Efficacy of antithrombin in the prevention of microvascular thrombosis during endotoxemia: An intravital microscopic study

Thrombosis Research (2007) 121, 241–248

intl.elsevierhealth.com/journals/thre

REGULAR ARTICLE

Efficacy of antithrombin in the prevention of microvascular thrombosis during endotoxemia: An intravital microscopic study Heiko Sorg a,⁎, 1 , Johannes N. Hoffmann b,1 , Rolando E. Rumbaut c , Michael D. Menger d , Nicole Lindenblatt a , Brigitte Vollmar a a

Institute for Experimental Surgery, University of Rostock, Schillingallee 69a, 18055 Rostock, Germany Department of Surgery, Klinikum Grosshadern, Ludwig-Maximilians-University of Munich, 81377 Munich, Germany c Department of Medicine and Pediatrics, Baylor College of Medicine, Houston 77030 TX, USA d Institute for Clinical and Experimental Surgery, University of Saarland, 66424 Homburg-Saar, Germany b

Received 5 January 2007; received in revised form 10 April 2007; accepted 11 April 2007 Available online 18 May 2007

KEYWORDS Cremaster muscle; Light/dye injury; Microcirculation; Mouse ear model; Sepsis; Disseminated intravascular coagulation

Abstract Introduction: The KyberSept trial in septic patients showed that antithrombin (AT) reduced 90-day mortality significantly in a subgroup of patients not receiving concomitant heparin for thrombosis prophylaxis. Microvascular thrombosis is a key pathophysiologic mechanism during sepsis, ischemia/reperfusion and disseminated intravascular coagulation (DIC). Therefore, this study investigated the antithrombotic property of AT as potential monotherapy in an experimental endotoxemia model in order to omit concomitant heparin. Materials and methods: Using a light/dye injury model in the ear and the cremaster muscle preparation of mice, we quantitatively assessed microvascular thrombus formation in a total of 30 endotoxemic mice by means of intravital fluorescence microscopy. Before thrombus induction animals received a single i.v. bolus of AT (100 or 250 IU/kg), heparin (100 IU/kg) or saline (NaCl). Results: In NaCl-treated endotoxemic animals, light/dye exposure led to complete thrombotic occlusion in arterioles and venules within b 450 s in the ear model. Heparin delayed thrombotic vessel occlusion by more than 50%. AT significantly prolonged times until thrombotic vessel occlusion in a dose-dependent manner and more effectively than heparin (p b 0.05 vs. NaCl and heparin). This anti-coagulative effect of AT was especially pronounced in arterioles. Upon light/dye exposure to cremaster muscle preparations in endotoxemic mice AT also caused a 4-fold delay in

⁎ Corresponding author. Tel.: +49 381 494 6238; fax: +49 381 494 6222. E-mail address: [email protected] (H. Sorg). 1 Both authors equally contributed to the study. 0049-3848/$ - see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2007.04.001

242

H. Sorg et al. microvascular thrombus growth with 827 ± 77 s until complete thrombotic occlusion. Conclusions: We could characterize for the first time AT-mediated antithrombotic activity during endotoxemia in two models of phototoxicity-induced microvascular thrombosis. Our results clearly demonstrate an additional AT mechanism of action that may be responsible for beneficial effects observed during endotoxemia and DIC. This AT profile may allow future high-dose AT application without giving heparin for thrombosis prophylaxis, an intriguing strategy that is to be tested under clinical conditions. © 2007 Elsevier Ltd. All rights reserved.

