Clinical and Biochemical Analysis in Infection-Associated Stroke
Background and Purpose Currently, recent infection (primarily bacterial infection) is discussed as a risk factor for cerebrovascular ischemia. The aim of this study was to investigate whether the association of ischemic stroke with recent infection is restricted to stroke subtypes and whether recent infection influences the severity of the postischemic deficit; we also aimed to define biochemical pathways linking infection and ischemic stroke.
Methods Analyzing the data of a prospective case-control study, we classified the etiology of cerebrovascular ischemia on the basis of clinical, neuroradiological, sonographical, cardiological, and biochemical data in 159 patients without and in 38 patients with infection within 1 week before ischemia. We assessed the severity of neurological deficits using the Scandinavian Stroke Scale.
Results In patients with recent infection compared with patients without infection, the neurological deficit on admission was more severe (median of scores, 41 versus 30.5; P<.005), cortical infarcts in the middle cerebral artery territory were more frequent (60% versus 26%; P<.001), the prevalence of extracranial artery stenoses was lower (9% versus 26%; P<.05), and definite or presumed cardioembolic stroke was more frequent (34% versus 19%; P<.05), as was stroke from cervical artery dissection (8% versus 1.3%; P=.05). Serum levels of C-reactive protein were higher in patients with (20.7±26.8 mg/L) than in those without infection (9.2±23.7 mg/L; P<.01).
Conclusions Recent infection may be associated with a more severe postischemic deficit and with an increased risk of stroke from cardioembolic origin and from cervical arterial dissection.
Recent infection is currently discussed as a risk factor for ischemic stroke.1 2 In a case-control study we found that within 1 week before cerebrovascular ischemia or examination, more patients (38 of 197; 19.3%) than control subjects randomly selected from the population (10 of 197; 5.1%) had suffered from an infection (odds ratio, 4.5; 95% confidence interval, 2.1 to 9.7). Respiratory infection and infection of bacterial origin dominated among patients. Infection remained a significant risk factor when other stroke risk factors were included as covariates in a logistic model. None of the patients had suffered from endocarditis, syphilis, bacterial (including tuberculous) meningitis or herpes zoster of neck and face, or infectious conditions known to be associated with ischemic stroke. The role of infection was not restricted to younger stroke patients, and the profile of vascular risk factors was similar in patients with and without recent infection.3 Therefore, infection may mainly further increase a preexisting higher risk for ischemic stroke. The pathogenetic link between recent infection and cerebrovascular ischemia is incompletely understood, however. In this respect it is important to investigate whether the association of infection and cerebral ischemia is restricted to anatomic and etiologic stroke subtypes. In the present study we analyze neuroradiological, neurosonological, and cardiological data and the etiology of cerebral ischemia in 197 patients of our recent case-control study. In a preliminary investigation with subgroups of our patients, we assess various inflammatory and coagulant parameters and try to illuminate biochemical and hematologic mechanisms linking infection and ischemic stroke. Results from animal experiments indicate that hyperthermia enhances ischemic brain damage.4 Therefore, we tested the hypothesis that a recent infection increases the postischemic neurological deficit and that increased body temperature mediates such aggravating effects.
Subjects and Methods
During 1 year we performed a case-control study to investigate whether a recent infection is a risk factor for cerebrovascular ischemia. We investigated 197 patients between 18 and 80 years of age who were residents of the city of Heidelberg or of two neighboring counties. The patients had been transferred to the neurological emergency unit of the University of Heidelberg for primary diagnostic and therapeutic procedures before being referred to hospitals in the vicinity or being admitted to wards of our hospital. The patients (83 women and 114 men; mean±SD age, 63.1±10.9 years) had suffered from transient (n=20) or persistent (n=177) cerebral ischemia. The control subjects were randomly selected from the population and were matched one-to-one with patients for sex, age (±2 years), and area of residence. In all participating subjects, we carefully evaluated the medical history with a particular focus on signs of recent infection. All subjects received a physical examination, and antimicrobial serum analysis was done on first examination and repeated 10 to 18 days later. We diagnosed an infection if fever (≥38.0°C), subfebrile temperature (37.5°C to 38.0°C), or a serological or cultural finding proving an infection was present in combination with at least one symptom typical for an infection. In addition, we acknowledged combinations of symptoms typical for a local infection. The design of this study has been reported in more detail previously.3 The study was approved by the institutional review committee of the University of Heidelberg, and subjects gave informed consent.
