THERE ARE APPROXIMATELY 300,000 to 500,000 new

Transcription

1 285 Screening for Venous Thromboembolism in Traumatic Brain Injury: Limitations of D-Dimer Assay Jay M. Meythaler, MD, JD, Winfield S. Fisher, MD, Loring W. Rue, MD, Alice Johnson, BSN, CCRC, Linda Davis, RN, MSN, CRNP, Robert C. Brunner, MD ABSTRACT. Meythaler JM, Fisher WS, Rue LW, Johnson A, Davis L, Brunner RC. Screening for venous thromboembolism in traumatic brain injury: limitations of D-dimer assay. Arch Phys Med Rehabil 2003;84: From the Department of Physical Medicine & Rehabilitation (Meythaler, Johnson, Davis, Brunner) and the Divisions of Neurological Surgery (Fisher) and Trauma Surgery (Rue), University of Alabama School of Medicine, Birmingham, AL. Supported by a grant from the National Institute on Disability and Rehabilitation Research, US Department of Education (grant no. H133A980010). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Jay M. Meythaler, MD, JD, University of Alabama, R157 Spain Rehabilitation Center, th St S, Birmingham, AL , JMeythal@uab.edu /03/ $35.00/0 doi: /apmr Objectives: To assess whether 2 different D-dimer fibrin degradation assays a second-generation latex immunosorbent agglutination (LIA) and an enzyme-linked immunosorbent assay (ELISA) are predictive for the development of deep venous thrombosis (DVT) at the currently accepted level of 500 g/l of D-dimer assay during the first weeks after traumatic brain injury (TBI) and to correlate over 8 weeks the second-generation LIA assay with the ELISA assay after acute TBI. Design: A case series of persons with TBI were screened for DVT at 2 weeks ( 3d) using real-time, spectral Doppler ultrasound, as well as D-dimer fibrin split products. All persons were rescreened at 4, 6, and 8 weeks ( 3d) after injury using D-dimer LIA and ELISA assays. Setting: A university hospital with a directly connected comprehensive in- and outpatient rehabilitation center that are part of the Traumatic Brain Injury Model Systems. Participants: Over 3 years, 35 TBI subjects with a mean Glasgow Coma Scale score of 6.5 were consecutively enrolled into the trial while on acute care. Persons were at least 16 years of age with no history of treatment for DVT. Interventions: Not applicable. Main Outcome Measures: Data were analyzed for the levels of D-dimer and risk as established by a predictive value of 500 g/l. Changes in D-dimer values over time and within subjects were assessed by analysis of variance (ANOVA) with repeated measures, and the methods were correlated. Results: The mean LIA level at 2 weeks was 4.3mg/L and averaged 1.6mg/L at 8 weeks from injury (P.012, ANOVA), and the ELISA dropped from 4748 g/l to 1695 g/l (P.0022, ANOVA). Except for 1 ELISA value in 1 patient, D-dimer levels were elevated beyond 500 g/l at 2 weeks. There was a very good correlation between the LIA and the ELISA at 2, 4, 6, and 8 weeks after TBI (P.0001). In individual cases, there were only occasional discrepancies between the LIA and ELISA methods. There were no positive DVTs at 2 weeks using ultrasound, so prediction of the sensitivity and the specificity of D-dimer with DVT was not possible. Conclusion: Using the currently recommended levels of D-dimer to predict DVT is not clinically useful in the acute TBI population. Key Words: Brain injuries; Fibrin fibrinogen degradation products; Rehabilitation; Thromboembolism; Ultrasonography, Doppler by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation THERE ARE APPROXIMATELY 300,000 to 500,000 new cases of traumatic brain injury (TBI) admitted to US hospitals each year, and those requiring hospitalization are estimated to be approximately 100,000 patients per year. 1 Patients with TBI are considered to be at risk for a deep venous thrombosis (DVT) because of venous insufficiency from immobilization, hypercoaguability, or trauma-related vascular complications as defined by Virchow s triad. 2 Pathophysiologically, DVT is believed to be the major cause of pulmonary embolism (PE). 