Introduction Sepsis, defined as a systemic response to an infection, is a complex disease and a major health problem considering its significant morbidity and mortality of up to 40% [1]. Considerable efforts have been undertaken to understand the pathogenesis of sepsis and to improve its therapy. Major therapeutic columns, i.e. (i) identification and control of the infection source, (ii) antibiotic therapy, (iii) organ supportive strategies and (iv) adjunctive treatment regimens, are addressing the optimum treatment of patients with sepsis, severe sepsis and septic shock along the surviving sepsis campaign guidelines [2]. Activation of coagulation and the generation of inflammatory mediators by the coagulation system, resulting in disturbance of the microcirculation, have been recognized to be an important part of the septic pro-inflammatory response [3]. Therefore, treatment of severe sepsis with natural coagulation inhibitors became a focus of considerable interest. Activated protein C and antithrombin (AT) are now known to not only exert an anticoagulant activity but also modulate inflammatory reactions [3]. AT inhibits inflammation by binding to glycosaminoglycans (GAGs) on endothelial surfaces, involving the subsequent release of prostacyclin [4–6] and the suppression of NFkappaB activation [7]. AT further suppresses the adhesion of leukocytes to the vascular endothelium [5] and thus can ameliorate organ failure [8]. In contrast to these promising results from animal studies, a large-scale clinical trial in patients with severe sepsis failed to prove efficacy of AT in terms of survival [9]. Interestingly, a pre-specified subgroup analysis of AT-treated patients not receiving concomitant heparin, however, revealed a significant benefit in terms of 90-day survival when compared with the control group [10]. Thus, an interference of AT with heparin administration has been suggested and confirmed by recent experimental and clinical findings [11–13]. Heparin is commonly used for venous thrombosis prophylaxis and is actually recommended by the surviving sepsis campaign guidelines [2]. Its actual

role in the treatment of intensive care unit (ICU) patients is not completely clear. There is thus far no large scale controlled trial about the use of heparin in critically ill patients. Moreover, heparin has been related to relevant adverse effects. Besides different types of heparin-induced thrombocytopenia, the combination of heparin with AT is known to increase bleeding risk [10] and has failed to reduce duration of DIC in critically ill patients in contrast to AT alone [14]. In line with this, AT alone highly effectively reduced 28-day and 90-day mortality in patients with DIC according to the International Society for thrombosis and hemostasis (ISTH) criteria [15]. Since microvascular thrombosis is known to be one key mechanism in DIC formation, sepsis and ischemia/reperfusion, it was the aim of the present study to investigate the efficacy of AT monotherapy in terms of microvascular thrombosis formation under septic conditions.

Materials and methods Animals A total of 20 homozygous (SKH-1-hr) hairless mice of either sex (10–12 weeks old) with a body weight (bw) of 22–40 g and 10 male C57Bl/6J mice (10–12 weeks old, 25–40 g bw) were used for the study. The animals were housed in standard laboratories with a 12-h lightdark cycle and had access to standard laboratory chow and water ad libitum. The experiments in each of the two participating institutions were conducted in accordance with guidelines for the Care and Use of Laboratory Animals and the Institutional Animal Care and Use Committee (University of Rostock, Medical Faculty, Rostock, Germany and Baylor College of Medicine, Houston, TX, USA).

Thrombus induction and microscopy in the ear of the hairless mouse For thrombus induction the light/dye injury model was used, as described by Roesken et al. [16] in the

AT and microvascular thrombosis ear of SKH-1-hr hairless mice. Mice were anesthetized with ketamine/xylazine (90/25 mg/kg bw i.p.) and placed prone on a Plexiglas pad with an integrated heating plate for maintenance of body temperature at ∼37.5 °C. Implantation of a fine polyethylene catheter (PE-10, 0.28 mm internal diameter) into the right jugular vein served for application of drugs and fluorescent dyes. The ear to be investigated was gently extended over a microscopic slide embedded into the pad and placed under an intravital fluorescence microscope equipped with a 100-W mercury lamp (Axiotech vario, Zeiss, Jena, Germany). After i.v. injection of 0.15 ml 5% fluorescein isothiocyanate (FITC)-labeled dextran (molecular weight 150 kDa; Sigma Chemical, Deisenhofen, Germany), photochemical thrombus formation was induced by continuous blue light (450–490 nm) exposure to the individual microvessels, using a ×63 water immersion objective. In each ear, up to three arterioles and up to five venules were studied. Light exposure was terminated after blood flow in the vessel ceased for at least 30 s due to complete vessel occlusion. Microscopic images were recorded by a video system (S-VHS Panasonic AG 7350-E, Matsushita, Tokyo, Japan) for off-line evaluation through a charge-coupled device video camera (FK 6990A-IQ, Pieper, Berlin, Germany) and were monitored on a television screen.