To classify phenomenological and etiologic subtypes of cerebrovascular ischemia, we used clinical findings and the results of ancillary diagnostic studies. All patients had a cerebral CT scan to exclude a primary cerebral hemorrhage. A second CT or an MRI scan was performed in 55 patients without an ischemic lesion corresponding to clinical symptoms on first examination. We performed at least one extracranial Doppler ultrasound examination in 177 patients and at least one transcranial Doppler ultrasound examination in 138 patients. Forty-nine patients received a cranial angiography. An electrocardiogram was available in 173 patients; in addition, 127 patients received a transthoracic or transesophageal echocardiogram.
In regard to etiology, we categorized the following subtypes of cerebrovascular ischemia: (1) embolism from large-artery atherosclerosis: stenosis with more than 50% diameter reduction of a brain-supplying artery at a site typical of atherosclerosis, as evidenced by Doppler ultrasound or angiography. Cerebral infarcts had to be of the territorial type and clearly larger than 1 cm in diameter. Clinical symptoms, stenosis, and area of infarction had to correspond. Cardiac sources of embolism had to be absent; (2) cardioembolism: high- or medium-risk sources of cardiac emboli according to the classification of the Trial of Org 10172 in Acute Stroke Treatment (TOAST) investigators5 with exclusion of other potential etiologies, eg, large-vessel disease; (3) thromboembolism of undetermined origin: cerebral infarction in supply territories of large intracranial arteries with either no or at least two identifiable sources of embolism (eg, combined cardioembolism and large-vessel disease); (4) small-vessel disease: presence of one of five lacunar syndromes (pure motor stroke, pure sensory stroke, sensorimotor stroke, ataxic hemiparesis, dysarthria–clumsy hand syndrome), absence of acute neuropsychological deficits, and infarction size of no more than 1 cm in diameter on cranial CT or MRI scans. A normal CT scan was accepted in case of expected lower brain stem lacunes. Other potential etiologies had to be absent; (5) cerebral ischemia of hemodynamic origin: high-grade stenosis (≥80%) or occlusion of an extracranial artery and extraterritorial infarcts (border-zone or terminal supply area infarcts) on neuroimaging. Insufficient collateral blood flow as evidenced by transcranial Doppler ultrasound or by angiography supported this diagnosis; (6) CAD: angiogram or MRI of the neck showing the typical findings of a dissection, such as a mural hematoma or a stenosis or occlusion with an irregular arterial vessel wall; (7) cerebral vasculitis: abnormal immunologic serum parameters, typical neuroradiological findings (MRI, angiogram) (and results of a postmortem examination in one case); (8) cerebrovascular ischemia of other determined origin: this category comprised one patient with cerebral thrombosis associated with cancer; and (9) cerebrovascular ischemia of unknown origin: patients with either incomplete work-up or undetermined etiology despite extensive investigations.
We assessed the severity of neurological deficits using the SSS (maximum points, 48).6 From the analysis of score values we excluded patients admitted later than 48 hours after ictus, patients with transient ischemic attack, and patients who had received fibrinolytic therapy (n=52). In a subgroup of patients who were not rapidly transferred to other hospitals, we evaluated the development of neurological deficits during the first week after ictus. Patients who had an infection after stroke, mainly a nosocomial infection, were excluded from the analysis of follow-up examinations.
