Hemostatic Therapy in Experimental Intracerebral Hemorrhage Associated With the Direct Thrombin Inhibitor Dabigatran
Background and Purpose—Dabigatran-etexilate (DE) recently has been approved for stroke prevention in atrial fibrillation. However, lack of effective antagonists represents a major concern in the event of intracerebral hemorrhage (ICH). The aims of the present study were to establish a murine model of ICH associated with dabigatran, and to test the efficacy of different hemostatic factors in preventing hematoma growth.
Methods—In C57BL/6 mice receiving DE (4.5 or 9.0 mg/kg), in vivo and in vitro coagulation assays and dabigatran plasma levels were measured repeatedly. Thirty minutes after inducing ICH by striatal collagenase injection, mice received an intravenous injection of saline, prothrombin complex concentrate (PCC; 100 U/kg), murine fresh-frozen plasma (200 μL), or recombinant human factor VIIa (8.0 mg/kg). ICH volume was quantified on brain cryosections 24 hours later.
Results—DE substantially prolonged tail vein bleeding time and ecarin clotting time for 4 hours corresponding to dabigatran plasma levels. Intracerebral hematoma expansion was observed mainly during the first 3 hours on serial T2* MRI. Anticoagulation with high doses of DE increased the hematoma volume significantly. PCC and, less consistently, fresh-frozen plasma prevented excess hematoma expansion caused by DE, whereas recombinant human factor VIIa was ineffective. Prevention of hematoma growth and reversal of tail vein bleeding time by PCC were dose-dependent.
Conclusions—The study provides strong evidence that PCC and, less consistently, fresh-frozen plasma prevent excess intracerebral hematoma expansion in a murine ICH model associated with dabigatran. The efficacy and safety of this strategy must be further evaluated in clinical studies.
Improving treatment strategies to prevent stroke in atrial fibrillation (AF) represents a key medical challenge worldwide. Oral anticoagulation (OAC) with vitamin K antagonists reduces the relative stroke risk by >60%,1 but vitamin K antagonist have multiple undesirable properties that have resulted in undertreatment of patients at risk.2 Consequently, new oral anticoagulants have been developed that directly inhibit the key coagulation factors thrombin or factor Xa, respectively.3 The direct thrombin inhibitor dabigatran-etexilate (DE) was recently approved for stroke prevention in AF after the RE-LY trial had shown that DE is not inferior or even superior to warfarin in preventing stroke without compromising bleeding.4–6
Although the benefits of OAC outweigh the risk in AF by several fold,7 intracerebral hemorrhage (ICH) remains the most serious and lethal complication of long-term use of OAC.8 The mortality for oral anticoagulant-associated ICH (OAC-ICH) is substantially higher than that of spontaneous ICH.9,10 A major goal of ICH management is to prevent secondary hematoma growth because hematoma size affects outcome after ICH substantially.11 Current guidelines for managing OAC-ICH associated with vitamin K antagonists recommend rapid replacement of deficient coagulation factors.
Currently, a major concern regarding the new OAC is that the optimal management of OAC-ICH is unknown.12 No specific antidote is available,10 and no preclinical model has been established to test potential hemostatic strategies.
Thus, to explore these mechanisms in a preclinical setting, we established a murine model of OAC-ICH that is sensitive to pretreatment with dabigatran and assessed the kinetics of hematoma expansion after collagenase-induced ICH. The effects of dabigatran and in vitro and in vivo coagulation assays at this dose were assessed. The primary goal of the study was to assess different reversal regimes that could prevent hematoma growth.
Materials and Methods
The study was conducted following national and international guidelines for use of experimental animals. The protocols were approved by the responsible committees for animal care and use (Regierungspraesidium Karlsruhe, Germany). Three hundred ninety one male mice were used (C57BL/6; Charles River Laboratories, Sulzfeld, Germany; 10–12 weeks old; body weight, 20–25g).
