the Platelet Reactivity containing drinks, and decaffein- coagulation and fibrinolysis vari- ated drinks did not have any effects ables related to atherosclerosis has Platelet reactivity also seems to be af- on hemostatic variables, such as fi- provided a plausible psychobiologi- fected by a variety of psychosocial brinogen level, FVII activity, FVIII cal link to CHD. The characteristic stressors. Acute mental stresses sig- antigen, and protein C and S levels. patterns of coagulation and fibrino- nificantly induce all platelet reactiv- Another experimental study271 seems lysis activation in response to vari- ity variables, such as platelet activa- to support such observations, but an- ous psychological stressors seem to tion or secretion and in vitro or ex other cross-sectional study272 re- follow closely that of physical or ex- vivo platelet aggregation, in parallel ported an increased in plasma fi- ercise-induced changes in markers of to a concomitant incremental in- brinogen levels with increased coffee thrombogenesis. Accordingly, as in crease of various hemodynamic in- consumption. short-term, strenuous exercise, acute dices that follow during mental- mental stress simultaneously acti- stress testing. Such a response has Fibrinolysis vates the coagulation system, with in- been consistently shown in almost creased levels of fibrinogen, total every study. Furthermore, these The available evidence seems to sug- plasma protein, hematocrit, FVII, and changes seem to be more pro- gest that coffee enhances fibrino- FVIII,259,260 and enhances fibrinoly- nounced in patients with atherovas- lytic potential as whole blood fibri- sis with increased activity of tPA cular disease37,267 compared with nolysis time is shortened273 and within a physiological range in healthy controls. Hence, such a re- PAI-1 levels are decreased, whereas healthy subjects.260,261 In patients with sponse may precipitate acute ische- tPA activity increases274 after con- atherosclerosis and impaired endo- mic coronary events in patients at sumption of coffee and such effects thelial anticoagulant function, how- high risk of cardiovascular events, in- are blunted during caffeine absti- ever, procoagulant responses to acute cluding individuals with sedentary nence. However, one study271 did not stressors may outweigh anticoagu- lifestyle. One group has reported that find an effect of abstinence from caf- lant mechanisms and thereby pro- dipyridamole attenuated the plate- feine on blood clot lysis time. mote a hypercoagulable state. let hyperreactivity in post-MI pa- Similarly, long-term psychoso- tients but had no effect on stress- Platelet Reactivity cial stressors, such as prolonged job induced increase of hemodynamic stress or low socioeconomic strain variables and epinephrine levels.267 The effects of caffeine intake on that provoked a state of vital exhaus- Aspirin has only a minimal platelet activity are more variable. tion, have been independently asso- effect on physical, psychosocial, or Several in vitro and in vivo stud- ciated with hypofibrinolysis, with an norepinephrine stress–induced ies275,276 have reported increased

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2384

Downloaded From: https://jamanetwork.com/ on 09/23/2021 platelet activation and release after habilitation with either exercise CONCLUSIONS coffee consumption, but others have alone or exercise as part of a com- found the opposite effects.271,272,277 prehensive rehabilitation program in The hemostatic system is assuming Again, this may be due to the lack post-MI and postrevascularization an increasingly prominent role in of standardization in the analysis patients significantly reduced all- the pathogenesis and progression of methods used to assess platelet re- cause or total cardiac mortality by atherovascular diseases. Human activity. at least 26% to 31%.287 In addition lifestyle or physical activities have to the reduction of cardiovascular diverse effects on coagulation, fibri- Tea: Coagulation, Fibrinolysis, morbidity and mortality, cardiac re- nolysis, and platelet reactivity. and Platelet Reactivity habilitation also significantly im- There have been abundant studies proves functional capacity and qual- of the effects of exercise, weight Previous epidemiologic studies have ity of life and lipid profile and blood loss, dietary lipids (especially n-3 suggested that tea consumption is as- pressure.288 PUFA), smoking, alcohol, and psy- sociated with a decreased risk of car- Thus, given all the evidence dis- chosocial stress on the 3 main sys- diovascular events, but a recent cussed herein, it is highly plausible tems of thrombogenesis. The data meta-analysis278 has reported no sig- that lifestyle modifications through from intervention and randomized nificant association and, in con- a program that incorporates step- clinical trials are largely frag- trast, the risk may be even in- wise increment of physical training mented, rarely complete, and creased for CHD in the United or exercise, patient education and inconsistent, mainly due to the dif- Kingdom and for stroke in Austra- advice, dietary modifications, and ferences in study design and the lia with increasing tea consump- psychosocial stress management inherent complexity of subjects’ tion. Indeed, the antioxidative poly- would have a significant impact on confounders and the lack of stan- phenolic flavonoids found in tea patients’ thrombogenic profile and dardization of the various analytical have been shown to prevent oxida- hence may beneficially influence the methods used in the assessment of tion of low-density lipoproteins both overall cardiovascular risk. coagulation, fibrinolysis, and plate- in vitro and in vivo279 and to inhibit However, so far, to our knowl- let function. The in vivo signifi- platelet aggregation in vitro.280-282 edge, no study has been reported on cance of examining one portion of However, recent randomized the overall effects of such a compre- the complex overall system is controlled trials283-285 of black or hensive cardiac rehabilitation pro- unclear. How much could one cor- green tea or tea extracts have found gram on changes of the various vari- relate the in vitro or ex vivo find- no effects on both hemostatic and fi- ables of hemostasis, fibrinolysis, and ings to the true in vivo biological brinolytic variables (eg, fibrinogen, platelet reactivity. Previous studies activities in many of the human vWf, or FVII and PAI-1, tPA, or uro- have mainly focused on the effects biological systems is largely kinase-type plasminogen activa- of short-term or regular physical ac- unknown. Nevertheless, these data tor) or on inflammatory markers tivities on fibrinolytic responses in have provided us with important such as C-reactive protein. Simi- post-MI or post–coronary artery by- preliminary explanations for the larly, no significant difference was pass grafting patients who partici- relative contribution of the various found from black tea consumption pated in cardiac rehabilitation ex- thrombogenic markers in relation on ex vivo platelet aggregation in pa- ercise programs (Table 7). to lifestyle habits to clinical out- tients with CHD286 and on in vitro In keeping with epidemio- comes reported in epidemiologic platelet aggregation in healthy sub- logic data, patients with CHD have studies. Available evidence from jects257 when compared with drink- higher basal levels of PAI-1 and tPA these studies support lifestyles that ing hot water. Interestingly, in the antigen, suggesting impaired fibri- adopt strategies to lose weight, stop same study, the latter group had nolytic activity compared with cigarette smoking, engage in regu- found significantly lower (15%) healthy subjects. Although healthy lar moderate exercise and relax- soluble P-selectin levels (but not subjects tended to have a marked fi- ation, and regularly consume light- other adhesion molecules) in those brinolytic response to exercise, pa- to-moderate alcohol and fatty fish who drank black tea; however, tients with CHD have a lower in- should significantly lower coagula- whether such a finding is of any crease in the fibrinolytic potential as bility, promote fibrinolysis, and clinical significance is unclear. Thus, evidenced by changes in tPA activ- reduce platelet reactivity. The over- it seems that the putative protec- ity and PAI-1 levels after regular all effects ought to translate into an tive effect of tea against develop- physical training.292 Although the in- improved cardiovascular or other ment of CHD may not be mediated crease of fibrinolytic capacity may be beneficial clinical outcome in through effects of tea consumption counterbalanced by an increase in healthy individuals, those with car- on hemostasis, fibrinolysis, or plate- blood coagulability and platelet ac- diovascular risk factors, or those let activity. tivity during short-term exercise,91 with established CHD. It follows lower plasma fibrinogen level has in that a cardiac rehabilitation pro- CARDIAC REHABILITATION fact been found in both post-MI and gram that incorporates a stepwise AND THROMBOGENESIS post–coronary bypass grafting pa- increment of physical training or tients who engage in regular aero- exercise, patient education and A recent Cochrane Systematic Re- bic exercise during cardiac rehabili- advice, dietary and personal habit view has concluded that cardiac re- tation.82,83 modifications, and psychosocial

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2385

Downloaded From: https://jamanetwork.com/ on 09/23/2021 Table 7. Controlled or Intervention Clinical Trials of Cardiac Rehabilitation Exercise on Thrombogenic Factors