To further characterize potential pathogenetic pathways in infection-associated stroke, we analyzed several routine biochemical variables. In subgroups of patients with and without previous infection, we determined additional parameters including the following: CRP (Beckman), fibrinogen, fibrin d-dimer, antithrombin-3 (Boehringer), thrombin-antithrombin complexes, prothrombin fragment F1+2, plasminogen, PAI-1 (Behring), and thrombomodulin (Diagnostica Stago). From these analyses we excluded all patients with a venipuncture later than 48 hours after ictus and all patients with noninfectious inflammatory or malignant diseases or with acute vascular diseases or trauma within the last month. Since we collected the samples as soon as possible after the admission of a patient, the time of venipuncture was not kept constant.
We compared technical results in patients with and without infection using Fisher’s exact test. To compare biochemical parameters and results from the SSS we used the Mann-Whitney U test, and for clinical follow-up assessments we applied the Wilcoxon signed rank test. We analyzed the influence of different factors on results of the SSS by an ANOVA model.
Patients with infection within the preceding week had a more severe neurological deficit on admission than patients without infection as assessed by the SSS (P=.0014). This difference was still present when patients with and without recent infection were reexamined 3 and 7 days after admission. Patients without infection showed higher SSS scores on both follow-up examinations compared with admission. Similarly, patients with nonfebrile infection improved during the first week, but this did not reach statistical significance, probably because of the low number of investigated subjects. The subgroup of patients with febrile infection who were admitted to our hospital and could be followed up did not improve during the first week (Table 1⇓). In an analysis of the influence of various factors on the extent of the neurological deficit on admission, recent infection was negatively correlated with clinical status (P=.0034), as were blood glucose levels above 1.5 g/L, leukocyte counts greater than 1010/L, and an age of 65 years or older (Table 2⇓).
There were no differences between groups in respect to the localization of cerebral ischemia in the territory of the left (47% versus 41%, patients with versus without infection, respectively) or right carotid (29% versus 35%) or the vertebrobasilar arteries (21% versus 22%). Stroke patients with recent infection relatively more often had a cortical infarct in the territory of the MCA (P<.001). No significant differences could be detected in respect to other anatomic infarct types (Table 3⇓). Patients with cortical MCA infarcts had a severe clinical deficit on admission regardless of whether or not they had a recent infection (Table 1⇑). In an ANOVA model with two factors, cortical MCA infarcts were significantly correlated with lower SSS values (P<.0001), whereas febrile infection had no significant impact (P=.056).
In Doppler ultrasound examinations, stenoses of proximal segments of brain-supplying arteries, which are almost exclusively of atherosclerotic origin, occurred significantly more often in patients without than in patients with previous infection (P=.039). In contrast, we found a pattern of increased distal resistance in the ICA relatively more often among patients with than among those without infection (P<.001) (Table 4⇓). In 18 of 20 patients with increased distal resistance in the ICA, this Doppler ultrasound pattern was related to the acute ischemic event; 6 of these patients (2 with and 4 without infection) had suffered from cardioembolism, 5 (3 with and 2 without infection) had emboli of undetermined origin, 5 (3 with and 2 without infection) had ICA dissections, and 2 (2 with and 0 without infection) had a cerebral vasculitis. There were no significant differences between groups in the results of transcranial ultrasound investigations and in the relatively small series of patients with angiography (n=49) (data not shown).