Establishing Effective Systemic Anticoagulation With DE in Mice
Twenty milligrams of DE, obtained by opening commercially available capsules (Pradaxa; Boehringer Ingelheim, Ingelheim, Germany), was dissolved in 25 mL phosphate-buffered saline under stirring at 300 rpm at 50°C for 1 hour. One, 2, 4, 8, and 12 hours after intraperitoneally injecting DE in different doses (2.25, 4.5, and 9.0 mg/kg), the effects of DE on the coagulation system were assessed using in vivo and in vitro assays. At each time point, 3 nonanticoagulated mice served as control.
In vivo coagulation was assessed by measuring the tail vein bleeding time (TVBT) as described13 (Supplemental Methods; https://stroke.ahajournals.org). For assessment of in vitro coagulation, the ecarin clotting time (ECT) was chosen14 (Supplemental Methods). In additional experiments, plasma concentrations of dabigatran were measured 1, 2, 4, 8, and 12 hours after injecting DE (4.5 or 9.0 mg/kg, 3 mice per dose and time point) using liquid chromatography tandem mass spectrometry as previously described15 (Supplemental Methods).
Surgical Induction of ICH and Measuring Hematoma Size
ICH was induced in anticoagulated mice (2.25, 4.5, or 9.0 mg/kg DE intraperitoneally 1 hour before ICH induction) by intrastriatal collagenase injection as described.16 Nonanticoagulated mice served as control. Twenty-four hours after inducing ICH, hematoma size in surviving mice was measured on cryosections as described16,17 (Supplemental Methods).
Hematoma Expansion Kinetics Using Repeated MRI T2* Imaging
Animals pretreated with DE (9.0 mg/kg) and nonanticoagulated controls were repeatedly scanned in a 9.4-T rodent MR scanner (Biospec 94/20USR; Bruker Medizintechnik, Rheinstetten, Germany) 1, 3, 6, and 24 hours after inducing ICH. Serial 0.16-mm-thick T2* images were acquired using the Flash-3D sequence (repetition time, 30 ms; echo time, 3.7 ms; field of view, 2.0×2.0×1.0 cm; matrix, 256×256×64; Supplemental Methods). Hematoma size was subsequently measured by an examiner blinded to treatment allocation using ImageJ. Total hematoma volume was determined by integrating measured hypointense areas and thickness of sections.16
Effect of Different Hemostatic Agents on Hematoma Volume
Mice were pretreated with either 9.0 mg/kg or 4.5 mg/kg DE for 1 hour before inducing ICH. Nonanticoagulated mice served as control. Thirty minutes after collagenase injection, 200 μL of either saline or 1 of 3 different hemostatic agents was injected via the femoral vein: (1) prothrombin complex concentrate (PCC; Beriplex; P/N 500; CSL Behring, Hattersheim am Main, Germany; 100 U/kg); (2) murine fresh-frozen plasma (FFP; produced by centrifugation of fresh mouse blood in an EDTA-coated tube at 1500 rpm for 10 minutes, as described17; or (3) factor VIIa (FVIIa; NovoSeven; Novo Nordisk, Bagsværd, Denmark; 8.0 mg/kg). Efficacy of the hemostatic agents was assessed by measuring hematoma volume on cryosections 24 hours after inducing ICH.
Dose-Dependency of Antagonization by Prothrombin Complex Concentrate
In mice pretreated with 9.0 mg/kg DE for 1 hour, the effects of different doses of PCC (25, 50, or 100 U/kg) on systemic coagulation and hematoma expansion were examined. First, TVBT was measured 30 minutes after injecting PCC. Second, ICH was induced in anticoagulated animals, followed by intravenous injection of PCC, and hematoma size was measured 24 hours later on cryosections.
All values are expressed as mean±standard deviation. Mean values were compared using analysis of variance for multiple comparisons with post hoc Bonferroni test, and mortality data were analyzed by Pearson χ2 test. All analyses were performed using SPSS 13.0 analysis software. P<0.05 was considered statistically significant.
Establishing a Model of Anticoagulation With Dabigatran in Mice
TVBT and ECT were determined at various time points after administrating DE in anticoagulated animals and controls (Figure 1A, B).