Experimental design and experimental groups All animals (n = 20) received an i.v. application of 3 mg/kg Escherichia coli lipopolysaccharide (LPS) (serotype 0128:B12; Sigma) at 1 h before light/dye

243 exposure was started (Fig. 1A). Five minutes before thrombus induction (Fig. 1A) animals received a single i.v. bolus of heparin (100 IU/kg bw; Liquemin N Hoffmann-La Roche AG Grenzach-Wyhlen, Germany; n = 5) or AT in dosages of 100 or 250 IU/kg bw (Kybernin HS, ZLB Behring, Marburg, Germany; each n = 5). Control animals received equivalent volumes of physiological saline (NaCl; 10 ml/kg bw; n = 5). After the experiment, animals were allowed to recover from anesthesia with access to food and water ad libitum. After 24 h, all vessels under investigation were reanalyzed for microvascular patency rate (Fig. 1A) as well as for blood and tissue sampling.

Thrombus induction and microscopy in the mouse cremaster muscle Male C57Bl/6J mice under pentobarbital anesthesia (30 mg/kg bw; n = 10) were used for the cremaster muscle preparation as originally described by Baez [17] in rats and applied by our group in mice [18,19]. For this purpose, mice were placed on a custom plexiglas tray and maintained at 37 °C with a homeothermic blanket monitored with a rectal temperature probe (FHC, ME, USA). A tracheotomy was performed to facilitate breathing, and the right external jugular vein was cannulated for administration of drugs and fluorescent dyes. After preparation and exposure of the cremaster muscle, tissue was superfused continuously with 37 °C warm bicarbonate-buffered saline (in mM: NaCl 127, KCl 4.7, CaCl2·2H2O 2.0, MgSO4 1.2, NaHCO3 28.0 and glucose 5.0). The buffer was bubbled continuously

Figure 1 Flow chart displaying the experimental protocols for hairless (A) and C57Bl/6J mice (B). One hour before thrombus induction (TI) anesthetized hairless mice received E. coli lipopolysaccharide (LPS; 3 mg/kg bw iv). After application of the respective medication (saline (NaCl), heparin (100 IU/kg bw) or antithrombin (AT) 100 or 250 IU/kg bw) 5 min prior to thrombus induction, intravital fluorescence microscopy (IVM) was performed (A). In anesthetized animals with the cremaster muscle preparation for IVM analysis of thrombosis (B), E. coli lipopolysaccharide (5 mg/kg) was administered 120 min prior to final experiment.

244 with 95% N2–5% CO2 gas mixture to maintain a pH between 7.35 and 7.45. The preparation was then transferred to the microscopic stage and allowed to equilibrate for 30 min during which the preparation was surveyed to select second- and third-order venules. After i.v. injection of 0.15 ml 5% FITClabeled dextran, photochemical thrombus formation was started as mentioned above.

Experimental design and experimental groups For induction of endotoxemia in animals (n = 10), which underwent the cremaster muscle preparation, E. coli LPS (5 mg/kg; serotype O111:B2; SigmaChemical, MO, USA) was administered i.p. 2 h before light/dye exposure (Fig. 1B). Five minutes before thrombus induction (Fig. 1B), animals received a single i.v. bolus of AT at a dosage of 250 IU/kg bw (Kybernin HS, ZLB Behring, Marburg, Germany; n = 5). Control animals received equivalent volumes of physiological saline (10 ml/kg bw; n = 5).

H. Sorg et al. nostica Stago, Asnières, France) and a STA-R coagulation analyzer (Diagnostica Stago).

Statistical analysis All data are given as mean ± SEM. Data were analyzed for normality and equal variance across groups. Differences between groups were assessed using one-way ANOVA followed by the appropriate post hoc comparison test. Overall statistical significance was set at p b 0.05. Statistics were performed using the software package SigmaStat (Jandel Corporation, SanRafael, CA, USA).

Results All endotoxemic hairless mice survived the experimental time period of 24 h but revealed signs

Microcirculatory analysis Kinetics of intravascular thrombus formation were quantified off-line by analysis of the videotaped images using the computer assisted image analysis system CapImage (Dr. Zeintl Software, Heidelberg, Germany). Thrombus formation was quantified by assessing the time until sustained cessation of blood flow due to complete vessel occlusion, as well as the percentage of vessels that were completely clogged after 25 min, determined as the acute occlusion rate of vessels. If a vessel did not clog within 25 min of continuous light exposure, observation was stopped and the vessel was considered as patent. Upon reanalysis 24 h after light/dye induced thrombus formation in the mouse ear model, the percentage of patent vessels was considered as patency rate of arterioles and venules, respectively.