In respect to the etiology of cerebrovascular ischemia, patients with recent infection suffered more often from cardioembolism (P=.040) than patients without infection (Table 5⇓). In the subgroup with febrile infection (≥38°C) (19 of 197), almost half of the patients (9 of 19) suffered from cardioembolic stroke. The diagnosis of a cardiac source of stroke was based on AF (n=33; 10 versus 23, patients with infection versus patients without infection), an akinetic or hypokinetic left ventricular segment (n=6; 2 versus 4), a left ventricular or atrial thrombus (n=2; 2 versus 0), sick sinus syndrome (n=2; 0 versus 2), mitral annulus calcification (n=2; 0 versus 2), recent myocardial infarction (n=2; 1 versus 1), mitral valve prolapse (n=1; 0 versus 1), patent foramen ovale (n=1; 0 versus 1), congestive heart failure (n=1; 1 versus 0), dilative cardiomyopathy (n=1; 0 versus 1), and a prosthetic cardiac valve (n=1; 0 versus 1). Four of the patients with AF had mitral regurgitation (0 versus 4), one had mitral stenosis (0 versus 1), and 6 (3 versus 3) showed a dilated left atrium without definite mitral valve abnormalities on echocardiogram; 7 patients (2 versus 5) suffered from intermittent AF. AF tended to occur more often in patients with (10 of 38) than in those without infection (23 of 159; P=.079). The diagnosis of cardioembolism was mainly made by means of clinical data, electrocardiography, and transthoracic echocardiography. Two of the patients with a diagnosis of cardioembolism (1 with and 1 without infection) received transesophageal echocardiography, in both of whom it was diagnostic. Patients with infection suffered less frequently from embolism caused by large-artery atherosclerotic disease than noninfected patients, but this difference did not reach significance. Small-vessel disease was diagnosed as the cause of acute cerebrovascular ischemia only in patients without recent infection.
Three of the 5 patients with CAD had suffered from a recent infection, and in 1 additional patient there was a probable preceding infection (Table 5⇑). Dissections affected the ICA in all cases and were diagnosed by angiography (n=3) or by MRI of the neck (n=2). None of the patients with dissection (3 women, 2 men, aged 36 to 56 years) suffered from hypertension, diabetes mellitus, or coronary heart disease, 3 of them were smokers, and 1 had a high consumption of alcohol (>50 g of pure alcohol daily). These patients had not suffered from any recent major trauma. The 3 patients with infection-associated CAD had a gastrointestinal infection with vomiting and diarrhea caused by Salmonella enteritidis, an infection with influenza A viruses with cough among the symptoms, and an upper respiratory tract infection with increased titers against respiratory syncytial viruses (no cough reported). One of the 2 other patients with a dissection had vomited several times on the day before stroke and developed diarrhea during the first days after stroke; a causative microbial agent could not be detected. This patient may have had a gastrointestinal infection; in our statistical analyses (Table 5⇑) she was part of the group without infection.
Both patients with infection and stroke from cerebral vasculitis did not suffer from preexisting autoimmune disorders. A 53-year-old male patient with a febrile gastrointestinal infection died soon after brain stem and cerebellar stroke. A postmortem examination revealed widespread vasculitis and intravascular coagulation in brain-supplying blood vessels. A 25-year-old female patient developed a right parietal infarction after bronchitis. Doppler ultrasound detected a right carotid siphon stenosis, MRI did not indicate a dissection, and biochemical analysis suggested an acute inflammatory and allergic reaction (eg, IgE, 1343 U/mL [normal range, <100 U/mL]).
Table 6⇓ summarizes the results from analyses of biochemical variables. Only the levels of CRP showed significant differences between patients with and without infection. CRP was higher in patients without infection (n=118; 9.2±23.7 mg/L) than in control subjects without infection (n=166; 2.5±4.5 mg/L; P<.0001). There was no significant difference in fibrinogen levels between both groups (P=.083). This may be due to a low number of subjects in the group of patients with infection. We had to exclude a high percentage of subjects from the analysis of fibrinogen levels because a heparin-sensitive method was used for fibrinogen assessment, and early treatment with heparin is frequently used in our hospital. The markers of coagulation and fibrinolysis (antithrombin-3, plasminogen, PAI-1, thrombin-antithrombin complexes, prothrombin fragment F1+2, and fibrin d-dimer) and the marker of endothelial damage (thrombomodulin) were not different between groups.