Source Subjects Exercise Main Results Conclusions Weiss et al,92 12 Males with CHD without Blood drawn before and CHD patients had ↓ PTF F1 + 2 and Patients and controls similarly ↑ 1998 MI within the preceding 6 after rehab group remained unchanged after exercise, coagulation and fibrinolysis mo (aged 55±9y)vs12 exercise session lasting whereas a significant ↑ occurred in induced by exercise. Rehab healthy male controls 1h controls. Postexercise platelet count, exercise in CHD patients (aged 52±7y) ␤-Tbg, TAT, and FPA ↑ in both groups beyond the immediate but was more pronounced in post-MI period has no controls. Exercise ↑ PAP complexes detrimental effects on in both groups. Repeated thrombin, fibrin, and plasmin experiments after aspirin given to formation controls did not alter results Suzuki et al,82 56 Active exercise post-MI Blood drawn before and Training ↓ Fg, FVIII, vWf, TAT, Training ↓ coagulation in 1992 patients vs 30 post-MI after 1 mo of systematic plasminogen activity, hematocrit, post-MI patients: ↓ in Fg, patients without training physical training platelet counts, and ␣2AP. APTT FVIII, vWf, FVII, and TAT. The prolonged with training. In 20 ↓ in plasminogen, tPA patients with training, resting ␣2AP, antigen, a2AP, PAI, and TAT, protein C, PAI, and FVII protein C after training may significantly ↓ after 1 mo, but tPA result from the ↓ coagulation activity unchanged Wosornu et al,83 55 Men (aged 32-70 y) 6 Months of aerobic or Both trained groups ↑ exercise capacity Aerobic training post-CABG 1992 within 12 mo of CABG power exercise on at 6 mo. Aerobic exercise improved improves exercise capacity randomized to aerobic or treadmill (3 times per earlier at 3 mo. Fg ↓ with aerobic and ↓ Fg, which is maintained power exercise or week) in training patients exercise but ↓ slightly with power with further training. Power controls exercise. Gradual ↑ in FVII activity in exercise causes delayed aerobic and control groups but small benefit in treadmill ↓ in the power group. No consistent performance and a small ↓ in changes in FPA Fg. These changes may ↓ cardiovascular morbidity 289 Fernhall et al, 13 Post-MI patients 2 Maximal exercise on V˙ O2max uptake, HR, and ventilation greater Fibrinolysis ↑ similarly in both 1998 treadmill and BE. Blood on treadmill than BE. Blood lactate modes. Exercise intensity, but drawn before and after similar between the 2 modes. tPA not the mode, seemed to be each test activity ↑ but a trend in ↓ PAI activity the primary determinant of with exercise fibrinolytic response to acute exercise Speiser et al,77 71 Males: 2 pairs of BE exercise At baseline, those doing regular sporting (1) Regular vigorous exercise↑ 1988 age-matched groups. (1) activities showed ↓ PAIvsthe fibrinolysis by ↓ PAI in healthy Athletes, (2) not engaged respective age-matched controls. individuals; (2) rehab sport in any sports, (3) regularly During exercise tPA antigen similar does not ↓ PAI in post-MI practicing sports, and (4) between the age-matched groups but patients compared with post-MI patients in a tPA activities ↑ after exercise in groups age-matched healthy subjects rehab sports program with lower pretest PAI. No change on regularly exercise; and (3) D-dimer in any group activation of fibrinolysis during exercise has no systemic fibrinolytic effect Estelles et al,290 Post-MI patients, active Rehab program with BE. At 6 mo, tPA activity ↓ in controls but Patients in rehab program 1989 exercise vs controls Blood drawn at end of slightly ↑ in active group. PAI ↑ in showed a slight ↑ in hospitalization and at 3 controls but remained constant or ↓ fibrinolytic capacity but ↓ and6mo slightly in active group significantly in controls Paramo et al,291 30 (M/F, 22/8; mean age, 47 9 Months of cardiac rehab. Marked ↓ in functional PAI after 3 and 9 Cardiac rehab improved 1998 y) Survivors of a first MI Blood drawn before and at mo in MI patients. Also significant ↑ of fibrinolysis and lipid profile in vs 30 healthy controls 3 and 9 mo after program HDL-C and ↓ of Lp(a) after the post-MI patients program

Abbreviations: For an explanation of abbreviations, see footnote to Table 1.

stress management would have a We acknowledge the support of REFERENCES significant impact on patients’ the City Hospital Research and De- hemostatic profiles and hence ben- velopment Programme for the Hae- 1. Fuster V, Badimon JJ, Chesebro JH. Athero- eficially influence the overall car- mostasis, Thrombosis, and Vascular thrombosis: mechanisms and clinical therapeu- diovascular risk. Biology Unit, Birmingham. tic approaches. Vasc Med. 1998;3:231-239. 2. Davies MJ. The pathophysiology of acute coro- Corresponding author: Gregory nary syndromes. Heart. 2000;83:361-366. Accepted for publication December 17, Y. H. Lip, MD, FRCP, Haemostasis, 3. Lip GY, Blann AD. Thrombogenesis and fibri- 2002. Thrombosis and Vascular Biology nolysis in acute coronary syndromes: impor- This study was funded by a Na- Unit, University Department of Medi- tant facets of a prothrombotic or hypercoagulable state? J Am Coll Cardiol. 2000;36:2044- tional Health Service Research and De- cine, City Hospital, Dudley Road, Bir- 2046. velopment Health Technology assess- mingham B18 7QH, England (e-mail: 4. Makris TK, Tsoukala C, Krespi P, et al. Haemo- ment project grant (Dr Lee). [email protected]). stasis balance disorders in patients with

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2386

Downloaded From: https://jamanetwork.com/ on 09/23/2021 essential hypertension. Thromb Res. 1997;88: minogen activator and risk of myocardial infarc- 38. Hurlen M, Seljeflot I, Arnesen H. Increased plate- 99-107. tion. Lancet. 1993;341:1165-1168. let aggregability during exercise in patients with 5. Woodward M, Lowe GD, Rumley A, et al. Epi- 23. Juhan-Vague I, Pyke SD, Alessi MC, Jespersen previous myocardial infarction: lack of inhibi- demiology of coagulation factors, inhibitors and J, Haverkate F, Thompson SG, European Con- tion by aspirin. Thromb Res. 2000;99:487-494. activation markers: The Third Glasgow MONICA certed Action on Thrombosis and Disabilities. Fi- 39. Willich SN, Lewis M, Lowel H, Arntz HR, Schu- Survey, II: relationships to cardiovascular risk brinolytic factors and the risk of myocardial in- bert F, Schroder R, Triggers and Mechanisms of factors and prevalent cardiovascular disease. Br farction or sudden death in patients with angina Myocardial Infarction Study Group. Physical ex- J Haematol. 1997;97:785-797. pectoris. Circulation. 1996;94:2057-2063. ertion as a trigger of acute myocardial infarc- 6. Lip GY, Blann AD, Jones AF, Lip PL, Beevers DG. 24. Junker R, Heinrich J, Schulte H, Erren M, Ass- tion. N Engl J Med. 1993;329:1684-1690. Relation of endothelium, thrombogenesis, and mann G. Hemostasis in normotensive and hy- 40. Larsson PT, Wallen NH, Hjemdahl P. Norepi- hemorheology in systemic hypertension to eth- pertensive men: results of the PROCAM study. nephrine-induced human platelet activation in nicity and left ventricular hypertrophy. Am J Car- J Hypertens. 1998;16:917-923. vivo is only partly counteracted by aspirin. Cir- diol. 1997;80:1566-1571. 25. Thompson SG, Kienast J, Pyke SD, Haverkate F, culation. 1994;89:1951-1957. 7. Colwell JA. Treatment for the procoagulant state van de Loo JC, European Concerted Action on 41. Kawano TA, Aoki N, Homori M, et al. Mental in type 2 diabetes. Endocrinol Metab Clin North Thrombosis and Disabilities Angina Pectoris stress and physical exercise increase platelet- Am. 2001;30:1011-1030. Study Group. Hemostatic factors and the risk of dependent thrombin generation. Heart Vessels. 