3,4 Both DVT and PE are believed to be complications that occur as a consequence of venous insufficiency associated with prolonged immobilization Trauma associated with brain injury is also believed to increase the risk for DVT. 2,7,8 Trauma to the lower extremities may produce vessel damage leading to clot formation at the injury site because of the interaction of platelets with the injured vessel wall. 2,11 Furthermore, blood loss, including that lost during surgery, and acute phase reactions generated in response to TBI lead to a hypercoaguable state contributing to the development of DVT. 12 Plasma D-dimers are generated when the endogenous fibrinolytic system degrades fibrin, as in the formation of a venous thrombosis, and they consist of 2 identical subunits derived from 2 fibrin molecules. 13 Unlike fibrinogen degradation products, which are derived from fibrinogen and fibrin, D-dimers are a specific cross-linked fibrin derivative. 14,15 The half-life is approximately 8 hours, with plasma clearance by urinary excretion and the action of the reticuloendothelial system. 13,16 The use of DVT screening tests has long been justified by the known pathophysiology of PE. 3,17 Real-time ultrasonography is presently regarded as the most sensitive and specific noninvasive test for the detection of proximal lower-extremity DVT These thrombi are typically located in the popliteal or more proximal deep veins of the legs. 21 The positive predictive value for an abnormal scan reveals a specificity of 97% to 98% for both symptomatic and asymptomatic patients, a sensitivity of 97% for symptomatic patients, but only a sensitivity of 59% for asymptomatic patients for all lower-extremity DVTs Real-time Doppler ultrasonography is a reliable noninvasive, but expensive, means of detecting DVT in the popliteal and more proximal veins, but it is less reliable in detecting thrombi in the calf. 19,21-23 Noninvasive screening with

2 286 D-DIMER IN TRAUMATIC BRAIN INJURY, Meythaler ultrasound has been justified because calf thrombi rarely lead to clinically significant emboli. 24,25 Screening for a proximal lower-extremity DVT is a recommended part of the routine clinical practice for inpatient rehabilitation TBI patients. 4 This is based on evidence that indicates the incidence of DVT is between 8.5% and 18% on admission to inpatient rehabilitation At an estimated DVT prevalence of 8.5% at time of rehabilitation admission for TBI patients, Meythaler et al 26 established that screening via ultrasound, despite its relatively high cost, was as cost effective per year of life saved as current mammography recommendations for women. With persons being transferred from acute care to rehabilitation more quickly than ever, a cost-effective method of screening for thromboembolic disease that can be used more frequently over an extended period of time would be useful in the prevention of PE. It has been suggested that screening for venous thromboembolic disease using D-dimer is a cost-effective method. 13,28-30 D-dimer levels are increased in any condition in which fibrin is formed and degraded by plasmin, and this includes trauma. 2,13,31 Although extremely sensitive, D-dimer tests are not specific, and if used as a sole method of screening or as part of a clinical paradigm, the result could be many more persons being treated for DVT than necessary. 13,26 Levels of 500 g/l of D-dimer assay have been established as being 97% sensitive for venous thromboembolic events. 13 Two laboratory procedures are available for conducting D-dimer assays: the routine latex immunosorbent agglutination (LIA) D-dimer and enzyme-linked immunosorbent assay (ELISA) D-dimer methods. The LIA D-dimer test is analyzed semiquantitatively and has a sensitivity of.25 to.50mg/l. The ELISA test gives quantitative values that are useful for comparison studies and are clinically relevant for early thrombosis detection. 28,32,33 The correlation between the 2 assays has been poor in the past, because the studies did not use the more accurate second-generation LIA assay. 