Compared with controls (1.3±0.3 minutes), 2.25 mg/kg of DE prolonged the TVBT slightly for 2 hours (2.0±0.7 minutes; P<0.05). In animals receiving 4.5 mg/kg DE, TVBT exceeded 20 minutes at 1 hour and returned into the normal range between 4 hours and 8 hours (1.5±0.6 minutes) after injection. Indeed, 9.0 mg/kg DE resulted in a consistently prolonged TVBT >20 minutes over 4 hours, which returned to baseline between 8 to 12 hours (1.6±0.4 minutes).
ECT in controls was 2.2±0.8 minutes and was prolonged beyond the measurable range of 16.5 minutes (1000 seconds) for at least 4 hours in both anticoagulated groups. Whereas ECT of mice receiving 4.5 mg/kg DE decreased to baseline between 8 and 12 hours (2.3±0.9 minutes), The values of animals receiving 9.0 mg/kg DE were still slightly elevated at 12 hours (3.6±0.6 minutes; P<0.01).
To depict the kinetics of dabigatran, plasma concentrations in our paradigm and, particularly because some TVBT and ECT values were beyond the upper limits of tests, plasma levels of dabigatran were determined in anticoagulated mice at various time points (Figure 1C). Dabigatran levels peaked within the first 2 hours and returned gradually to baseline within 8 hours.
Hematoma Size Depends on Strength of Anticoagulation/Kinetics of Hematoma Expansion
Twenty-four hours after inducing ICH, the mean size of the hematoma was 7.8±2.9 mm3 in controls (Figure 2). Hematoma size in mice receiving 2.25 mg/kg DE (11.1±1.9 mm3) did not differ from that of the controls. In contrast, hematoma volumes in mice pretreated with 4.5 mg/kg (14.8±3.3 mm3) or 9.0 mg/kg DE (16.7±4.5 mm3), respectively, were significantly larger.
To depict the kinetics of hematoma expansion, we used repeated T2* MRI (Figure 3A). In anticoagulated mice (9.0 mg/kg), the maximal hematoma size (17.3±3.6 mm3) was reached 6 hours after inducing ICH (Figure 3B). Interestingly, 77% of the maximal volume was already reached within 1 hour and 86.7% within 3 hours. Thus, the majority of excess hematoma growth caused by DE in this model occurred in the first hour after collagenase injection. Hematoma volume measured on MR T2* images and brain cryosections, respectively, correlated well (r=0.959; P<0.01), which is consistent with our previous report.16
Effect of Different Hemostatic Agents on Hematoma Volume and Mortality
Twenty-four hours after inducing ICH, mean hematoma volume in controls was 11.9±2.7 mm3 (Figure 4C). Pretreatment with DE 9.0 mg/kg increased the hematoma size to 17.0±4.1 mm3. Hematoma growth could be effectively prevented by administering PCC (11.7±3.0 mm3; P<0.05), whereas FFP (15.0±3.2 mm3) and FVIIa (17.9±4.7 mm3) failed to reduce hematoma expansion. In the second series of experiments (Figure 4D), hematoma size was 9.8±2.0 mm3 in controls. Pretreatment with 4.5 mg/kg DE increased the hematoma volume to 13.6±2.9 mm3 (P<0.01). Injecting either FFP (9.3±1.8 mm3; P<0.001) or PCC (9.0±1.7 mm3; P<0.001) prevented hematoma growth. In contrast, FVIIa did not reduce hematoma size significantly (12.0±2.9 mm3).
The 24-hour mortality after ICH differed significantly among groups (Figure 4E). Approximately 5% of control mice died within 24 hours. In contrast, the 24-hour mortality was 30% (P<0.05) in the anticoagulated mice that had not received any hemostatic agent. Only administrating PCC significantly reduced the 24-hour mortality (4%) to the level of controls, whereas neither FFP (14%) nor FVIIa (15%) had an impact on mortality.