Our results indicate that a recent infection is associated with more severe stroke on hospital admission. It is known from animal experiments that raised body temperature is a deleterious factor in ischemic stroke, and hypothermia has been discussed as a potential therapy in ischemic stroke.4 Therefore, increased body temperature may have caused the greater severity of stroke associated with infection. Our data suggest, however, that on admission the worse clinical status of patients with infection is rather caused by the predominance of cortical MCA infarcts in this group compared with patients without infection. Ameriso et al2 also reported a trend for a predominance of cortical infarcts over small deep infarcts in patients with recent infection but no difference in the extent of the neurological deficit between patients with and without prior infection. The results of our follow-up examinations may suggest that fever delays or reduces the recovery after stroke, but further studies with larger numbers of patients are required to confirm this hypothesis. We acknowledge that the SSS is not a metric scale and that the interpretation of the estimates in Table 2⇑ is somewhat arbitrary; however, we think that the analysis correctly shows that infection within 1 week significantly decreases the SSS value after simultaneous adjustment for other factors. In accordance with previous reports, our analysis demonstrates an association of high blood glucose levels7 and high leukocyte counts8 with a poor clinical status in acute stroke.
The occlusion of the MCA or its branches, causing cortical infarction, appears to be due mostly to embolic mechanisms and only rarely to local atherothrombotic disease. Therefore, the high percentage of cortical MCA infarcts may indicate a dominant role of embolic infarcts in infection-associated stroke. Additionally, the results from Doppler ultrasound studies point to the distal part of the ICA as a common site of occlusion or stenosis in infection-associated stroke; this localization of the vascular pathology is in our experience primarily associated with embolism, dissection, or vasculitis. Moyamoya disease is another entity with an involvement of the distal ends of the ICA and a frequent observation of recent infection.9 Moyamoya disease is rare in western Europe, and there was no case with this diagnosis among our patients.
Atheromatous plaques in the large extracranial arteries are an important source of cerebral emboli. Atherosclerotic lesions contain considerable numbers of macrophages, which play an important role in the pathogenesis of atherosclerosis.10 Recently, Tipping et al11 reported a high procoagulant potential of macrophages from carotid plaques, which had previously generated cerebral emboli. We had hypothesized that during infection the macrophage-associated pathways of coagulation may be stimulated and that artery-to-artery emboli may occur more frequently in infection-associated stroke. Our data, however, do not support this hypothesis. Patients with infection less frequently had stenoses of the large extracranial vessels, and artery-to-artery emboli tended to be less common among patients with than among those without infection. Recent infection was significantly associated with cardioembolic stroke. The single most important cardiac source of embolism was AF, a common and important risk factor for stroke.12 Alterations in hemostatic function can be detected in patients with nonvalvular AF, and these coagulant abnormalities may contribute to the increased risk of stroke in AF.13 We hypothesize that the procoagulant state in patients with AF is increased during infection and that infection may therefore temporarily increase the risk for stroke in AF and possibly in other cardiac sources of embolism. It can be argued that the diagnosis of a cardiac source of stroke may be overestimated in the subgroup of patients with infection because physicians may ask more easily for transesophageal echocardiography in patients with fever or with a history of recent infection. In our study only two patients with cardioembolic stroke had received a transesophageal echocardiogram. Therefore, different diagnostic decisions in patients with and without infection have not influenced the results to a relevant extent. Infection involving the myocardium can induce arrhythmias such as AF. Although myocarditis was not diagnosed among our patients, we cannot exclude that mainly in patients with intermittent AF, infection may have contributed to cardiac arrhythmia. Further studies are required to investigate the link between infection and cardiac sources of cerebral embolism.
In our series two patients developed stroke from cerebral vasculitis after febrile infection; both patients had not suffered from autoimmune disorders. These observations suggest that cerebral vasculitis may in some cases be caused by infection-associated autoimmune mechanisms. In a recent study on stroke in systemic lupus erythematosus, cerebral vasculitis occurred only in association with infection invading cerebral blood vessels.14 Therefore, infection may also play a role in the pathogenesis of vasculitis associated with autoimmune disorders.