13,34 The goals of this study were to assess whether D-dimer assays are predictive of DVT during the first 2 weeks after TBI at a predictive value of 500 g/l (0.5mg/L) and to correlate over 8 weeks the second-generation LIA assay with the ELISA after acute TBI. METHODS Design A prospective longitudinal study was conducted at a Traumatic Brain Injury Model Systems center. The design was a consecutive case series of persons with TBI enrolled in the study. Subjects were screened for a DVT at 2 weeks ( 3d) using real-time, spectral Doppler ultrasound and had both LIA and ELISA studies for D-dimer levels performed at weeks 2, 4, 6, and 8 ( 3d) after injury. Participants Over a 3-year period, 35 TBI subjects with a mean Glasgow Coma Scale score of 6.5 (range, 3 15) were consecutively enrolled into the study; 33 subjects had D-dimer levels drawn at 2 weeks and 21 subjects were followed for 8 weeks after injury. There were 25 men and 10 women who had an average age of 33 years (range, 17 66y), and 10 (29%) were African Americans. All persons were consecutively recruited from acute care at University of Alabama at Birmingham (UAB) hospital if they were a direct admission through the UAB emergency department to either the trauma burn intensive care unit (ICU) or the neurologic ICU. Subjects were included in the study if they were at least 16 years of age with no history of treatment for DVT, had injuries severe enough to likely require inpatient rehabilitation after 48 hours, and if informed consent was obtained. All persons were treated in the 22-bed neurologic ICU or the 12-bed trauma burn ICU with intermittent pneumatic compression, unless contraindicated by other medical treatments (eg, long bone fractures with external fixators). Setting The study was performed within the vertically integrated trauma system that includes the Birmingham Regional Emergency Medical System through the UAB emergency department. UAB Hospital, an 800-bed university teaching hospital, is a vertically integrated level I trauma center as designated by the American College of Surgeons. It is all-inclusive on a single campus, with all facilities physically linked. This includes Spain Rehabilitation Center, a 52-bed comprehensive in- and outpatient rehabilitation facility. Main Outcome Measures Data were analyzed for the levels of D-dimer and risk as established by previous clinical normative values. 13 Patients were also evaluated for DVT in the legs at 2 weeks after injury using duplex ultrasound. All testing was performed at the UAB Hospital laboratory with the commercial tests described later. D-dimer assays. For purposes of this study, the LIA D- dimer test used was the serum D-dimer assay (STA -Liatest line a ). This LIA method is a second-generation D-dimer test for the quantitative determination of D-dimer in blood plasma by the immunoturbidimetric method (for use with STA analyzers a ). The detection threshold is 22 /ml. The test was performed on citrated venous blood plasma. The ELISA method used in this study was the serum D- dimer assay (Asserachrom a ). This is an accurate quantitative ELISA method with a detection limit of 5ng/mL. The test was also performed on citrated venous blood plasma. All persons were screened at 2, 4, 6, and 8 weeks ( 3d) after injury using the serum D-dimer assay via the LIA and ELISA methods. The plasma was drawn into a citrated tube by a laboratory technician while the study participant was an inpatient or an outpatient if the person had been discharged. All D-dimer tests were run on the same day using the methods outlined by the commercial manufacturer, and all plasma tests were performed at the University Hospital coagulation laboratory. Generally, the ELISA methods are considered to be more accurate than the LIA methods for detection of D-dimers assays. However, the second-generation LIA methods, which use whole blood agglutination or immunoturbidimetric techniques, have been clinically useful. 33 Duplex Ultrasonography All persons with TBI enrolled in the study were screened for DVT at 2 weeks ( 3d) using real-time spectral Doppler and color Doppler ultrasonography, as well as D-dimer assay. These ultrasound tests were performed either while the person was an inpatient or an outpatient if the person had been discharged. Ultrasonography was performed by a certified technician, and the ultrasounds were read by an experienced boardcertified radiologist. The sensitivity and specificity for DVT detection from the popliteal area to the groin is 97% at our institution. Statistical Analysis The statistical design is an open linear study with each patient used as his/her own control. Data for the LIA and