8. Osende JI, Badimon JJ, Fuster V, et al. Blood myocardial infarction or sudden death in pa- 2000;15:280-288. thrombogenicity in type 2 diabetes mellitus patients with angina pectoris. N Engl J Med. 1995; 42. Zanettini R, Bettega D, Agostoni O, et al. Exer- tients is associated with glycemic control. JAm 332:635-641. cise training in mild hypertension: effects on Coll Cardiol. 2001;38:1307-1312. 26. Thaulow E, Erikssen J, Sandvik L, Stormorken blood pressure, left ventricular mass and co- 9. Makris TK, Stavroulakis GA, Krespi PG, et al. Fi- H, Cohn PF. Blood platelet count and function agulation factor VII and fibrinogen. Cardiology. brinolytic/hemostatic variables in arterial hyper- are related to total and cardiovascular death in 1997;88:468-473. tension: response to treatment with irbesartan apparently healthy men. Circulation. 1991;84: 43. El Sayed MS, Lin X, Rattu AJ. Blood coagula- or atenolol. Am J Hypertens. 2000;13:783-788. 613-617. tion and fibrinolysis at rest and in response to 10. Ernst E, Resch KL. Therapeutic interventions to 27. Trip MD, Cats VM, van Capelle FJ, Vreeken J. maximal exercise before and after a physical con- lower plasma fibrinogen concentration. Eur Heart Platelet hyperreactivity and prognosis in survi- ditioning programme. Blood Coagul Fibrinoly- J. 1995;16(suppl A):47-52. vors of myocardial infarction. N Engl J Med. 1990; sis. 1995;6:747-752. 11. Sacks FM, Svetkey LP, Vollmer WM, et al, DASH- 322:1549-1554. 44. El Sayed MS. Effects of high and low intensity Sodium Collaborative Research Group. Effects 28. Koenig W, Sund M, Lowel H, Doring A, Ernst E. aerobic conditioning programs on blood fibri- on blood pressure of reduced dietary sodium and Associationbetweenplasmaviscosityandall-cause nolysis and lipid profile. Blood Coagul Fibrino- the Dietary Approaches to Stop Hypertension mortality: results from the MONICA-Augsburg lysis. 1996;7:484-490. (DASH) diet. N Engl J Med. 2001;344:3-10. Cohort Study 1984-92. Br J Haematol. 2000;109: 45. van den Burg PJ, Hospers JE, van Vliet M, Mo- 12. Dietary supplementation with n-3 polyunsatu- 453-458. sterd WL, Bouma BN, Huisveld IA. Effect of en- rated fatty acids and vitamin E after myocardial 29. Sweetnam PM, Thomas HF, Yarnell JW, Be- durance training and seasonal fluctuation on co- infarction: results of the GISSI-Prevenzione trial. swick AD, Baker IA, Elwood PC. Fibrinogen, vis- agulation and fibrinolysis in young sedentary Lancet. 1999;354:447-455. cosity and the 10-year incidence of ischaemic men. J Appl Physiol. 1997;82:613-620. 13. Burr ML, Fehily AM, Gilbert JF, et al. Effects of heart disease. Eur Heart J. 1996;17:1814- 46. Stratton JR, Chandler WL, Schwartz RS, et al. changes in fat, fish, and fibre intakes on death 1820. Effects of physical conditioning on fibrinolytic and myocardial reinfarction: diet and reinfarc- 30. Lowe GD, Lee AJ, Rumley A, Price JF, Fowkes variables and fibrinogen in young and old healthy tion trial (DART). Lancet. 1989;2:757-761. FG. Blood viscosity and risk of cardiovascular adults. Circulation. 1991;83:1692-1697. 14. Marchioli R, Barzi F, Bomba E, et al. Early pro- events: the Edinburgh Artery Study. Br J Hae- 47. Schuit AJ, Schouten EG, Kluft C, de Maat M, Men- tection against sudden death by n-3 polyunsat- matol. 1997;96:168-173. heere PP, Kok FJ. Effect of strenuous exercise on urated fatty acids after myocardial infarction: 31. Thogersen AM, Jansson JH, Boman K, et al. High fibrinogen and fibrinolysis in healthy elderly men time-course analysis of the results of the Gruppo plasminogen activator inhibitor and tissue plas- and women. Thromb Haemost. 1997;78:845- Italiano per lo Studio della Sopravvivenza minogen activator levels in plasma precede a first 851. nell’Infarto Miocardico (GISSI)-Prevenzione. Cir- acute myocardial infarction in both men and 48. Ponjee GA, Janssen GM, van Wersch JW. Pro- culation. 2002;105:1897-1903. women: evidence for the fibrinolytic system as longed endurance exercise and blood coagula- 15. Davies MJ. The investigation of sudden cardiac an independent primary risk factor. Circulation. tion:a9month prospective study. Blood Co- death. Histopathology. 1999;34:93-98. 1998;98:2241-2247. agul Fibrinolysis. 1993;4:21-25. 16. Wiman B. Predictive value of fibrinolytic factors 32. Nordt TK, Peter K, Ruef J, Kubler W, Bode C. Plas- 49. Vaisanen SB, Humphries SE, Luong LA, Pent- in coronary heart disease. Scand J Clin Lab In- minogen activator inhibitor type-1 (PAI-1) and tila I, Bouchard C, Rauramaa R. Regular exer- vest Suppl. 1999;230:23-31. its role in cardiovascular disease. Thromb Hae- cise, plasminogen activator inhibitor-1 (PAI-1) 17. Lip GY. Fibrinogen and cardiovascular disor- most. 1999;82(suppl 1):14-18. activity and the 4G/5G promoter polymorphism ders. Q J Med. 1995;88:155-165. 33. Juhan-Vague I, Alessi MC. Fibrinolysis and risk in the PAI-1 gene. Thromb Haemost. 1999;82: 18. Lowe GD, Rumley A, Sweetnam PM, Yarnell JW, of coronary artery disease. Fibrinolysis. 1996; 1117-1120. Rumley J. Fibrin D-dimer, markers of coagula- 10:127-136. 50. Hegde SS, Goldfarb AH, Hegde S. Clotting and tion activation and the risk of major ischaemic 34. Meade TW, Ruddock V, Stirling Y, Chakrabarti fibrinolytic activity change during the 1 h after a heart disease in the caerphilly study. Thromb Hae- R, Miller GJ. Fibrinolytic activity, clotting fac- submaximal run. Med Sci Sports Exerc. 2001; most. 2001;86:822-827. tors, and long-term incidence of ischaemic heart 33:887-892. 19. Rumley A, Lowe GD, Sweetnam PM, Yarnell JW, disease in the Northwick Park Heart Study. Lan- 51. Lin X, El Sayed MS, Waterhouse J, Reilly T. Ac- Ford RP. Factor VIII, von Willebrand factor and cet. 1993;342:1076-1079. tivation and disturbance of blood haemostasis the risk of major ischaemic heart disease in the 35. Wallen NH, Goodall AH, Li N, Hjemdahl P. Acti- following strenuous physical exercise. Int J Caerphilly Heart Study. Br J Haematol. 1999; vation of haemostasis by exercise, mental stress Sports Med. 1999;20:149-153. 105:110-116. and adrenaline: effects on platelet sensitivity to 52. Rankinen T, Vaisanen S, Penttila I, Rauramaa R. 20. Smith FB, Lee AJ, Fowkes FG, Price JF, Rumley thrombin and thrombin generation. Clin Sci Acute dynamic exercise increases fibrinolytic ac- A, Lowe GD. Hemostatic factors as predictors of (Lond). 1999;97:27-35. tivity. Thromb Haemost. 1995;73:281-286. ischemic heart disease and stroke in the Edin- 36. von Kanel R, Dimsdale JE. Effects of sympa- 53. Prisco D, Paniccia R, Bandinelli B, et al. Evalu- burgh Artery Study. Arterioscler Thromb Vasc thetic activation by adrenergic infusions on he- ation of clotting and fibrinolytic activation after Biol. 1997;17:3321-3325. mostasis in vivo. Eur J Haematol. 2000;65:357- protracted physical exercise. Thromb Res. 1998; 21. Meade TW, Mellows S, Brozovic M, et al. Hae- 369. 89:73-78. mostatic function and ischaemic heart disease: 37. Wallen NH, Held C, Rehnqvist N, Hjemdahl P. Ef- 54. Cerneca F, Crocetti G, Gombacci A, Simeone R, principal results of the Northwick Park Heart fects of mental and physical stress on platelet Tamaro G, Mangiarotti MA. Variations in hemo- Study. Lancet. 1986;2:533-537. function in patients with stable angina pectoris static parameters after near-maximum exercise 22. Ridker PM, Vaughan DE, Stampfer MJ, Manson and healthy controls. Eur Heart J. 1997;18:807- and specific tests in athletes. J Sports Med Phys JE, Hennekens CH. Endogenous tissue-type plas- 815. Fitness. 1999;39:31-36.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2387