CAD is a cause of stroke that is diagnosed with increasing frequency. Its etiology, however, remains obscure in most patients. The rate of recurrence of CAD is probably low in patients without known underlying connective tissue disorders.15 16 Therefore, it is an attractive hypothesis that a temporary trigger mechanism leads to CAD in patients with or without preexisting arteriopathy. Infection may be such a trigger factor precipitating CAD by mechanical as well as inflammatory mechanisms. Cough and vomiting may place mechanical stress on cervical arteries. Several inflammatory agents that may be increased during infection can activate and damage the endothelium and may generate a first lesion leading to CAD. Those agents may include leukocyte-derived proteases such as elastase, which can degrade proteins of the extracellular matrix and can injure the endothelium.17 We recently described increased levels of elastase-inhibitor complexes after acute stroke, in subjects with vascular risk factors, and particularly in patients with infection-associated stroke.18 A deficiency of α1-antitrypsin, the main antagonist of elastase, was recently described in a patient with CAD.19 An imbalance between proteolytic enzymes and their inhibitors could result in damage to the arterial wall with consecutive dissection. Such imbalance may be enhanced by infection. Our series of patients with CAD is small, and additional data are required to confirm the role of infection in CAD. Possibly, mechanical, cytotoxic, and predisposing factors must be present for CAD to occur.
Rheological impairment caused by dehydration after fever or diarrhea may play a role in the pathogenesis of infection-associated stroke. In accordance with a previous study,2 the hematocrit was not different between groups, indicating that dehydration may not play an important role in infection-associated stroke. There was a trend to higher fibrinogen in association with infection, but there was no difference in the other coagulant parameters between both groups. In contrast to our results, Ameriso and coworkers2 found higher fibrin d-dimer levels in patients with infection compared with patients without infection, whereas their results in regard to PAI-1 were similar to ours. In a small series of patients we had detected increased levels of tumor necrosis factor–α and of its soluble receptors in infection-associated stroke.20 Tumor necrosis factor–α is a potential mediator in infection-associated stroke because it inhibits the anticoagulant (eg, downregulation of thrombomodulin) and stimulates the procoagulant features (eg, upregulation of PAI-1) of endothelial cells.21 22 Different pathways may stimulate coagulation during the pathogenesis of stroke, and the result of activated coagulation after stroke may be mainly independent of initial causative events. But there are some limitations in our work with respect to biochemical analyses. The numbers of subjects were small in the group with infection, and measurements were done only once during the first 48 hours after stroke. We therefore cannot exclude that measurements done earlier and repeatedly after stroke may still reveal particular coagulant features in infection-associated stroke that we did not find.
Levels of CRP were higher in patients with infection compared with those without infection. CRP is an acute-phase reactant that rapidly increases in inflammatory and other conditions.23 The increased level of CRP in patients without infection indicates that CRP participates in the acute-phase reaction after stroke, which is characterized by a transient rise of parameters such as interleukin-624 and fibrinogen.25 The even higher values in infection-associated stroke may be the sum of two factors causing an acute-phase reaction: the cerebral infarct and the infection. However, CRP may also play a pathogenetic role in infection-associated stroke. CRP induces human peripheral blood monocytes to synthesize tissue factor and thus contributes to a procoagulant state during inflammation.26 In a recent study we had not detected an increased coagulant potential of peripheral mononuclear leukocytes after stroke, but infection-associated stroke was not analyzed in this investigation.27
None of the studies performed thus far can prove a causal role of infection in ischemic stroke. However, several mechanisms by which infection (mainly bacterial infection) can increase the risk for thrombosis support such a hypothesis. Studies from Scandinavia showed that the risk for cerebral infarction in bacteremic patients is extremely high compared with the corresponding risk in the general population. Valtonen and coworkers28 estimate that approximately 10% of all stroke cases in Finland are associated with bacteremic infection. The role of recent infection in the pathogenesis of cerebrovascular ischemia deserves further investigations with larger numbers of patients. For preventive purposes, it may be important to elucidate whether subgroups of patients with stroke risk factors may benefit from an early treatment of bacterial infection.
Selected Abbreviations and Acronyms
|CAD||=||cervical artery dissection|
|ICA||=||internal carotid artery|
|MCA||=||middle cerebral artery|
|PAI-1||=||plasminogen activator inhibitor–1|
|SSS||=||Scandinavian Stroke Scale|
This study was funded by the financial resources of the contributing departments. We gratefully acknowledge the skillful technical assistance of Anette Immel, Marburg, FRG.