3 D-DIMER IN TRAUMATIC BRAIN INJURY, Meythaler 287 Fig 1. The regression coefficients between the LIA and the ELISA methods at (A) 2 weeks (P.0029), (B) 4 weeks (P.0008), (C) 6 weeks (P<.0001), and (D) 8 weeks (P<.0001). This data show that there are times when some rare discrepancies arise between the LIA and ELISA methods. The LIA method is presented in milligrams per liter; the ELISA method is presented in micrograms per liter. ELISA D-dimer tests were evaluated with descriptive statistics and simple regression analysis. Changes over time were analyzed with repeated-measures analysis of variance (ANOVA) over the 8-week period and with the paired Student t test for specific periods of time. A P value less than.05 was considered significant. Data are presented as means with standard deviations to facilitate the interpretation of the magnitude and clinical significance of our findings. Laboratory standards reported at the University Hospital for each test have established that a level of 0.5mg/L via the LIA method or 500 g/l via the ELISA method are indicative of a thromboembolic event. Statistical analysis was performed with StatView, version 5.0, b for Macintosh computers. Protocol Review The study was reviewed and approved by the University of Alabama at Birmingham Institutional Review Board for Human Use. After being familiarized with the experimental method and the potential risks as well as the potential benefits of the procedure, each patient or his/her respective legal guardian as determined by Alabama state law signed an informed consent form before testing. RESULTS Evaluation of D-Dimer Testing With Time At 2 weeks, all 29 patients evaluated by the LIA method were above 0.5mg/L using the ELISA method; 19 patients (n 20) had a D-dimer level above 500 g/l (fig 1A). No DVTs were detected by ultrasound evaluation at 2 weeks, so prediction of the sensitivity and the specificity of D-dimer with DVT was not possible with the D-dimer assay. The majority of patients at 4 and 6 weeks after injury sustained elevated levels of D-dimer (figs 1B, 1C). At 8 weeks, there were still 12 subjects above 0.5mg/L (57%, n 21) via the LIA test and 7 subjects above 500 g/l (43%, n 16) via the ELISA test, levels that are indicative of a possible thromboembolic event (fig 1D).

4 288 D-DIMER IN TRAUMATIC BRAIN INJURY, Meythaler Fig 2. Change over time of the LIA measurement for D-dimer over 2 to 8 weeks. LIA is represented in milligrams per liter. D-dimer testing using the LIA and the ELISA methods revealed a continuous drop in D-dimer levels as the time interval between trauma and test date increased (figs 2, 3). The LIA levels fell from mg/L at 2 weeks postinjury to mg/L (P.012, ANOVA) at 8 weeks postinjury, and the ELISA levels dropped from g/l at 2 weeks postinjury to g/l at 8 weeks postinjury (P.0022, ANOVA). There was a significant decrease in the D-dimer levels over time during each successive 2-week period (figs 1A D). The D-dimer level obtained with the LIA test dropped from mg/L to mg/L at 2 weeks (P.0156, paired Student t test), and those obtained with the ELISA method dropped from g/l to g/l (P.0046, paired Student t test). Correlation Between the LIA and ELISA Methods One goal of this study was to correlate the D-dimer results obtained with the LIA and ELISA methods of D-dimer detection in the TBI population. There was a strong correlation at 2 weeks (P.0001, r.948), 4 weeks (P.0001, r.741), 6 weeks (P.0001, r.954), and 8 weeks (P.0001, r.935) after TBI (figs 1A D). In individual cases, there were only rare discrepancies between the LIA and ELISA methods (figs 1A D). Many of the patients maintained D-dimer levels indicative of a thromboembolic event at our institution by either test (figs 2, 3). DISCUSSION This study indicates that the currently accepted D-dimer levels for thromboembolic events of 500 g/l 13 may not be useful in the acute TBI population. This is in contrast to previous reports that D-dimer levels might be useful in the prediction of venous thromboembolism in TBI patients. 