Hematological analysis and bleeding time By retrobulbar puncture using a glass capillary, blood was sampled for subsequent laboratory analysis at 24 h after thrombus induction in hairless mice. Blood cell count and hematocrit were determined using an automatic cell counter (Sysmex KX21, Sysmex GmbH, Norderstedt, Germany). After severing a segment of the tail, the amputated tail was immersed in 37 °C warm water and the time required for the bloodstream to stop was defined as the bleeding time. In addition, blood samples served for measurement of AT activities in citrated plasma using the amidolytic anti-factor IIa method (STA Antithrombin III; Diag-

Figure 2 Acute occlusion rate (A) and chronic patency rate (B) of arterioles and venules after induction of light/dye injury for thrombus formation in endotoxemic SKH-1-hr mice treated with either saline (NaCl), heparin (100 IU/kg) or AT (100 and 250 IU/kg). Data are given as means ± SEM; ANOVA, post hoc comparison; ⁎p b 0.05 vs. NaCl; #p b 0.01 vs. heparin; § p b 0.01 vs. AT-100.

AT and microvascular thrombosis

245

of sickness, i.e. closed eyes, scrubby coat and reduced explorative activity. Bleeding disorders could not be observed.

Effects of AT and heparin on microvascular thrombotic occlusion rates Application of both physiologic saline and heparin led to a thrombotic occlusion rate of 100% (saline: arterioles n = 10, venules n = 36; heparin: arterioles n = 10, venules n = 32). AT-100 could prevent thrombus formation in 20% of the arterioles (n = 14) and 2.5% of the venules (n = 37). The AT dose of 250 IU/kg was far more effective with respect to the fact that in only 20% of all arterioles (n = 11) and 46% of all venules (n = 33) under investigation thrombotic vessel occlusion could be induced (Fig. 2A; p b 0.05 vs. NaCl; p b 0.01 vs. heparin). At 24 h after thrombus induction, almost 80% and 45% of all vessels in the AT-250 (p b 0.05 vs. NaCl; p b 0.01 vs. heparin and AT-100) and the AT-100 group were still open (Fig. 2B), while in heparin- and saline-treated animals patency rate of vessels was well below 40% (Fig. 2B).

Effects of AT and heparin on microvascular thrombus kinetics In saline-treated animals, arteriolar and venular vessel occlusions were observed after approximately 440 s and 290 s of light/dye exposure (Fig. 3). Heparin prolonged the time until complete vessel occlusion by 1.5-fold in both arterioles and venules (Fig. 3; p b 0.05 vs. NaCl).

Figure 4 Tail bleeding times 24 h after induction of light/ dye injury for thrombus formation in endotoxemic SKH-1-hr mice treated with either saline (NaCl), heparin (100 IU/kg) or AT (100 and 250 IU/kg). Data are given as means ± SEM; ANOVA, post hoc comparison; ⁎p b 0.01 vs. NaCl; #p b 0.01 vs. heparin; §p b 0.01 vs. AT-100.

AT dose-dependently delayed microvascular thrombus growth (Fig. 3). AT-100 increased the time until complete vessel occlusion comparably to heparin in arterioles, but significantly more in venules. Notably, in the AT-250 group thrombus growth was markedly delayed compared to all other groups (p b 0.05). Thrombotic vessel occlusion was observed after 1336 ± 106 s in arterioles and 1078 ± 144 s in venules (Fig. 3). Comparably, AT-250 caused a 4-fold increase of the time necessary for complete occlusion of venules (n = 10) in the cremaster muscle of endotoxemic mice (827 ± 77 s vs. NaCl: 220 ± 19 s (n = 10); p b 0.05).

Effects of ATand heparin on tail bleeding times Tail bleeding time in endotoxemic hairless mice treated with 250 IU/kg AT was significantly longer when compared to all other groups (Fig. 4). In

Table 1 Peripheral blood cell counts as well as AT activities 24 h after induction of light/dye injury for thrombus formation in mice treated with either saline, heparin or increasing doses of AT in hairless mice Groups RBC WBC PLT HCT AT (× 1012/l) (× 109/l) (×109/l) (%) (%)

Figure 3 Time until complete occlusion of arterioles and venules after induction of light/dye injury for thrombus formation in endotoxemic SKH-1-hr mice treated with either saline (NaCl), heparin (100 IU/kg) or AT (100 and 250 IU/kg). Data are given as means ± SEM; ANOVA, post hoc comparison; ⁎p b 0.05 vs. NaCl; #p b 0.05 vs. heparin.