Downloaded From: https://jamanetwork.com/ on 09/23/2021 55. Watts EJ. Haemostatic changes in long- Nuno MA, Gonzalez-Gallego J. Changes in the bin generation after physical exercise. Thromb distance runners and their relevance to the pre- fibrinolytic system associated with physical con- Res. 1993;69:159-164. vention of ischaemic heart disease. Blood Co- ditioning. Eur J Appl Physiol Occup Physiol. 1992; 90. Weiss C, Welsch B, Albert M, et al. Coagulation agul Fibrinolysis. 1991;2:221-225. 65:388-393. and thrombomodulin in response to exercise of 56. Wang JS, Jen CJ, Chen HI. Effects of exercise 73. Szymanski LM, Pate RR, Durstine JL. Effects of different type and duration. Med Sci Sports Ex- training and deconditioning on platelet function maximal exercise and venous occlusion on fi- erc. 1998;30:1205-1210. in men. Arterioscler Thromb Vasc Biol. 1995; brinolytic activity in physically active and inac- 91. Weiss C, Seitel G, Bartsch P. Coagulation and fi- 15:1668-1674. tive men. J Appl Physiol. 1994;77:2305-2310. brinolysis after moderate and very heavy exer- 57. Li N, Wallen NH, Hjemdahl P. Evidence for pro- 74. De Paz JA, Lasierra J, Villa JG, Vilades E, Gonza- cise in healthy male subjects. Med Sci Sports Ex- thrombotic effects of exercise and limited pro- lez-Gallego J. Effects of aerobic and anaerobic erc. 1998;30:246-251. tection by aspirin. Circulation. 1999;100:1374- physical conditioning on fibrinolysis. J Sports 92. Weiss C, Velich T, Niebauer J, et al. Activation 1379. Med Phys Fitness. 1995;35:263-267. of coagulation and fibrinolysis after rehabilita- 58. Andreotti F, Lanza GA, Sciahbasi A, et al. Low- 75. Gris JC, Schved JF, Feugeas O, et al. Impact of tive exercise in patients with coronary artery dis- grade exercise enhances platelet aggregability in smoking, physical training and weight reduc- ease. Am J Cardiol. 1998;81:672-677. patients with obstructive coronary disease in- tion on FVII, PAI-1 and hemostatic markers in 93. Bartsch P, Welsch B, Albert M, Friedmann B, Levi dependently of myocardial ischemia. Am J Car- sedentary men. Thromb Haemost. 1990;64:516- M, Kruithof EK. Balanced activation of coagula- diol. 2001;87:16-20. 520. tion and fibrinolysis after a 2-h triathlon. Med Sci 59. Mustonen P, Lepantalo M, Lassila R. Physical 76. Lindahl B, Nilsson TK, Jansson JH, Asplund K, Sports Exerc. 1995;27:1465-1470. exertion induces thrombin formation and fibrin Hallmans G. Improved fibrinolysis by intense life- 94. Grant PJ. Hormonal regulation of the acute hae- degradation in patients with peripheral athero- style intervention: a randomized trial in sub- mostatic response to stress. Blood Coagul Fi- sclerosis. Arterioscler Thromb Vasc Biol. 1998; jects with impaired glucose tolerance. J Intern brinolysis. 1990;1:299-306. 18:244-249. Med. 1999;246:105-112. 95. Chandler WL, Veith RC, Fellingham GW, et al. Fi- 60. Li-Saw-Hee FL, Blann AD, Edmunds E, Gibbs CR, 77. Speiser W, Langer W, Pschaick A, et al. In- brinolytic response during exercise and epineph- Lip GY. Effect of acute exercise on the raised creased blood fibrinolytic activity after physical rine infusion in the same subjects. J Am Coll Car- plasma fibrinogen, soluble P-selectin and von exercise: comparative study in individuals with diol. 1992;19:1412-1420. Willebrand factor levels in chronic atrial fibrilla- different sporting activities and in patients after 96. El Sayed MS, Davies B. Effect of two formula- tion. Clin Cardiol. 2001;24:409-414. myocardial infarction taking part in a rehabilita- tions of a beta blocker on fibrinolytic response 61. Gibbs CR, Blann AD, Edmunds E, Watson RD, tion sports program. Thromb Res. 1988;51:543- to maximum exercise. Med Sci Sports Exerc. Lip GY. Effects of acute exercise on hemorheo- 555. 1989;21:369-373. logical, endothelial, and platelet markers in pa- 78. van den Burg PJ, Hospers JE, Mosterd WL, 97. Davis RB, Boyd DG, McKinney ME, Jones CC. tients with chronic heart failure in sinus rhythm. Bouma BN, Huisveld IA. Aging, physical condi- Effects of exercise and exercise conditioning on Clin Cardiol. 2001;24:724-729. tioning, and exercise-induced changes in hemo- blood platelet function. Med Sci Sports Exerc. 62. Eliasson M, Asplund K, Evrin PE. Regular lei- static factors and reaction products. J Appl 1990;22:49-53. sure time physical activity predicts high activity Physiol. 2000;88:1558-1564. 98. Rauramaa R, Salonen JT, Seppanen K, et al. In- of tissue plasminogen activator: The Northern 79. Rock G, Tittley P, Pipe A. Coagulation factor hibition of platelet aggregability by moderate- Sweden MONICA Study. Int J Epidemiol. 1996; changes following endurance exercise. Clin J intensity physical exercise: a randomized clini- 25:1182-1188. Sport Med. 1997;7:94-99. cal trial in overweight men. Circulation. 1986; 63. El Sayed MS. Effects of exercise on blood co- 80. Hansen JB, Wilsgard L, Olsen JO, Osterud B. For- 74:939-944. agulation, fibrinolysis and platelet aggregation. mation and persistence of procoagulant and fi- 99. Kestin AS, Ellis PA, Barnard MR, Errichetti A, Ros- Sports Med. 1996;22:282-298. brinolytic activities in circulation after strenu- ner BA, Michelson AD. Effect of strenuous ex- 64. Rydzewski A, Sakata K, Kobayashi A, et al. ous physical exercise. Thromb Haemost. 1990; ercise on platelet activation state and reactivity. Changes in plasminogen activator inhibitor 1 and 64:385-389. Circulation. 1993;88:1502-1511. tissue-type plasminogen activator during exer- 81. Boman K, Hellsten G, Bruce A, Hallmans G, Nils- 100. Wang JS, Jen CJ, Kung HC, Lin LJ, Hsiue TR, cise in patients with coronary artery disease. Hae- son TK. Endurance physical activity, diet and fi- Chen HI. Different effects of strenuous exercise mostasis. 1990;20:305-312. brinolysis. Atherosclerosis. 1994;106:65-74. and moderate exercise on platelet function in 65. Imhof A, Koenig W. Exercise and thrombosis. 82. Suzuki T, Yamauchi K, Yamada Y, et al. Blood men. Circulation. 1994;90:2877-2885. Cardiol Clin. 2001;19:389-400. coagulability and fibrinolytic activity before and 101. Wang JS, Jen CJ, Chen HI. Effects of chronic ex- 66. Szymanski LM, Pate RR. Effects of exercise in- after physical training during the recovery phase ercise and deconditioning on platelet function in tensity, duration, and time of day on fibrinolytic of acute myocardial infarction. Clin Cardiol. 1992; women. J Appl Physiol. 1997;83:2080-2085. activity in physically active men. Med Sci Sports 15:358-364. 102. Bartsch P. Platelet activation with exercise and Exerc. 1994;26:1102-1108. 83. Wosornu D, Allardyce W, Ballantyne D, Tansey risk of cardiac events. Lancet. 1999;354:1747- 67. Fernhall B, Szymanski LM, Gorman PA, Milani P. Influence of power and aerobic exercise train- 1748. J, Paup DC, Kessler CM. Fibrinolytic activity is ing on haemostatic factors after coronary ar- 103. Streiff M, Bell WR. Exercise and hemostasis in similar in physically active men with and with- tery surgery. Br Heart J. 1992;68:181-186. humans. Semin Hematol. 1994;31:155-165. out a history of myocardial infarction. Arterio- 84. Vanninen E, Laitinen J, Uusitupa M. Physical ac- 104. Rao SV, Donahue M, Pi-Sunyer FX, Fuster V. Re- scler Thromb Vasc Biol. 1997;17:1106-1113. tivity and fibrinogen concentration in newly di- sults of expert meetings: obesity and cardiovas- 68. DeSouza CA, Dengel DR, Rogers MA, Cox K, agnosed NIDDM. Diabetes Care. 1994;17:1031- cular disease: obesity as a risk factor in coronary Macko RF. Fibrinolytic responses to acute physi- 1038. artery disease. Am Heart J. 2001;142:1102-1107. cal activity in older hypertensive men. J Appl 85. Ernst E. Regular exercise reduces fibrinogen lev- 105. Vega GL. Results of expert meetings: obesity and Physiol. 1997;82:1765-1770. els: a review of longitudinal studies. Br J Sports cardiovascular disease: obesity, the metabolic 69. Arai M, Yorifuji H, Ikematsu S, et al. Influences Med. 1993;27:175-176. syndrome, and cardiovascular disease. Am Heart of strenuous exercise (triathlon) on blood co- 86. Gill JM, Frayn KN, Wootton SA, Miller GJ, Hard- J. 2001;142:1108-1116. agulation and fibrinolytic system. Thromb Res. man AE. Effects of prior moderate exercise on 106. Van Gaal LF, Wauters MA, De Leeuw IH. The ben- 1990;57:465-471. exogenous and endogenous lipid metabolism and eficial effects of modest weight loss on cardio- 70. Montgomery HE, Clarkson P, Nwose OM, et plasma factor VII activity. Clin Sci (Lond). 2001; vascular risk factors. Int J Obes Relat Metab Dis- al. The acute rise in plasma fibrinogen con- 100:517-527. ord. 1997;21(suppl 1):S5-S9. centration with exercise is influenced by the 87. van den Burg PJ, Hospers JE, van Vliet M, Mo- 107. Yarnell JW, Sweetnam PM, Rumley A, Lowe GD. G-453-A polymorphism of the beta-fibrinogen sterd WL, Huisveld IA. Unbalanced haemo- Lifestyle and hemostatic risk factors for ische- gene. Arterioscler Thromb Vasc Biol. 1996;16: static changes following strenuous physical ex- mic heart disease: the Caerphilly Study. Arterio- 386-391. ercise: a study in young sedentary males. Eur scler Thromb Vasc Biol. 2000;20:271-279. 71. Bartsch P, Haeberli A, Straub PW. Blood coagu- Heart J. 1995;16:1995-2001. 108. Landin K, Stigendal L, Eriksson E, et al. Abdomi- lation after long distance running: antithrombin 88. Connelly JB, Cooper JA, Meade TW. Strenuous nal obesity is associated with an impaired fibri- III prevents fibrin formation. Thromb Haemost. exercise, plasma fibrinogen, and factor VII ac- nolytic activity and elevated plasminogen acti- 1990;63:430-434. tivity. Br Heart J. 1992;67:351-354. vator inhibitor-1. Metabolism. 1990;39:1044- 72. De Paz JA, Lasierra J, Villa JG, Vilades E, Martin- 89. Prisco D, Paniccia R, Guarnaccia V, et al. Throm- 1048.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2388