Dose-Dependency of Prothrombin Complex Concentrate
After pretreatment with DE (9.0 mg/kg) for 1 hour, TVBT was measured to assess whether the antagonistic effect of PCC was dose-dependent. Here, 30 minutes after intravenous injection of saline, TVBT exceeded 20 minutes. Compared with controls (1.2±0.3 minutes), TVBT in anticoagulated animals receiving 100 U/kg PCC was slightly prolonged (2.6±1.2 minutes; P<0.01). Lower doses of PCC (25 or 50 U/kg) tended to reduce TVBT, but some values were still beyond 20 minutes (Figure 5A).
Despite their different effects on TVBT, both 100 U/kg (10.2±2.2 mm3; P<0.05) and 50 U/kg (10.7±3.6 mm3; P<0.05) of PCC significantly reduced hematoma growth as compared to anticoagulated mice receiving saline (15.6±4.2 mm3; Figure 5B).
DE recently has been approved for stroke prevention in AF and has potential to become a long-term preventive medication for millions of patients with AF worldwide. However, despite its lower risk of hemorrhagic complications compared to warfarin,4 lack of an antidote or an effective antagonist is a major concern in the event of severe bleeding, including ICH.8 The present study has 4 major findings: (1) a previously established model of OAC-ICH16,18 was modified to demonstrate OAC-ICH to dabigatran; (2) high doses of DE caused an early excess hematoma expansion in this model; (3) administration of coagulation factors (particularly PCC) prevented excess hematoma expansion caused by dabigatran; and (4) PCC reversed the prolongation of bleeding time and prevented excess hematoma expansion in a dose-dependent manner and was associated with improved survival.
The experimental OAC-ICH model in the present study is based on the collagenase injection model introduced by Rosenberg.19 This model was recently modified by Foerch et al18 and further characterized by our group16 to allow the examination of pathophysiologic processes of OAC-ICH associated with warfarin. Although there can be slight variations of hematoma volume among experiments (control groups in Figures 2⇑⇑–5) that are caused by different batches of collagenase, the excess hematoma expansion caused by anticoagulants is highly reproducible. Previous studies revealed that hematoma growth caused by warfarin can be prevented by substituting hemostatic factors.17,20 Consistency between these findings and current clinical guidelines21 supports the translational relevance of this model.
To establish a similar model of OAC-ICH sensitive to dabigatran, several challenges had to be met. First, a reproducible and sufficiently long-lasting anticoagulant effect of dabigatran had to be verified in mice. Dabigatran is not enterally absorbed and the enteral absorption of its prodrug DE is low in rodents.22,23 Moreover, no bedside test is available to verify equally effective anticoagulation in individual mice. Intraperitonealy injecting the compound ensured consistent anticoagulant efficacy of dabigatran according to both ECT and TVBT, which corresponded with dabigatran plasma levels. The duration of the anticoagulant effect of a single high dose of dabigatran in this setting was 4 to 8 hours, which is shorter than the 12 to 16 hours reported in humans. Of note, the peak plasma levels in mice exceeded human therapeutic plasma levels ≈5-fold to 10-fold.23 Eventually, intraperitoneal injection of 9.0 mg/kg or 4.5 mg/kg DE was chosen for these coagulation and pharmacokinetic studies because a lower dose did not increase TVBT and hematoma volume consistently.
In patients, intracerebral hematomas frequently expand over a more prolonged period in OAC-ICH associated with warfarin compared to spontaneous ICH.24 It is unknown whether prolonged hematoma expansion also occurs in OAC-ICH associated with dabigatran. To depict the window of potential antagonistic therapies in our model, kinetics of hematoma expansion were determined using repeated in vivo MRI. We have previously shown that hypointensity in T2* MRI closely correlates with hematoma volume measured on unstained cryosections.16 Interestingly, hematoma expansion in the present study occurred mainly in the first 1 to 3 hours after collagenase injection similar to warfarin.16 Therefore, antagonistic treatments were initiated as early as 30 minutes after collagenase injection.