NaCl Heparin AT-100 AT-250

8.0 ± 0.5 8.4 ± 0.1 8.4 ± 0.2 8.1 ± 0.4

2.3 ± 0.5 3.1 ± 0.9 2.3 ± 0.2 2.0 ± 0.4

527 ± 62 640 ± 49 509 ± 34 518 ± 24

40 ± 3 66 ± 5 40 ± 1 68 ± 6 50 ± 1 98 ± 2⁎,# 40 ± 3 121 ± 4⁎,#,§

RBC, red blood cells; WBC, white blood cells; PLT, platelets; HCT, hematocrit; AT, antithrombin plasma activity. Data are given as means ± SEM; ANOVA, post hoc comparison; ⁎p b 0.001 vs. NaCl; #p b 0.05 vs. heparin; §p b 0.05 vs. AT-100.

246 parallel, endotoxemic mice used for the cremaster muscle preparation showed a significant prolongation of bleeding time from 70 ± 10 s (NaCl) to 170 ± 25 s in AT-250-treated animals (p b 0.05).

Effect of AT and heparin on blood cell counts as well as plasma AT activity Neither heparin nor AT significantly affected systemic blood cell counts and hematocrit when compared to those of saline-treated controls (Table 1). However, animals of the AT-treated groups showed a dose-dependent increase of ATactivities at 24 h after application of 100 and 250 IU/kg AT, while ATactivities were found b70% in saline- and heparintreated animals (Table 1).

Discussion Microvascular thrombosis may be a major pathophysiological mechanism contributing to the development and progression of DIC in endotoxemia, thus initiating and promoting septic organ dysfunction. The present data clearly show that AT in two different models of microvascular thrombosis formation could effectively prevent thrombotic microvascular occlusion when given without heparin under the conditions of experimental endotoxemia.

Methodological considerations To test the effects of ATon thrombus formation under endotoxemic conditions, two well-established models were used to study thrombus induction in the microcirculation. The hairless mouse ear model is well characterized by analysis of leukocyte-endothelial cell interaction, capillary perfusion and thrombus formation in a variety of experimental settings [16,20,21]. It allows direct visualization of the process of arteriolar and venular thrombus formation by intravital fluorescence microscopy. Of interest, there is no need of surgical preparation, which is known to induce per se a local pro-inflammatory response, potentially interfering with the process of thrombus formation [22]. In addition, we used the mouse cremaster muscle preparation to confirm our findings and to exclude possible model-specific or model-induced aspects of microvascular thrombosis depending on the animal model and interfering with the properties of the used anticoagulants. The cremaster muscle allows in vivo microscopy via both trans- and epi-illumination technique, however, requires microsurgical exposure of tissue and thus does not permit repetitive analysis over a prolonged period of time. Yielding comparable results in two

H. Sorg et al. different experimental approaches underlines the reproducibility and strength of the present findings. Thrombus formation was initiated by a photochemical reaction based on the transmural blue light exposure and i.v. administration of high molecular weight FITC-labeled dextran. The phototoxicity of the fluorescently labeled agent is known to cause endothelial injury due to the release of reactive oxygen species. This local endothelial cell damage supports deposition of blood cellular components and represents the ideal site for thrombus formation. In line with this, light/dye-exposure to induce thrombi represents a common and widely used model, in particular to study molecular and cellular mechanisms underlying thrombosis and impaired hemostasis [23–26]. ATwas administered as a single bolus in a maximum dosage (250 IU/kg) that has previously been shown to inhibit microvascular thrombosis in non-endotoxemic animals [21] and to improve outcome in several experimental short-term [27] and long-term sepsis models [5]. These high AT-concentrations were attained because various animal studies showed that supranormal ATactivities are necessary to be efficient during endotoxemia [28]. In addition, doses were chosen that would result in plasma concentrations comparable to those achieved in clinical practice [2]. Most surprisingly, heparin was not associated with changes in thrombosis and bleeding times, contrasting previous results from our group [21] with a prolongation of these two parameters by heparin (100 IU/kg). In line with the present findings, other groups [29,30] described that heparin at dosages of 100 to even 250 IU/kg had no significant effect in irradiation-induced mesenteric platelet thrombus formation. It remains unclear so far, as to whether the lack of anticoagulatory action of heparin is caused by species-specific proneness to heparin action or whether endotoxemic conditions would have compensated anticoagulatory activity of heparin.