Downloaded From: https://jamanetwork.com/ on 09/23/2021 109. Eliasson M, Evrin PE, Lundblad D. Fibrinogen and 125. Juhan-Vague I, Vague P, Alessi MC, et al. Rela- 141. Boberg M, Pollare T, Siegbahn A, Vessby B. fibrinolytic variables in relation to anthropom- tionships between plasma insulin triglyceride, Supplementation with n-3 fatty acids reduces tri- etry, lipids and blood pressure: The Northern body mass index, and plasminogen activator in- glycerides but increases PAI-1 in non-insulin- Sweden MONICA Study. J Clin Epidemiol. 1994; hibitor 1. Diabetes Metab. 1987;13:331-336. dependent diabetes mellitus. Eur J Clin Invest. 47:513-524. 126. Mehta J, Mehta P, Lawson D, Saldeen T. Plasma 1992;22:645-650. 110. Ernst E, Resch KL. Fibrinogen as a cardiovas- tissue plasminogen activator inhibitor levels in 142. Radack K, Deck C, Huster G. The comparative cular risk factor: a meta-analysis and review of coronary artery disease: correlation with age and effects of ␻-3 and n-6 polyunsaturated fatty ac- the literature. Ann Intern Med. 1993;118:956- serum triglyceride concentrations. JAmCollCar- ids on plasma fibrinogen levels: a controlled clini- 963. diol. 1987;9:263-268. cal trial in hypertriglyceridemic subjects. JAm 111. Ernst E, Resch KL. Therapeutic interventions to 127. Elkeles RS, Chakrabarti R, Vickers M, Stirling Y, Coll Nutr. 1990;9:352-357. lower plasma fibrinogen concentration. Eur Heart Meade TW. Effect of treatment of hyperlipidae- 143. Prisco D, Paniccia R, Filippini M, et al. No changes J. 1995;16(suppl A):47-52. mia on haemostatic variables. BMJ. 1980;281: in PAI-1 levels after four-month n-3 PUFA ethyl 112. Charles MA, Morange P, Eschwege E, Andre P, 973-974. ester supplementation in healthy subjects. Vague P, Juhan-Vague I. Effect of weight change 128. Andersen P, Smith P, Seljeflot I, Brataker S, Ar- Thromb Res. 1994;76:237-244. and metformin on fibrinolysis and the von nesen H. Effects of gemfibrozil on lipids and hae- 144. Myrup B, Rossing P, Jensen T, et al. Lack of effect Willebrand factor in obese nondiabetic sub- mostasis after myocardial infarction. Thromb of fish oil supplementation on coagulation and jects: the BIGPRO1 Study. Diabetes Care. 1998; Haemost. 1990;63:174-177. transcapillary escape rate of albumin in insulin- 21:1967-1972. 129. Executive Summary of The Third Report of The dependent diabetic patients with diabetic ne- 113. Rissanen P, Vahtera E, Krusius T, Uusitupa M, National Cholesterol Education Program (NCEP) phropathy. Scand J Clin Lab Invest. 2001;61: Rissanen A. Weight change and blood coagula- Expert Panel on Detection, Evaluation, and Treat- 349-356. bility and fibrinolysis in healthy obese women. ment of High Blood Cholesterol in Adults (Adult 145. Toft I, Bonaa KH, Ingebretsen OC, Nordoy A, Int J Obes Relat Metab Disord. 2001;25:212- Treatment Panel III). JAMA. 2001;285:2486- Jenssen T. Fibrinolytic function after dietary 218. 2497. supplementation with omega3 polyunsaturated 114. Marckmann P, Toubro S, Astrup A. Sustained im- 130. de Lorgeril M, Salen P, Martin JL, Monjaud I, fatty acids. Arterioscler Thromb Vasc Biol. 1997; provement in blood lipids, coagulation, and fi- Boucher P, Mamelle N. Mediterranean dietary pat- 17:814-819. brinolysis after major weight loss in obese sub- tern in a randomized trial: prolonged survival and 146. Hansen J, Grimsgaard S, Nordoy A, Bonaa KH. jects. Eur J Clin Nutr. 1998;52:329-333. possible reduced cancer rate. Arch Intern Med. Dietary supplementation with highly purified 115. Folsom AR, Qamhieh HT, Wing RR, et al. Im- 1998;158:1181-1187. eicosapentaenoic acid and docosahexaenoic acid pact of weight loss on plasminogen activator in- 131. Kris-Etherton P, Eckel RH, Howard BV, St Jeor does not influence PAI-1 activity. Thromb Res. hibitor (PAI-1), factor VII, and other hemostatic S, Bazzarre TL. AHA science advisory: Lyon Diet 2000;98:123-132. factors in moderately overweight adults. Arte- Heart Study: benefits of a Mediterranean-style, 147. Nilsen DW, Dalaker K, Nordoy A, et al. Influ- rioscler Thromb. 1993;13:162-169. National Cholesterol Education Program/ ence of a concentrated ethylester compound of 116. Primrose JN, Davies JA, Prentice CR, Hughes R, American Heart Association step I dietary pat- n-3 fatty acids on lipids, platelets and coagula- Johnston D. Reduction in factor VII, fibrinogen tern on cardiovascular disease. Circulation. 2001; tion in patients undergoing coronary bypass sur- and plasminogen activator inhibitor-1 activity af- 103:1823-1825. gery. Thromb Haemost. 1991;66:195-201. ter surgical treatment of morbid obesity. Thromb 132. Kristensen SD, Iversen AM, Schmidt EB. n-3 poly- 148. Hellsten G, Boman K, Saarem K, Hallmans G, Nils- Haemost. 1992;68:396-399. unsaturated fatty acids and coronary thrombo- son TK. Effects on fibrinolytic activity of corn oil 117. Berntorp E, Berntorp K, Brorson H, Frick K. Li- sis. Lipids. 2001;36(suppl):S79-S82. and a fish oil preparation enriched with omega- posuction in Dercum’s disease: impact on hae- 133. Hornstra G. Influence of dietary fat type on ar- 3-polyunsaturated fatty acids in a long-term mostatic factors associated with cardiovascu- terial thrombosis tendency. J Nutr Health Ag- study. Curr Med Res Opin. 1993;13:133-139. lar disease and insulin sensitivity. J Intern Med. ing. 2001;5:160-166. 149. Rogers S, James KS, Butland BK, Etherington 1998;243:197-201. 134. Hoak JC. Fatty acids in animals: thrombosis and MD, O’Brien JR, Jones JG. Effects of a fish oil 118. Sudi KM, Gallistl S, Trobinger M, et al. The in- hemostasis. Am J Clin Nutr. 1997;65:1683S- supplement on serum lipids, blood pressure, fluence of weight loss on fibrinolytic and meta- 1686S. bleeding time, haemostatic and rheological vari- bolic parameters in obese children and adoles- 135. Nobukata H, Ishikawa T, Obata M, Shibutani Y. ables: a double blind randomised controlled trial cents. J Pediatr Endocrinol Metab. 2001;14: Long-term administration of highly purified in healthy volunteers. Atherosclerosis. 1987;63: 85-94. eicosapentaenoic acid ethyl ester prevents dia- 137-143. 119. Mavri A, Stegnar M, Krebs M, Sentocnik JT, Gei- betes and abnormalities of blood coagulation in 150. Li XL, Steiner M. Fish oil: a potent inhibitor of ger M, Binder BR. Impact of adipose tissue on male WBN/Kob rats. Metabolism. 2000;49:912- platelet adhesiveness. Blood. 1990;76:938- plasma plasminogen activator inhibitor-1 in di- 919. 945. eting obese women. Arterioscler Thromb Vasc 136. Archer SL, Green D, Chamberlain M, Dyer AR, 151. Andrioli G, Carletto A, Guarini P, et al. Differen- Biol. 1999;19:1582-1587. Liu K. Association of dietary fish and n-3 fatty tial effects of dietary supplementation with fish 120. Charles MA, Morange P, Eschwege E, Andre P, acid intake with hemostatic factors in the coro- oil or soy lecithin on human platelet adhesion. Vague P, Juhan-Vague I. Effect of weight change nary artery risk development in young adults Thromb Haemost. 1999;82:1522-1527. and metformin on fibrinolysis and the von (CARDIA) study. Arterioscler Thromb Vasc Biol. 152. Freese R, Mutanen M. Alpha-linolenic acid and ma- Willebrand factor in obese nondiabetic sub- 1998;18:1119-1123. rine long-chain n-3 fatty acids differ only slightly jects: the BIGPRO1 Study. Diabetes Care. 1998; 137. Scarabin PY, Aillaud MF, Luc G, et al. Haemo- in their effects on hemostatic factors in healthy sub- 21:1967-1972. stasis in relation to dietary fat as estimated by jects. Am J Clin Nutr. 1997;66:591-598. 121. Calles-Escandon J, Ballor D, Harvey-Berino J, erythrocyte fatty acid composition: the PRIME 153. Barcelli U, Glas-Greenwalt P, Pollak VE. Enhanc- Ades P, Tracy R, Sobel B. Amelioration of the in- study. Thromb Res. 2001;102:285-293. ing effect of dietary supplementation with hibition of fibrinolysis in elderly, obese sub- 138. Shahar E, Folsom AR, Wu KK, et al. Associa- omega-3 fatty acids on plasma fibrinolysis in nor- jects by moderate energy intake restriction. Am tions of fish intake and dietary n-3 polyunsatu- mal subjects. Thromb Res. 1985;39:307-312. J Clin Nutr. 1996;64:7-11. rated fatty acids with a hypocoagulable profile: 154. Mehta J, Lawson D, Saldeen TJ. Reduction in 122. Loskutoff DJ, Samad F. The adipocyte and he- the Atherosclerosis Risk in Communities (ARIC) plasminogen activator inhibitor-1 (PAI-1) with mostatic balance in obesity: studies of PAI-1. Ar- Study. Arterioscler Thromb. 1993;13:1205- omega-3 polyunsaturated fatty acid (PUFA) in- terioscler Thromb Vasc Biol. 1998;18:1-6. 1212. take. Am Heart J. 1988;116:1201-1206. 123. Lundgren CH, Brown SL, Nordt TK, Sobel BE, Fu- 139. Iacoviello L, Amore C, De Curtis A, et al. Modu- 155. Schmidt EB, Varming K, Ernst E, Madsen P, Dy- jii S. Elaboration of type-1 plasminogen activa- lation of fibrinolytic response to venous occlu- erberg J. Dose-response studies on the effect of tor inhibitor from adipocytes: a potential patho- sion in humans by a combination of low-dose n-3 polyunsaturated fatty acids on lipids and hae- genetic link between obesity and cardiovascular aspirin and n-3 polyunsaturated fatty acids. Ar- mostasis. Thromb Haemost. 1990;63:1-5. disease. Circulation. 1996;93:106-110. terioscler Thromb. 1992;12:1191-1197. 156. Miller ME, Anagnostou AA, Ley B, Marshall P, 124. Hamsten A, De Faire U, Walldius G, et al. Plas- 140. Oosthuizen W, Vorster HH, Jerling JC, et al. Both Steiner M. Effect of fish oil concentrates on minogen activator inhibitor in plasma: risk fac- fish oil and olive oil lowered plasma fibrinogen hemorheological and hemostatic aspects of dia- tor for recurrent myocardial infarction. Lancet. in women with high baseline fibrinogen levels. betes mellitus: a preliminary study. Thromb Res. 1987;2:3-9. Thromb Haemost. 1994;72:557-562. 1987;47:201-214.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2389