28 Almost all TBI patients had D-dimer levels indicative of a thromboembolic event at 2 weeks after injury and the condition continued for up to 8 weeks in many of the patients (figs 2, 3). Yet the incidence of DVT in this population was much lower Most patients who develop DVT after TBI do so in the first weeks after injury. 2,26-28 There must be other secondary biochemical and medical factors in acute TBI patients that contribute to a relatively low 8% to 18% DVT rate noted in the severe TBI population The decrease in D-dimer levels over time was expected, yet most of the patients with TBI still had elevated D-dimer levels at 8 weeks postinjury (fig 1D). This is of clinical significance because the use of D-dimer assay has been expanded to other patient populations based on the supposition that this test can be generalized to all patient populations that are predisposed to PE and DVT. 13,29,33 However, at the mean D-dimer levels detected at 2 weeks and 2 months from injury, the 500 g/l D-dimer level had no utility in diagnosing DVT in the acute TBI population. There was a strong correlation between the LIA and ELISA methods (r 0.9) at each 2-week interval of the study (figs 1A D). This contradicts some studies and may be a reflection of the high levels of D-dimer in these patients or attributable to the more accurate second-generation LIA test using microagglutination techniques with immunoflouescence. 13 It is possible that, at these high levels of D-dimer, there will be a good correlation between the ELISA and most LIA tests. Clearly the previously established levels are not relevant in the population with TBI. 13,15,28-30,34 It is possible that, at a new higher cutoff level, D-dimer would be more predictive. A study in patients admitted to rehabilitation from stroke was also noted to have similarly increased D-dimer levels. 35 Harvey et al 35 estimated that a D-dimer level of 1591 g/l had a 96% negative predictive value. However, when detailed statistical analyses were performed, Harvey concluded that D-dimer only had a moderate ability to identify patients with and without DVT. 35 There were not enough subjects in the present study to attempt to develop new standards to assess sensitivity and specificity for this TBI patient population using either D-dimer method. At an incidence of 8.5% at our institution, 26 it would require years of data collection to develop new standards with D-dimer in the TBI population. A larger multicenter study will be required to determine new sensitivity and specificity values of D-dimer assay in acute TBI. Prophylactic anticoagulation is the most cost-effective way to prevent DVT development in many conditions, 17,24 but patients with TBI are often not considered candidates for anticoagulation until several weeks postinjury because they have undergone neurosurgical procedures or have had a recent intracerebral hemorrhage. Some investigators 36,37 have suggested that anticoagulation in the first 2 weeks post-tbi may be feasible. But there are no studies to establish that early use of anticoagulants in patients with TBI does not carry an unwarranted risk of increased intracerebral hemorrhage. 36 With re- Fig 3. Change over time of the measurement for D-dimer over 2 to 8 weeks using the ELISA method of analysis. ELISA is represented in micrograms per liter