Microvascular thrombus formation and AT In many preclinical sepsis models AT has been shown to reduce mortality [27,31,32], most probably by a significant attenuation of microcirculatory dysfunctions and leukocytic inflammation [5,8]. Recent investigations of our group and others [6,33] substantiated that the beneficial AT action critically depends on its binding to endothelial GAGs with subsequent release of anti-inflammatory prostacyclin [34,35]. Direct interaction of AT with granulocytes and endothelial cells is antagonized by heparin, which blocks AT-binding regions and thus prevents the antiinflammatory and anti-adhesive action of AT on the microvascular endothelium [11,12]. In support of an existing heparin-AT antagonism, no effect of high-

AT and microvascular thrombosis dose ATon mortality was seen in the cohort of patients with concomitant heparin for deep venous thrombosis prophylaxis in the KyberSept trial [10]. Of particular interest within the benefit/risk profile of high-dose AT in patients with severe sepsis is the fact that rates of thromboembolic events were similar when AT was given with or without concomitant heparin, but an increased incidence of bleeding was reported with AT plus concomitant heparin [10]. The KyberSept trial was clearly not designed to investigate venous thrombosis incidence and does not fulfill criteria of a respective trial. However, the finding that only one patient in the high-dose AT group without heparin versus four patients in the placebo group without heparin developed thrombosis as a serious adverse event over a 14-day period [10] may indicate a potential protective AT effect in terms of thromboembolic events. Also, a recent clinical study performed in patients using high-dose AT in simultaneous kidney–pancreas transplantation could substantiate a relevant protection from pancreatic thrombosis, an event often leading to graft loss and subsequent morbidity [36]. Therefore, a detailed analysis of AT actions under septic conditions seemed to be appropriate. We recently analyzed the anticoagulant potential of AT monotherapy in a thrombosis model without endotoxemia and found it highly protective against microvascular thrombosis [21]. The present study now extends this information in that high-dose AT effectively prevented the development of microvascular thrombi also during systemic endotoxemia. By affecting the microcirculation AT may be important in preserving microvascular perfusion. Under clinical conditions, there was a clear correlation between low AT levels and poor outcome of patients with severe sepsis [37,38] and patients after cardiac surgery [39]. In these patients low AT levels at ICU admission were significantly associated with prolonged mechanical ventilation and increased duration of ICU treatment. Interestingly, low AT activities also predicted higher incidence of allogenic blood product use and bleeding events as well as an increase in thromboembolic events [39]. Though not specifically studied, it might be assumed that not only the anti-inflammatory action but also the anticoagulant property of AT contribute to the beneficial effects mentioned above. In conclusion, the here shown AT profile of anticoagulative action under endotoxemic conditions may allow future AT application without applying concomitant heparin and thereby, full maintenance of its highly effective anti-inflammatory potential. The avoidance of heparin during high-dose AT therapy in septic patients could represent a promising strategy which however needs to be tested under clinical conditions.

247

Acknowledgments The authors kindly thank Dorothea Frenz, Maren Nerowski, Berit Blendow and Doris Butzlaff (Institute for Experimental Surgery, University of Rostock) as well as Kimberly Langlois (Department of Medicine & Pediatrics, Baylor College of Medicine, Houston) for their excellent technical assistance. The study was in part supported by a grant of the Medical faculty of the University of Rostock (FORUN; H.S.) and an NIH grant HL-079368 from the National Heart Lung and Blood Institute (R.E.R.).