The main goal of the present study was to compare the efficacy of different hemostatic factors. In our previous study in OAC-ICH associated with warfarin, PCC and FFP were equally effective in preventing hematoma expansion, whereas FVIIa had only intermediate efficacy.17 In the present study, PCC was the most consistently effective hemostatic agent. With both examined doses of DE, PCC prevented excess bleeding and reduced the mortality to the level of nonanticoagulated controls. This is consistent with the results of previous studies in systemic bleeding models.12
In contrast, FVIIa failed to reduce hematoma expansion in the present experiments. Similar results were reported for the direct thrombin inhibitor melagatran in an experimental systemic bleeding model.25 The dose of human FVIIa used in the present study (8.0 mg/kg) was most effective in a murine experimental study of hemophilia.26 Inefficiency of FVIIa in the present study may result from its inability to counteract the inhibitive effect of dabigatran on factor II.
The present study has strengths and limitations. The efficacy of the different hemostatic agents was tested in 2 independent experiments using 2 different doses of DE. Furthermore, the dose-dependent effect of PCC was shown in additional experiments because PCC appeared most promising in our synopsis of experimental and clinical evidence. Although in these studies 100 U/kg of PCC reversed the effect of dabigatran on TVBT more effectively than 50 or 25 U/kg, both 100 U/kg and 50 U/kg PCC prevented excess hematoma growth. However, we did not examine whether long-term treatment with lower doses of DE could cause comparable excess hematoma expansion in the event of ICH. The mechanisms of how PCC prevents hematoma growth were not examined. We speculate that administration of PCC or FFP increases the availability of factor II, which reconstitutes the coagulation cascade. Another limitation is that we did not examine the neurological outcome. Finally, our study was not designed to answer whether repeated administration of coagulation factors is necessary in OAC-ICH associated with dabigatran in patients.
In conclusion, we provide strong evidence that administering PCC can prevent excess hematoma expansion in a murine OAC-ICH model associated with dabigatran. Because the pathophysiology of ICH in patients is only partially reflected in current experimental ICH models, the efficacy and safety of this strategy must be further evaluated in appropriate clinical studies.
Sources of Funding
R.V. is supported by an Else-Kröner-Memorial Scholarship. S.I. is supported by a Deutscher-Akademischer-Austausch-Dienst (DAAD) Scholarship.
The study was investigator-initiated. It was planned, performed, and financed independently of Boehringer Ingelheim, except for determining dabigatran plasma levels, which were measured by a third party under contract of Boehringer Ingelheim. Prothrombin complex concentrate (Beriplex) was a kind gift of CSL Behring, Germany. R.V. was an investigator in the RE-LY trial and has received speaker's honoraria, travel support, and consulting fees from Boehringer Ingelheim in the past. J.v.R is an employee of Boehringer Ingelheim.
Prothrombin complex concentrate (Beriplex) was a kind gift of CSL Behring, Germany.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.624650/-/DC1.
- Received April 28, 2011.
- Accepted August 9, 2011.
- © 2011 American Heart Association, Inc.
- Garcia D,
- Libby E,
- Crowther MA
- Steiner T,
- Rosand J,
- Diringer M
- Sjoblom L,
- Hardemark HG,
- Lindgren A,
- Norrving B,
- Fahlen M,
- Samuelsson M,
- et al
- Davis SM,
- Broderick J,
- Hennerici M,
- Brun NC,
- Diringer MN,
- Mayer SA,
- et al
- Illanes S,
- Zhou W,
- Schwarting S,
- Heiland S,
- Veltkamp R
- Foerch C,
- Arai K,
- Jin G,
- Park KP,
- Pallast S,
- van Leyen K,
- et al
- Rosenberg GA,
- Mun-Bryce S,
- Wesley M,
- Kornfeld M
- Morgenstern LB,
- Hemphill JC III.,
- Anderson C,
- Becker K,
- Broderick JP,
- Connolly ES Jr.,
- et al
- Eisert WG,
- Hauel N,
- Stangier J,
- Wienen W,
- Clemens A,
- van Ryn J
- Flaherty ML,
- Tao H,
- Haverbusch M,
- Sekar P,
- Kleindorfer D,
- Kissela B,
- et al