5 D-DIMER IN TRAUMATIC BRAIN INJURY, Meythaler 289 gard to spontaneous intracerebral hemorrhage, it has been estimated 36,37 that between 9% to 23% of persons were on anticoagulation treatment at the time of hemorrhage. Because the timing of safe anticoagulation after a brain injury needs to be delineated, early diagnosis of thromboembolic disease is still the first line of defense in the acute TBI population. 2,26 Early use of sequential intermittent compression on the lower extremities, although potentially useful in those with orthopedic conditions, also has not been specifically tested in the TBI patient population. All of our patients had intermittent compression, if possible, in the ICUs. The acute TBI patient population may have stasis or hypercoaguable syndromes as the predominant cause of DVT rather than endothelial damage to the venous system from direct trauma, as is noted in the orthopedic trauma patients. The effectiveness of these devices has not been systematically studied in the TBI patient population. Indeed, leg compression devices in the population with acquired brain injury are controversial because the evidence as to their effectiveness is conflicting, depending on the type of device used. 38 Hence, early DVT diagnosis and treatment with anticoagulation, or inferior vena cava filter placement, is still the primary method to avoid fatal pulmonary embolus in this population. 26 CONCLUSION The elevated D-dimer levels are indicative of a coagulopathy after TBI and may be of some value regarding neurologic outcome. 39,40 It is likely that D-dimer levels are a marker indicating the level of the secondary inflammatory processes involving neuronal cell death after TBI. Patients with higher levels had poorer outcomes. 39,40 The fact that there is not a more significant number of venous thromboembolic disease in the TBI population is puzzling. 26,27,36 More research is required on the factors that cause coagulopathy with and without proximal leg DVT in the population with TBI. References 1. Thurman DJ, Guerrero J. Trends in hospitalization associated with traumatic brain injury. JAMA 1999;282: Geerts WH, Code KI, Jay RM, Chen E, Szalai JP. A prospective study of venous thromboembolism after major trauma. N Engl J Med 1994;331: National Institutes of Health Consensus Development Conference. Prevention of venous thrombosis and pulmonary embolism. JAMA 1986;256: Brandstater ME, Roth EJ, Siebens HC. Venous thromboembolism in stroke: literature review and implications for clinical practice. Arch Phys Med Rehabil 1992;73(5 Suppl):S Green D, Lee MY, Ito VY, et al. Fixed- vs adjusted-dose heparin in the prophylaxis of thromboembolism in spinal cord injury. JAMA 1988;260: Lamb GC, Tomski MA, Kaufman J, Maiman DJ. Is chronic spinal cord injury associated with increased risk of venous thromboembolism? J Am Paraplegia Soc 1993;16: Coon WW. Venous thromboembolism: prevalence, risk factors, and prevention. Clin Chest Med 1984;5: Sevitt S, Gallagher N. Venous thrombosis and pulmonary embolism: a clinico-pathological study in injured and burned patients. Br J Surg 1971;48: Sevitt S. Venous thrombosis in injured patients with some observations on pathogenesis. In: Sherry S, Brinkhouse KM, Genton E, et al, editors. Thrombosis. Washington (DC): Natl Acad Sci; p Sasahara AA, Sharma GV, Paris AF. New developments in detection and prevention of venous thromboembolism. Am J Cardiol 1979;43: Stamatikis JD, Kakkar VV, Sagar S, Lawrence D, Nairn K, Bentley PG. Femoral vein thrombosis and total hip replacement. BMJ 1977;2: Gitel SN, Salvati EA, Wessler S, Robinson HJ, Worth MH. The effect of total hip replacement and general surgery on antithrombin III in relation to venous thrombosis. J Bone Joint Surg Am 1979;61: Kelley J, Rudd A, Lewis RR, Hunt BJ. Plasma D-dimers in the diagnosis of venous thromboembolism. Arch Intern Med 2002; 162: Kario K, Matsuo T, Kobayashi H. Which factors affect D-dimer levels in the elderly? Thromb Res 1991;62: Crippa L, D Angelo A. The utility and cost-effectiveness of D- dimer measurements in the diagnosis of deep vein thrombosis. Haemotologica 1997;82: Hager K, Platt D. Fibrin degeneration product concentrations (D-dimers) in the course of aging. Gerontology 1995;41: Oster G, Tuden RL, Colditz GA. A cost-effectiveness analysis of prophylaxis against deep-vein thrombosis in major orthopedic surgery. JAMA 1987;257: Oudkerk M, van Beek EJ, van Putten WL, Buller HR. Costeffectiveness analysis of various strategies in the diagnostic management of pulmonary embolism. Arch Intern Med 1993;153: Lensing AW, Prandoni P, Brandjes DP, et al. Detection of deepvein thrombosis by real-time B-mode ultrasonography. N Engl J Med 1989;320: Pedersen DM, Aslaksen A, Vik-Mo H, Bassoe AM. Compression ultrasonography in hospitalized patients with suspected deep venous thrombosis. Arch Intern Med 1991;151: Weinmann EE, Salzman EW. Deep-vein thrombosis. N Engl J Med 1994;331: Heijboer H, Buller HR, Lensing AW, Turpie AG, Colly LP, Cate JW. A comparison of real-time compression ultrasonography with impedance plethysmography for the diagnosis of deep-vein thrombosis in symptomatic outpatients. N Engl J Med 1993;329: Verstraite M. The diagnosis and treatment of deep-vein thrombosis. N Engl J Med 1993;329: Hull RD, Hirsh J, Carter CJ, et al. Diagnostic efficacy of impedance plethysmography for clinically suspected deep-vein thrombosis: a randomized trial. Ann Intern Med 1985;102: Rosberg B. Blood coagulation during and after normovolemic hemodilution in elective surgery. Ann Clin Res 1981;13 Suppl 33: Meythaler JM, DeVivo MJ, Hayne JB. Cost-effectiveness of routine screening for proximal deep venous thrombosis in acquired brain injury patients admitted to rehabilitation. Arch Phys Med Rehabil 1996;77: Cifu DX, Kaelin DL, Wall BE. Deep venous thrombosis: incidence on admission to a brain injury program. Arch Phys Med Rehabil 1996;77: Yablon SA, McGrath CM. Utility of the D-dimer assay in the screening of venous thromboembolic disease amount patients with acquired brain injury [abstract]. Arch Phys Med Rehabil 1997:78: Perrier A, Buswell L, Bounameaux H, et al. Cost-effectiveness of noninvasive diagnostic aids in suspected pulmonary embolism. Arch Intern Med 1997;157: Perrier A, Bounameaux H, Marabia A, et al. Diagnosis of pulmonary embolism by a decision analysis based strategy including clinical probability, D-dimer levels, and ultrasonography: a management study. Arch Intern Med 1996;156: Meythaler JM, DeVivo MJ, Renfroe SG, Fisher W, Rue LW. Increased incidence of DVT on admission of inpatient rehabilitation in acquired brain injury secondary to intermittent pneumatic compression hose [abstract]. Arch Phys Med Rehabil 1998;79: Becker DM, Philbrick JT, Bachhuber TL, Humphries JE. D-dimer testing and acute venous thromboembolism. A shortcut to accurate diagnosis? Arch Intern Med 1996;156: Freyburger G, Trillaud H, Labrouche S, et al. D-dimer strategy in thrombosis exclusion a gold standard study in 100 patients suspected of deep venous thrombosis or pulmonary embolism: 8 DD methods compared. Thromb Haemost 1998;79:32-7.

6 290 D-DIMER IN TRAUMATIC BRAIN INJURY, Meythaler 34. van der Graaf F, van den Borne H, van der Kolk M, de Wild PJ, Janssen GW, van Uum SH. Exclusion of deep venous thrombosis with D-dimer testing. Thromb Haemost 2000;83: Harvey RL, Roth EJ, Yarnold PR, Durham JR, Green D. Deep vein thrombosis in stroke: the use of plasma D-dimer level as a screening test in the rehabilitation setting. Stroke 1996;27: Steinberg DP, Green D. Prophylactic anticoagulation in patients with intracranial hemorrhage: a retrospective review of bleeding risk [abstract]. Arch Phys Med Rehabil 1993;64: Fogelholm R, Eskola K, Kiminkinen T, Kunnamo I. Anticoagulation treatment as a risk factor for primary intracerebral haemorrhage. J Neurol Neurosurg Psychiatry 1992:55: Yablon SA, McGrath CM. Utility of the D-dimer assay in the screening of venous thromboembolic disease amount patients with acquired brain injury [abstract]. Arch Phys Med Rehabil 1997:78: Takahashi H. Fibrinolytic parameters as an admission prognostic marker of head injury in patients who talk and deteriorate. J Neurosurg 1997;86: Bredbacka S. Soluble fibrin and D-dimer as detectors of hypercoaguability in patients with isolated brain trauma. J Neurosurg Anesthesiol 1994;6: Suppliers a. Diagnostica Stago SA, 9 rue des Freres Chausson, Asnieres F France. b. Abacus Concepts Inc, 23 Ballas Ln, Ste 100, Piscataway, NJ