References [1] Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001;29:1303–10. [2] Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, et al. Surviving Sepsis Campaign Management Guidelines Committee. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 2004;32:858–73. [3] Amaral A, Opal SM, Vincent JL. Coagulation in sepsis. Intensive Care Med 2004;30:1032–40. [4] Isobe H, Okajima K, Uchiba M, Harada N, Okabe H. Antithrombin prevents endotoxin-induced hypotension by inhibiting the induction of nitric oxide synthase in rats. Blood 2002;99:1638–45. [5] Hoffmann JN, Vollmar B, Inthorn D, Schildberg FW, Menger MD. Antithrombin reduces leukocyte adhesion during chronic endotoxemia by modulation of the cyclooxygenase pathway. Am J Physiol Cell Physiol 2000;279: C98–C107. [6] Hoffmann JN, Vollmar B, Romisch J, Inthorn D, Schildberg FW, Menger MD. Antithrombin effects on endotoxin-induced microcirculatory disorders are mediated mainly by its interaction with microvascular endothelium. Crit Care Med 2002;30:218–25. [7] Oelschlager C, Romisch J, Staubitz A, Stauss H, Leithauser B, Tillmanns H, et al. Antithrombin III inhibits nuclear factor kappaB activation in human monocytes and vascular endothelial cells. Blood 2002;99:4015–20. [8] Iba T, Kidokoro A, Fukunaga M, Fuse S, Suda M. Antithrombin modulates the leukocyte-endothelial cell interaction in the staphylococcal enterotoxin B-challenged mouse. J Trauma 2003;55:546–50. [9] Warren BL, Eid A, Singer P, Pillay SS, Carl P, Novak I, et al. KyberSept Trial Study Group. Caring for the critically ill patient. High-dose antithrombin III in severe sepsis: a randomized controlled trial. JAMA 2001;286:1869–78. [10] Hoffmann JN, Wiedermann CJ, Juers M, Ostermann H, Kienast J, Briegel J, et al. KyberSept investigators. Benefit/ risk profile of high-dose antithrombin in patients with severe sepsis treated with and without concomitant heparin. Thromb Haemost 2006;95:850–6. [11] Hoffmann JN, Vollmar B, Laschke MW, Inthorn D, Kaneider NC, Dunzendorfer S, et al. Adverse effect of heparin on antithrombin action during endotoxemia: microhemodynamic and cellular mechanisms. Thromb Haemost 2002;88: 242–52. [12] Pulletz S, Lehmann C, Volk T, Schmutzler M, Ziemer S, Kox WJ, et al. Influence of heparin and hirudin on endothelial binding of antithrombin in experimental thrombinemia. Crit Care Med 2000;28:2881–6.

248 [13] Opal SM, Kessler CM, Roemisch J, Knaub S. Antithrombin, heparin, and heparan sulfate. Crit Care Med 2002;30: S325–31. [14] Blauhut B, Kramar H, Vinazzer H, Bergmann H. Substitution of antithrombin III in shock and DIC: a randomized study. Thromb Res 1985;39:81–9. [15] Kienast J, Juers M, Wiedermann CJ, Hoffmann JN, Ostermann H, Strauss R, et al. KyberSept investigators. Treatment effects of high-dose antithrombin without concomitant heparin in patients with severe sepsis with or without disseminated intravascular coagulation. J Thromb Haemost 2006;4:90–7. [16] Roesken F, Ruecker M, Vollmar B, Boeckel N, Morgenstern E, Menger MD. A new model for quantitative in vivo microscopic analysis of thrombus formation and vascular recanalisation: the ear of the hairless (hr/hr) mouse. Thromb Haemost 1997;78:1408–14. [17] Baez S. An open cremaster muscle preparation for the study of blood vessels by in vivo microscopy. Microvasc Res 1973;5:384–94. [18] Lindenblatt N, Menger MD, Klar E, Vollmar B. Sustained hypothermia accelerates microvascular thrombus formation in mice. Am J Physiol Heart Circ Physiol 2005;289: H2680–7. [19] Vollmar B, Schmits R, Kunz D, Menger MD. Lack of in vivo function of CD31 in vascular thrombosis. Thromb Haemost 2001;85:160–4. [20] Roesken F, Vollmar B, Ruecker M, Seiffge D, Menger MD. In vivo analysis of antithrombotic effectiveness of recombinant hirudin on microvascular thrombus formation and recanalization. J Vasc Surg 1998;28:498–505. [21] Sorg H, Hoffmann JN, Menger MD, Lindenblatt N, Goehring P, Vollmar B. Antithrombin is as effective as heparin and hirudin to prevent formation of microvascular thrombosis in a murine model. Thromb Haemost 2006;96:371–7. [22] Fiebig E, Ley K, Arfors KE. Rapid leukocyte accumulation by “spontaneous” rolling and adhesion in the exteriorized rabbit mesentery. Int J Microcirc Clin Exp 1991;10:127–44. [23] Denis C, Methia N, Frenette PS, Rayburn H, Ullman-Cullere M, Hynes RO, et al. A mouse model of severe von Willebrand disease: defects in hemostasis and thrombosis. Proc Natl Acad Sci U S A 1998;95:9524–9. [24] Weiss EJ, Hamilton JR, Lease KE, Coughlin SR. Protection against thrombosis in mice lacking PAR3. Blood 2002;100: 3240–4. [25] Schafer K, Konstantinides S, Riedel C, Thinnes T, Muller K, Dellas C, et al. Different mechanisms of increased luminal stenosis after arterial injury in mice deficient for urokinaseor tissue-type plasminogen activator. Circulation 2002;106: 1847–52.