Downloaded From: https://jamanetwork.com/ on 09/23/2021 157. Schmidt EB, Lervang HH, Varming K, Madsen MC, Bas BM, Hornstra G. Effect of a moderate rated fatty acids and human platelet function: a P, Dyerberg J. Long-term supplementation with fish intake on haemostatic parameters in healthy review with particular emphasis on implica- n-3 fatty acids, I: effect on blood lipids, haemo- males. Thromb Haemost. 1989;61:468-473. tions for cardiovascular disease. J Intern Med stasis and blood pressure. Scand J Clin Lab In- 173. Sanders TA, Oakley FR, Miller GJ, Mitropoulos Suppl. 1989;225:141-150. vest. 1992;52:221-228. KA, Crook D, Oliver MF. Influence of n-6 versus 188. Nelson GJ, Schmidt PS, Bartolini GL, Kelley DS, 158. Smith P, Arnesen H, Opstad T, Dahl KH, Eritsland n-3 polyunsaturated fatty acids in diets low in Kyle D. The effect of dietary docosahexaenoic acid J. Influence of highly concentrated n-3 fatty ac- saturated fatty acids on plasma lipoproteins and on platelet function, platelet fatty acid compo- ids on serum lipids and hemostatic variables in hemostatic factors. Arterioscler Thromb Vasc sition, and blood coagulation in humans. Lip- survivors of myocardial infarction receiving ei- Biol. 1997;17:3449-3460. ids. 1997;32:1129-1136. ther oral anticoagulants or matching placebo. 174. Conquer JA, Cheryk LA, Chan E, Gentry PA, Holub 189. Mezzano D, Kosiel K, Martinez C, et al. Cardio- Thromb Res. 1989;53:467-474. BJ. Effect of supplementation with dietary seal vascular risk factors in vegetarians: normaliza- 159. Li D, Sinclair A, Wilson A, et al. Effect of dietary oil on selected cardiovascular risk factors and tion of hyperhomocysteinemia with vitamin B(12) alpha-linolenic acid on thrombotic risk factors hemostatic variables in healthy male subjects. and reduction of platelet aggregation with n-3 in vegetarian men. Am J Clin Nutr. 1999;69:872- Thromb Res. 1999;96:239-250. fatty acids. Thromb Res. 2000;100:153-160. 882. 175. Berrettini M, Parise P, Ricotta S, Iorio A, Pei- 190. Pirich C, Gaszo A, Granegger S, Sinzinger H. Ef- 160. Emeis JJ, van Houwelingen AC, van den Hoo- rone C, Nenci GG. Increased plasma levels of tis- fects of fish oil supplementation on platelet sur- gen CM, Hornstra G. A moderate fish intake in- sue factor pathway inhibitor (TFPI) after n-3 poly- vival and ex vivo platelet function in hypercho- creases plasminogen activator inhibitor type-1 unsaturated fatty acids supplementation in lesterolemic patients. Thromb Res. 1999;96: in human volunteers. Blood. 1989;74:233-237. patients with chronic atherosclerotic disease. 219-227. 161. Nordoy A, Hatcher L, Goodnight S, FitzGerald GA, Thromb Haemost. 1996;75:395-400. 191. Vericel E, Calzada C, Chapuy P, Lagarde M. The Conner WE. Effects of dietary fat content, satu- 176. Hansen JB, Olsen JO, Wilsgard L, Osterud B. Ef- influence of low intake of n-3 fatty acids on plate- rated fatty acids, and fish oil on eicosanoid pro- fects of dietary supplementation with cod liver lets in elderly people. Atherosclerosis. 1999;147: duction and hemostatic parameters in normal oil on monocyte thromboplastin synthesis, co- 187-192. men. J Lab Clin Med. 1994;123:914-920. agulation and fibrinolysis. J Intern Med Suppl. 192. Mori TA, Beilin LJ, Burke V, Morris J, Ritchie J. 162. Saynor R, Gillott T. Changes in blood lipids and 1989;225:133-139. Interactions between dietary fat, fish, and fish oils fibrinogen with a note on safety in a long term 177. Marckmann P, Bladbjerg EM, Jespersen J. Di- and their effects on platelet function in men at study on the effects of n-3 fatty acids in sub- etary fish oil (4 g daily) and cardiovascular risk risk of cardiovascular disease. Arterioscler jects receiving fish oil supplements and fol- markers in healthy men. Arterioscler Thromb Thromb Vasc Biol. 1997;17:279-286. lowed for seven years. Lipids. 1992;27:533- Vasc Biol. 1997;17:3384-3391. 193. Prisco D, Filippini M, Francalanci I, Paniccia R, 538. 178. Schmidt EB, Nielsen LK, Pedersen JO, Kor- Gensini GF, Serneri GG. Effect of ␻-3 fatty acid 163. Goodnight SH Jr, Harris WS, Connor WE. The nerup HJ, Dyerberg J. The effect of n-3 polyun- ethyl ester supplementation on fatty acid com- effects of dietary omega 3 fatty acids on plate- saturated fatty acids on lipids, platelet function, position of the single platelet phospholipids and let composition and function in man: a prospec- coagulation, fibrinolysis and monocyte chemo- on platelet functions. Metabolism. 1995;44:562- tive, controlled study. Blood. 1981;58:880- taxis in patients with hypertension. Clin Chim 569. 885. Acta. 1990;189:25-32. 194. Tremoli E, Maderna P, Marangoni F, et al. Pro- 164. Solomon SA, Cartwright I, Pockley G, et al. A pla- 179. Tremoli E, Eligini S, Colli S, et al. n-3 fatty acid longed inhibition of platelet aggregation after n-3 cebo-controlled, double-blind study of eicosa- ethyl ester administration to healthy subjects and fatty acid ethyl ester ingestion by healthy volun- pentaenoic acid-rich fish oil in patients with stable to hypertriglyceridemic patients reduces tissue teers. Am J Clin Nutr. 1995;61:607-613. angina pectoris. Curr Med Res Opin. 1990;12: factor activity in adherent monocytes. Arterio- 195. Mutanen M, Freese R. Polyunsaturated fatty ac- 1-11. scler Thromb. 1994;14:1600-1608. ids and platelet aggregation. Curr Opin Lipidol. 165. Bach R, Schmidt U, Jung F, et al. Effects of fish 180. Takimoto G, Galang J, Lee GK, Bradlow BA. 1996;7:14-19. oil capsules in two dosages on blood pressure, Plasma fibrinolytic activity after ingestion of 196. Owens MR, Cave WT Jr. Dietary fish lipids do platelet functions, haemorheological and clini- omega-3 fatty acids in human subjects. Thromb not diminish platelet adhesion to subendothe- cal chemistry parameters in apparently healthy Res. 1989;54:573-582. lium. Br J Haematol. 1990;75:82-85. subjects. Ann Nutr Metab. 1989;33:359-367. 181. Marckmann P, Sandstrom B, Jespersen J. Fast- 197. Berg KJ, Skaga E, Skjaeggestad O, Stormorken 166. Cartwright IJ, Pockley AG, Galloway JH, Greaves ing blood coagulation and fibrinolysis of young H. Effect of linseed oil on platelet adhesiveness M, Preston FE. The effects of dietary omega-3 adults unchanged by reduction in dietary fat con- and bleeding-time in patients with coronary heart- polyunsaturated fatty acids on erythrocyte mem- tent. Arterioscler Thromb. 1992;12:201-205. disease. Lancet. 1965;2:980-982. brane phospholipids, erythrocyte deformability 182. Eritsland J, Arnesen H, Seljeflot I, Kierulf P. Long- 198. Natvig H, Borchgrevink CF, Dedichen J, Owren and blood viscosity in healthy volunteers. Ath- term effects of n-3 polyunsaturated fatty acids PA, Schiotz EH, Westlund K. A controlled trial of erosclerosis. 1985;55:267-281. on haemostatic variables and bleeding epi- the effect of linolenic acid on incidence of coro- 167. Blonk MC, Bilo HJ, Nauta JJ, Popp-Snijders C, sodes in patients with coronary artery disease. nary heart disease: the Norwegian vegetable oil Mulder C, Donker AJ. Dose-response effects of Blood Coagul Fibrinolysis. 1995;6:17-22. experiment of 1965-66. Scand J Clin Lab Invest fish-oil supplementation in healthy volunteers. 183. Marckmann P, Jespersen J, Leth T, Sandstrom Suppl. 1968;105:1-20. Am J Clin Nutr. 1990;52:120-127. B. Effect of fish diet versus meat diet on blood 199. Christensen JH, Schmidt EB. n-3 fatty acids and 168. Harris WS, Windsor SL, Dujovne CA. Effects of lipids, coagulation and fibrinolysis in healthy the risk of sudden cardiac death. Lipids. 2001; four doses of n-3 fatty acids given to hyperlip- young men. J Intern Med. 1991;229:317-323. 36(suppl):S115-S118. idemic patients for six months. J Am Coll Nutr. 184. Dunstan DW, Mori TA, Puddey IB, et al. A ran- 200. Siscovick DS, Raghunathan T, King I, et al. Di- 1991;10:220-227. domised, controlled study of the effects of aero- etary intake of long-chain n-3 polyunsaturated 169. Rillaerts EG, Engelmann GJ, Van Camp KM, De bic exercise and dietary fish on coagulation and fatty acids and the risk of primary cardiac ar- Leeuw I. Effect of omega-3 fatty acids in diet of fibrinolytic factors in type 2 diabetics. Thromb rest. Am J Clin Nutr. 2000;71(1 suppl):208S- type I diabetic subjects on lipid values and Haemost. 1999;81:367-372. 212S. hemorheological parameters. Diabetes. 1989; 185. Armstrong RA, Chardigny JM, Beaufrere B, et al. 201. Connor SL, Connor WE. Are fish oils beneficial 38:1412-1416. No effect of dietary trans isomers of alpha- in the prevention and treatment of coronary ar- 170. Nelson GJ, Schmidt PC, Corash L. The effect of linolenic acid on platelet aggregation and hae- tery disease? Am J Clin Nutr. 1997;66(4 suppl): a salmon diet on blood clotting, platelet aggre- mostatic factors in european healthy men: the 1020S-1031S. gation and fatty acids in normal adult men. Lip- TRANSLinE study. Thromb Res. 2000;100:133- 202. Beaglehole R, Jackson R. Alcohol, cardiovascu- ids. 1991;26:87-96. 141. lar diseases and total mortality: the epidemiologi- 171. Allman-Farinelli MA, Hall D, Kingham K, Pang D, 186. Almendingen K, Seljeflot I, Sandstad B, Peder- cal evidence. N Z Med J. 1991;104:249-251. Petocz P, Favaloro EJ. Comparison of the ef- sen JI. Effects of partially hydrogenated fish oil, 203. Suh I, Shaten BJ, Cutler JA, Kuller LH, The Mul- fects of two low fat diets with different alpha- partially hydrogenated soybean oil, and butter on tiple Risk Factor Intervention Trial Research linolenic:linoleic acid ratios on coagulation and hemostatic variables in men. Arterioscler Thromb Group. Alcohol use and mortality from coro- fibrinolysis. Atherosclerosis. 1999;142:159- Vasc Biol. 1996;16:375-380. nary heart disease: the role of high-density li- 168. 187. Kristensen SD, Schmidt EB, Dyerberg J. Di- poprotein cholesterol. Ann Intern Med. 1992; 172. Muller AD, van Houwelingen AC, Dam-Mieras etary supplementation with n-3 polyunsatu- 116:881-887.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2390