- Received May 2, 1995.
- Revision received June 6, 1995.
- Accepted June 8, 1995.
- Copyright © 1995 by American Heart Association
Syrjänen J, Valtonen VV, Iivanainen M, Kaste M, Huttunen JK. Preceding infection as an important risk factor for ischaemic brain infarction in young and middle aged patients. Br Med J. 1988;296:1156-1160.
Ameriso SF, Wong VLY, Quismorio FP Jr, Fisher M. Immunohematologic characteristics of infection-associated cerebral infarction. Stroke. 1991;22:1004-1009.
Grau AJ, Buggle F, Heindl S, Steichen-Wiehn C, Banerjee T, Maiwald M, Rohlfs M, Suhr H, Fiehn W, Becher H, Hacke W. Recent infection as a risk factor for cerebrovascular ischemia. Stroke. 1995;26:373-379.
Adams HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE, and the TOAST Investigators. Classification of subtype of acute stroke: definitions for use in a multicenter clinical trial. Stroke. 1993;24:35-41.
Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke: background and study protocol. Stroke. 1985;16:885-890.
O’Neill PA, Davies I, Fullerton KJ, Bennett D. Stress hormone and blood glucose response following acute stroke in the elderly. Stroke. 1991;22:842-847.
Suzuki J, Kodama N. Moyamoya disease: a review. Stroke. 1983;14:104-109.
Wolf PA, Dawber TR, Emerson T Jr, Kannel WB. Epidemiologic assessment of chronic atrial fibrillation and risk of stroke: the Framingham Study. Neurology. 1978;28:973-977.
Gustafsson C, Blombäck M, Britton M, Hamsten A, Svensson J. Coagulation factors and the increased risk of stroke in nonvalvular atrial fibrillation. Stroke. 1990;21:47-51.
Futrell N, Millikan C. Frequency, etiology, and prevention of stroke in patients with systemic lupus erythematosus. Stroke. 1989;20:583-591.
Leys D, Moulin Th, Stojkovic T, Begey S, Chavot D. Follow-up of patients with history of cervical artery dissection. Cerebrovasc Dis. 1995;5:43-49.
Smedly LA, Tonnesen MG, Sandhaus RA, Haslett C, Guthrie LA, Johnston RB Jr, Henson PM, Worthen GS. Neutrophil-mediated injury to endothelial cells: enhancement by endotoxin and essential role of neutrophil elastase. J Clin Invest. 1986;77:1233-1243.
Grau AJ, Heilig B, Heindl S, Buggle F, Hacke W. Tumor necrosis factor-alpha and its soluble 55-kD and 75-kD receptors in acute ischemic stroke. Neurology. 1993;43S:A228. Abstract.
Schleef RR, Bevilacqua MP, Sawdey M, Gimbrone MA, Loskutoff DJ. Cytokine activation of vascular endothelium: effects on tissue-type plasminogen activator and type 1 plasminogen activator inhibitor. J Biol Chem. 1988;263:5797-5803.
Fassbender K, Rossol S, Kammer TH, Daffertshofer M, Wirth ST, Dollmann M, Hennerici M. Proinflammatory cytokines in serum of patients with acute cerebral ischemia: kinetics of secretion and relation to the extent of brain damage and outcome of disease. J Neurol Sci. 1994;122:135-139.
Elliot FA, Buckell M. Fibrinogen changes in relation to cerebrovascular accidents. Neurology. 1961;11:120-124.
Cermak J, Key NS, Bach RR, Balla J, Jacob HS, Vercellotti GM. C-reactive protein induces human peripheral blood monocytes to synthesize tissue factor. Blood. 1993;82:513-520.
Valtonen V, Kuikka A, Syrjänen J. Thrombo-embolic complications in bacteraemic infections. Eur Heart J. 1993;14(suppl K):20-23.