H. Sorg et al. [26] Andre P, Prasad KS, Denis CV, He M, Papalia JM, Hynes RO, et al. CD40L stabilizes arterial thrombi by a beta3 integrin-dependent mechanism. Nat Med 2002;8:247–52. [27] Dickneite G. Antithrombin III in animal models of sepsis and organ failure. Semin Thromb Haemost 1998;24:61–9. [28] Roemisch J, Gray E, Hoffmann JN, Wiedermann CJ. Antithrombin: a new look at the actions of a serine protease inhibitor. Blood Coagul Fibrinolysis 2002;13:657–70. [29] Chang MC, Huang TF. In vivo effect of a thrombin-like enzyme on platelet plug formation induced in mesenteric microvessels of mice. Thromb Res 1994;73:31–8. [30] Sheu JR, Chao SH, Yen MH, Huang TF. In vivo antithrombotic effect of triflavin, an Arg-Gly-Asp containing peptide on platelet plug formation in mesenteric microvessels of mice. Thromb Haemost 1994;72:617–21. [31] Triantaphyllopoulos DC. Effects of human antithrombin III on mortality and blood coagulation induced in rabbits by endotoxin. Thromb Haemost 1984;51:232–5. [32] Emerson Jr TE, Fournel MA, Redens TB, Taylor Jr FB. Efficacy of antithrombin III supplementation in animal models of fulminant Escherichia coli endotoxemia or bacteremia. Am J Med 1989;87(3B):27S–33S. [33] Dunzendorfer S, Kaneider N, Rabensteiner A, Meierhofer C, Reinisch C, Romisch J, et al. Cell-surface heparan sulfate proteoglycan-mediated regulation of human neutrophil migration by the serpin antithrombin III. Blood 2001;97: 1079–85. [34] Horie S, Ishii H, Kazama M. Heparin-like glycosaminoglycan is a receptor for antithrombin III-dependent but not for thrombin-dependent prostacyclin production in human endothelial cells. Thromb Res 1990;59:895–904. [35] Uchiba M, Okajima K, Murakami K, Okabe H, Takatsuki K. Attenuation of endotoxin-induced pulmonary vascular injury by antithrombin III. Am J Physiol 1996;270:L921–30. [36] Fertmann JM, Wimmer CD, Arbogast HP, Illner WD, Tarabichi A, Calasan I, et al. Single-shot antithrombin in human pancreas-kidney transplantation: reduction of reperfusion pancreatitis and prevention of graft thrombosis. Transpl Int 2006;19:458–65. [37] Inthorn D, Hoffmann JN, Hartl WH, Muhlbayer D, Jochum M. Antithrombin III supplementation in severe sepsis: beneficial effects on organ dysfunction. Shock 1997;8:329–34. [38] Inthorn D, Hoffmann JN, Hartl WH, Muhlbayer D, Jochum M. Effect of antithrombin III supplementation on inflammatory response in patients with severe sepsis. Shock 1998;10: 90–6. [39] Ranucci M, Frigiola A, Menicanti L, Ditta A, Boncilli A, Brozzi S. Postoperative antithrombin levels and outcome in cardiac operations. Crit Care Med 2005;33:355–60.