Downloaded From: https://jamanetwork.com/ on 09/23/2021 204. Gaziano JM, Hennekens CH, Godfried SL, et al. type plasminogen activator and plasminogen 239. Pernerstorfer T, Stohlawetz P, Stummvoll G, et Type of alcoholic beverage and risk of myocar- activator inhibitor. Thromb Haemost. 1990;63: al. Low-dose aspirin does not lower in vivo plate- dial infarction. Am J Cardiol. 1999;83:52-57. 345-348. let activation in healthy smokers. Br J Haema- 205. Ridker PM, Vaughan DE, Stampfer MJ, Glynn RJ, 222. Lacoste L, Hung J, Lam JY. Acute and delayed tol. 1998;102:1229-1231. Hennekens CH. Association of moderate alco- antithrombotic effects of alcohol in humans. Am 240. Blache D, Bouthillier D, Davignon J. Acute influ- hol consumption and plasma concentration of J Cardiol. 2001;87:82-85. ence of smoking on platelet behaviour, endo- endogenous tissue-type plasminogen activa- 223. Rimm EB, Williams P, Fosher K, Criqui M, thelium and plasma lipids and normalization by tor. JAMA. 1994;272:929-933. Stampfer MJ. Moderate alcohol intake and lower aspirin. Atherosclerosis. 1992;93:179-188. 206. Hendriks HF, van der Gaag MS. Alcohol, coagu- risk of coronary heart disease: meta-analysis of 241. Blann AD, Kirkpatrick U, Devine C, Naser S, Mc- lation and fibrinolysis. Novartis Found Symp. effects on lipids and haemostatic factors. BMJ. Collum CN. The influence of acute smoking on 1998;216:111-120. 1999;319:1523-1528. leucocytes, platelets and the endothelium. Ath- 207. Mukamal KJ, Jadhav PP, D’Agostino RB, et al. 224. Renaud SC, Ruf JC. Effects of alcohol on plate- erosclerosis. 1998;141:133-139. Alcohol consumption and hemostatic factors: let functions. Clin Chim Acta. 1996;246:77-89. 242. Gleerup G, Winther K. Smoking further in- analysis of the Framingham Offspring cohort. Cir- 225. Ruf JC. Wine and polyphenols related to plate- creases platelet activity in patients with mild hy- culation. 2001;104:1367-1373. let aggregation and atherothrombosis. Drugs Exp pertension. Eur J Clin Invest. 1996;26:49-52. 208. Lee AJ, Flanagan PA, Rumley A, Fowkes FGR, Clin Res. 1999;25:125-131. 243. Davis JW, Shelton L, Eigenberg DA, Hignite CE, Lowe GDO. Relationship between alcohol in- 226. Puddey IB, Rakic V, Dimmitt SB, Beilin LJ. In- Watanabe IS. Effects of tobacco and non- take and tissue-plasminogen activator antigen fluence of pattern of drinking on cardiovascular tobacco cigarette smoking on endothelium and and other hemostatic factors in the general popu- disease and cardiovascular risk factors—a re- platelets. Clin Pharmacol Ther. 1985;37:529- lation. Fibrinolysis. 1995;9:49-54. view. Addiction. 1999;94:649-663. 533. 209. Ridker PM, Hennekens CH, Stampfer MJ, Man- 227. Marques-Vidal P, Arveiler D, Evans A, Amouyel 244. Simpson AJ, Booth NA, Moore NR, Gray RS. son JE, Vaughan DE. Prospective study of en- P, Ferrieres J, Ducimetiere P. Different alcohol Does chronic smoking influence fibrinolytic po- dogenous tissue plasminogen activator and risk drinking and blood pressure relationships in tential in type 1 diabetes mellitus? Diabet Med. of stroke. Lancet. 1994;343:940-943. France and Northern Ireland: The PRIME Study. 1998;15:683-687. 210. Gorinstein S, Zemser M, Lichman I, et al. Mod- Hypertension. 2001;38:1361-1366. 245. Yarnell JW, Sweetnam PM, Rogers S, et al. Some erate beer consumption and the blood coagula- 228. Eliasson M, Asplund K, Evrin PE, Lundblad D. long term effects of smoking on the haemo- tion in patients with coronary artery disease. J Relationship of cigarette smoking and snuff dip- static system: a report from the Caerphilly and Intern Med. 1997;241:47-51. ping to plasma fibrinogen, fibrinolytic variables Speedwell Collaborative Surveys. J Clin Pathol. 211. Dimmitt SB, Rakic V, Puddey IB, et al. The ef- and serum insulin: The Northern Sweden 1987;40:909-913. fects of alcohol on coagulation and fibrinolytic MONICA Study. Atherosclerosis. 1995;113:41- 246. Feher MD, Rampling MW, Brown J, et al. Acute factors: a controlled trial. Blood Coagul Fibrino- 53. changes in atherogenic and thrombogenic fac- lysis. 1998;9:39-45. 229. Meade TW, Imeson J, Stirling Y. Effects of tors with cessation of smoking. J R Soc Med. 212. Pellegrini N, Pareti FI, Stabile F, Brusamolino A, changes in smoking and other characteristics on 1990;83:146-148. Simonetti P. Effects of moderate consumption clotting factors and the risk of ischaemic heart 247. ECAT angina pectoris study: baseline associa- of red wine on platelet aggregation and haemo- disease. Lancet. 1987;2:986-988. tions of haemostatic factors with extent of coro- static variables in healthy volunteers. EurJClin 230. Simpson AJ, Gray RS, Moore NR, Booth NA. The nary arteriosclerosis and other coronary risk fac- Nutr. 1996;50:209-213. effects of chronic smoking on the fibrinolytic po- tors in 3000 patients with angina pectoris 213. Elmer O, Goransson G, Zoucas E. Impairment of tential of plasma and platelets. Br J Haematol. undergoing coronary angiography. Eur Heart J. primary hemostasis and platelet function after 1997;97:208-213. 1993;14:8-17. alcohol ingestion in man. Haemostasis. 1984; 231. Enderle MD, Pfohl M, Kellermann N, Haering HU, 248. Smith FB, Lee AJ, Rumley A, Fowkes FG, Lowe 14:223-228. Hoffmeister HM. Endothelial function, variables of GD. Tissue-plasminogen activator, plasmino- 214. Pikaar NA, Wedel M, van der Beek EJ, et al. Ef- fibrinolysis and coagulation in smokers and healthy gen activator inhibitor and risk of peripheral ar- fects of moderate alcohol consumption on plate- controls. Haemostasis. 2000;30:149-158. terial disease. Atherosclerosis. 1995;115:35- let aggregation, fibrinolysis, and blood lipids. Me- 232. Newby DE, Wright RA, Labinjoh C, et al. Endo- 43. tabolism. 1987;36:538-543. thelial dysfunction, impaired endogenous fibri- 249. Chia S, Newby DE. Atherosclerosis, cigarette smok- 215. El Sayed MS, Nieuwenhuizen W. The effect of al- nolysis, and cigarette smoking: a mechanism for ing, and endogenous fibrinolysis: is there a direct cohol ingestion on the exercise-induced changes arterial thrombosis and myocardial infarction. Cir- link? Curr Atheroscler Rep. 2002;4:143-148. in fibrin and fibrinogen degradation products in culation. 1999;99:1411-1415. 250. Hung J, Lam JY, Lacoste L, Letchacovski G. Ciga- man. Blood Coagul Fibrinolysis. 2000;11:359- 233. Newby DE, McLeod AL, Uren NG, et al. Im- rette smoking acutely increases platelet throm- 365. paired coronary tissue plasminogen activator re- bus formation in patients with coronary artery 216. van de WA, van Golde PM, Kraaijenhagen RJ, lease is associated with coronary atherosclero- disease taking aspirin. Circulation. 1995;92: dem Borne PA, Bouma BN, Hart HC. Acute in- sis and cigarette smoking: direct link between 2432-2436. hibitory effect of alcohol on fibrinolysis. EurJClin endothelial dysfunction and atherothrombosis. 251. Davis JW, Hartman CR, Shelton L, Ruttinger HA. Invest. 2001;31:164-170. Circulation. 2001;103:1936-1941. A trial of dipyridamole and aspirin in the preven- 217. Johansen KM, Skorpe S, Olsen JO, Osterud B. 234. Allen RA, Kluft C, Brommer EJ. Effect of chronic tion of smoking-induced changes in platelets and The effect of red wine on the fibrinolytic system smoking on fibrinolysis. Arteriosclerosis. 1985; endothelium in men with coronary artery dis- and the cellular activation reactions before and 5:443-450. ease. Am J Cardiol. 1989;63:1450-1454. after exercise. Thromb Res. 1999;96:355-363. 235. Belch JJ, McArdle BM, Burns P, Lowe GD, Forbes 252. Blann AD, Steele C, McCollum CN. The influ- 218. Hendriks HF, Veenstra J, Velthuis-te Wierik EJ, CD. The effects of acute smoking on platelet be- ence of smoking on soluble adhesion mol- Schaafsma G, Kluft C. Effect of moderate dose haviour, fibrinolysis and haemorheology in ha- ecules and endothelial cell markers. Thromb Res. of alcohol with evening meal on fibrinolytic fac- bitual smokers. Thromb Haemost. 1984;51: 1997;85:433-438. tors. BMJ. 1994;308:1003-1006. 6-8. 253. Bain BJ, Rothwell M, Feher MD, Robinson R, 219. Numminen H, Syrjala M, Benthin G, Kaste M, Hill- 236. Haire WD, Goldsmith JC, Rasmussen J. Abnor- Brown J, Sever PS. Acute changes in haemato- bom M. The effect of acute ingestion of a large mal fibrinolysis in healthy male cigarette smok- logical parameters on cessation of smoking. J dose of alcohol on the hemostatic system and ers: role of plasminogen activator inhibitors. Am R Soc Med. 1992;85:80-82. its circadian variation. Stroke. 2000;31:1269- J Hematol. 1989;31:36-40. 254. Davis JW, Shelton L, Watanabe IS, Arnold J. Pas- 1273. 237. Brockmann MA, Beythien C, Magens MM, Wilck- sive smoking affects endothelium and plate- 220. McConnell MV, Vavouranakis I, Wu LL, Vaughan ens V, Kuehnl P, Gutensohn K. Platelet hemo- lets. Arch Intern Med. 1989;149:386-389. DE, Ridker PM. Effects of a single, daily alco- stasis capacity in smokers: in vitro function analy- 255. Williams JE, Nieto FJ, Sanford CP, Tyroler HA. holic beverage on lipid and hemostatic markers ses with 3.2% citrated whole blood. Thromb Res. Effects of an angry temperament on coronary of cardiovascular risk. Am J Cardiol. 1997;80: 2001;104:333-342. heart disease risk: The Atherosclerosis Risk in 1226-1228. 238. Modesti PA, Abbate R, Gensini GF, Colella A, Neri Communities Study. Am J Epidemiol. 2001;154: 221. Veenstra J, Kluft C, Ockhuizen TH, vd Pol H, We- Serneri GG. Platelet thromboxane A2 receptors 230-235. del M, Schaafsma G. Effects of moderate alco- in habitual smokers. Thromb Res. 1989;55:195- 256. Whiteman MC, Deary IJ, Fowkes FG. Personal- hol consumption on platelet function, tissue- 201. ity and social predictors of atherosclerotic pro-

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2391

Downloaded From: https://jamanetwork.com/ on 09/23/2021 gression: Edinburgh Artery Study. Psychosom 269. Larsson PT, Wallen NH, Hjemdahl P. Norepi- 282. Janssen K, Mensink RP, Cox FJ, et al. Effects of Med. 2000;62:703-714. nephrine-induced human platelet activation in the flavonoids quercetin and apigenin on hemo- 257. Everson SA, Lynch JW, Chesney MA, et al. In- vivo is only partly counteracted by aspirin. Cir- stasis in healthy volunteers: results from an in teraction of workplace demands and cardiovas- culation. 1994;89:1951-1957. vitro and a dietary supplement study. AmJClin cular reactivity in progression of carotid athero- 270. Bak AA, van Vliet HH, Grobbee DE. Coffee, caf- Nutr. 1998;67:255-262. sclerosis: population based study. BMJ. 1997; feine and hemostasis: results from two random- 283. Vorster H, Jerling J, Oosthuizen W, et al. Tea 314:553-558. ized studies. Atherosclerosis. 1990;83:249- drinking and haemostasis: a randomized, pla- 258. Everson SA, Lynch JW, Kaplan GA, Lakka TA, 255. cebo-controlled, crossover study in free-living Sivenius J, Salonen JT. Stress-induced blood 271. Naismith DJ, Akinyanju PA, Szanto S, Yudkin J. subjects. Haemostasis. 1996;26:58-64. pressure reactivity and incident stroke in middle- The effect, in volunteers, of coffee and decaf- 284. de Maat MP, Pijl H, Kluft C, Princen HM. Con- aged men. Stroke. 2001;32:1263-1270. feinated coffee on blood glucose, insulin, plasma sumption of black and green tea had no effect 259. Patterson SM, Krantz DS, Gottdiener JS, Hecht lipids and some factors involved in blood clot- on inflammation, haemostasis and endothelial G, Vargot S, Goldstein DS. Prothrombotic ef- ting. Nutr Metab. 1970;12:144-151. markers in smoking healthy individuals. Eur J Clin fects of environmental stress: changes in plate- 272. Happonen P, Salonen JT, Seppanen K, Raura- Nutr. 2000;54:757-763. let function, hematocrit, and total plasma pro- maa R. Association of coffee consumption with 285. Hodgson JM, Puddey IB, Mori TA, Burke V, Baker tein. Psychosom Med. 1995;57:592-599. plasma lipoprotein, fibrinogen and platelet ag- RI, Beilin LJ. Effects of regular ingestion of black 260. Jern C, Eriksson E, Tengborn L, Risberg B, Wad- gregability in middle-aged men. In: Proceed- tea on haemostasis and cell adhesion mol- envik H, Jern S. Changes of plasma coagulation ings of the Meeting of the International Epide- ecules in humans. Eur J Clin Nutr. 2001;55:881- and fibrinolysis in response to mental stress. miological Association; August 8-13, 1987; 886. Thromb Haemost. 1989;62:767-771. Helsinki, Finland. 286. Duffy SJ, Vita JA, Holbrook M, Swerdloff PL, 261. Urano T, Cho M, Takahashi S, et al. Changes of 273. Samarrae WA, Truswell AS. Short-term effect of Keaney JF Jr. Effect of acute and chronic tea con- parameters in fibrinolytic system caused by men- coffee on blood fibrinolytic activity in healthy sumption on platelet aggregation in patients with tal stress. Thromb Res. 1990;60:501-507. adults. Atherosclerosis. 1977;26:255-260. coronary artery disease. Arterioscler Thromb 262. Raikkonen K, Lassila R, Keltikangas-Jarvinen L, 274. Wojta J, Kirchheimer JC, Peska MG, Binder BR. Vasc Biol. 2001;21:1084-1089. Hautanen A. Association of chronic stress with Effect of caffein ingestion on plasma fibrinolytic 287. plasminogen activator inhibitor-1 in healthy potential. Thromb Haemost. 1988;59:337-338. Jolliffe JA, Rees K, Taylor RS, Thompson D, Ol- middle-aged men. Arterioscler Thromb Vasc Biol. 275. Ammaturo V, Perricone C, Canazio A, et al. Caf- dridge N, Ebrahim S. Exercise-based rehabilita- 1996;16:363-367. feine stimulates in vivo platelet reactivity. Acta tion for coronary heart disease. Cochrane Data- 263. Vrijkotte TG, van Doornen LJ, De Geus EJ. Work Med Scand. 1988;224:245-247. base Syst Rev. 2001;1:CD001800. stress and metabolic and hemostatic risk fac- 276. Paoletti R, Corsini A, Tremoli E, Fumagalli R, Cata- 288. Lavie CJ, Milani RV. Benefits of cardiac rehabili- tors. Psychosom Med. 1999;61:796-805. pano AL. Effects of coffee on plasma lipids, li- tation and exercise training. Chest. 2000;117: 264. Wamala SP, Murray MA, Horsten M, et al. So- poproteins and apolipoproteins. Pharmacol Res. 5-7. cioeconomic status and determinants of hemo- 1989;21:27-38. 289. Fernhall B, Szymanski LM, Gorman PA, Milani static function in healthy women. Arterioscler 277. Bydlowski SP, Yunker RL, Rymaszewski Z, Sub- J, Paup DC, Kessler CM. Fibrinolytic activity is Thromb Vasc Biol. 1999;19:485-492. biah MT. Coffee extracts inhibit platelet aggre- not dependent upon exercise mode in post- 265. Frimerman A, Miller HI, Laniado S, Keren G. gation in vivo and in vitro. Int J Vitam Nutr Res. myocardial infarction patients. Eur J Appl Physiol Changes in hemostatic function at times of cy- 1987;57:217-223. Occup Physiol. 1998;78:247-252. clic variation in occupational stress. Am J Car- 278. Peters U, Poole C, Arab L. Does tea affect car- 290. Estelles A, Aznar J, Tormo G, Sapena P, Tormo diol. 1997;79:72-75. diovascular disease? a meta-analysis. Am J Epi- V, Espana F. Influence of a rehabilitation sports 266. Ishizaki M, Tsuritani I, Noborisaka Y, Yamada Y, demiol. 2001;154:495-503. programme on the fibrinolytic activity of pa- Tabata M, Nakagawa H. Relationship between job 279. Yoshida H, Ishikawa T, Hosoai H, et al. Inhibi- tients after myocardial infarction. Thromb Res. stress and plasma fibrinolytic activity in male tory effect of tea flavonoids on the ability of cells 1989;55:203-212. Japanese workers. Int Arch Occup Environ Health. to oxidize low density lipoprotein. Biochem Phar- 291. Paramo JA, Olavide I, Barba J, et al. Long-term 1996;68:315-320. macol. 1999;58:1695-1703. cardiac rehabilitation program favorably influ- 267. Grignani G, Pacchiarini L, Zucchella M, et al. Effect 280. Neiva TJ, Morais L, Polack M, Simoes CM, ences fibrinolysis and lipid concentrations in of mental stress on platelet function in normal D’Amico EA. Effects of catechins on human blood acute myocardial infarction. Haematologica. subjects and in patients with coronary artery dis- platelet aggregation and lipid peroxidation. Phy- 1998;83:519-524. ease. Haemostasis. 1992;22:138-146. tother Res. 1999;13:597-600. 292. Hamouratidis ND, Pertsinidis TE, Bacharoudis 268. Li N, Wallen NH, Hjemdahl P. Evidence for pro- 281. Kang WS, Lim IH, Yuk DY, et al. Antithrombotic GP, Papazachariou GS. Effects of exercise on thrombotic effects of exercise and limited pro- activities of green tea catechins and (-)- plasma fibrinolytic activity in patients with isch- tection by aspirin. Circulation. 1999;100:1374- epigallocatechin gallate. Thromb Res. 1999;96: aemic heart disease. Int J Cardiol. 1988;19: 1379. 229-237. 39-45.

(REPRINTED) ARCH INTERN MED/ VOL 163, OCT 27, 2003 WWW.ARCHINTERNMED.COM 2392

Downloaded From: https://jamanetwork.com/ on 09/23/2021