Histopathological Analysis of the Mechanisms of Intracranial Hemorrhage Complicating Infective Endocarditis

Transcription

1 843 Histopathological Analysis of the Mechanisms of Intracranial Hemorrhage Complicating Infective Endocarditis Junichi Masuda, MD; Chikao Yutani, MD; Riichiro Waki, MD; Jun Ogata, MD; Yoshihiro Kuriyama, MD; and Takenori Yamaguchi, MD Background and Purpose: We conducted the present study to elucidate the pathological mechanisms leading to intracranial hemorrhage complicating infective endocarditis. Methods: Neurological, neuroradiological, and histopathological analyses were performed in 16 patients (one surgical and 15 autopsy cases), 12 men and four women 2668 years of age, who had demonstrated central nervous system complications during the course of infective endocarditis. Results: Intracranial hemorrhage was found in all cases; parenchymal hematomas were found in 12 cases, hemorrhagic infarcts in four cases, and primary subarachnoid hemorrhages in two cases. Chronological analysis of neurological examination and computed tomographic scan of the brain confirmed that antecedent cerebral ischemic events had occurred in five of 12 patients showing parenchymal hematomas at autopsy. Hemorrhagic infarct, indicated by petechial or diffuse hemorrhages within the infarct, was seen in another four patients, so that hemorrhagic transformation of the ischemic infarct was confirmed in nine patients. Although mycotic aneurysms were found in five patients, only three of these were ruptured; the other two were occluded with septic emboli. Pyogenic arteritis without aneurysm was found to be distributed in the small cortical arterial branches located in the spaces of cortical sulci, with rupture occurring in five patients. Conclusions: These results suggest that hemorrhagic transformation of the ischemic infarct due to septic emboli is the most frequent mechanism leading to intracerebral hemorrhage encountered in patients dying of infective endocarditis and that rupture of pyogenic arteritis may be responsible for such hemorrhage in many cases, with ruptures of mycotic aneurysms as an alternative mechanism. (Stroke 1992^3:843850) KEY WORDS cerebral hemorrhage embolism endocarditis, bacterial Intracranial hemorrhage (ICH) occurs in about 5% of patients with infective endocarditis and is an important complication because it often leads to death. 1 8 Previous studies have observed cerebral mycotic aneurysms in 210% of patients with infective endocarditis, and these aneurysms are usually considered to be responsible for such hemorrhagic events. 4 " 13 There are, however, many cases in which mycotic aneurysms are not detectable on cerebral angiograms or at autopsy. Yock reported a case of septic saddle embolus causing basilar artery rupture without mycotic aneurysm. 14 Furthermore, Hart et al' 3 reported that mycotic aneurysm was present in only two of 17 patients with ICH complicating infective endocarditis; these authors suggested the following different mechanisms leading to ICH in patients with infective endocarditis: 1) secondary hemorrhagic transformation of ischemic infarcts From the Research Institute (J.M., R.W., J.O.), Department of Pathology (C.Y.), and Cerebrovascular Division (Y.K., T.Y.), National Cardiovascular Center, Osaka, Japan. Supported by the Research Grant for Cardiovascular Diseases (2A6) from the Ministry of Health and Welfare of Japan. Address for reprints: Junichi Masuda, MD, National Cardiovascular Center, Research Institute, 571, Fujishirodai, Suita, Osaka 565, Japan. Received October 10, 1991; accepted February 20, caused by sterile emboli in anticoagulated patients, and 2) rupture of pyogenic arteritis following septic embolization during uncontrolled infection. However, the direct relation between vascular changes and hemorrhagic events was not clearly demonstrated in their study because pathological examinations were performed in only six patients. To clarify the mechanisms leading to ICH in patients with infective endocarditis, we reviewed the autopsy and surgery materials from the National Cardiovascular Center, Osaka, Japan, from 1978 to We examined 16 patients who experienced central nervous system complications during the course of infective endocarditis, and we analyzed the pathological changes of the brain and cerebral vessels. Subjects and Methods Fortytwo patients with infective endocarditis were autopsied at the National Cardiovascular Center from January 1978 to December Central nervous system complications during the course of infective endocarditis were confirmed in 15 patients. In addition, one surgical specimen was obtained from the evacuated brain tissue of a patient with infective endocarditis, and a pathological examination was performed on it. Thus,

2 844 Stroke Vol 23, No 6 June 1992 we analyzed 16 patients, 12 men and four women, 2668 (mean±sd, 4913) years of age (Table 1). Careful pathological examination of each case was carried out with special attention to the brain, heart, and arterial tree supplying the brain. The clinical records of the patients, including brain computed tomographic (CT) scan and cerebral angiography, were also collected and analyzed. The parenchymal hematoma was defined as a massive hemorrhage in the brain parenchyma with surrounding compression of tissues and flattening of adjacent cortical sulci and ganglial components. Hemorrhagic infarct was defined as petechial or diffuse hemorrhages (coalescent petechiae) within an area of the infarct without evidence of surrounding compression by the hemorrhage. The primary subarachnoid hemorrhage was defined as a hemorrhage in the subarachnoid space that could not be explained as secondary to parenchymal hemorrhage. Mycotic aneurysm was distinguished from pyogenic arteritis by aneurysmal dilatation of the inflamed arterial wall. Results Table 1 summarizes the clinical profiles of the 16 patients with infective endocarditis and neurological complications. Various congenital or acquired forms of valvular disease were found as a predisposing cardiovascular or systemic disease in 12 patients. Rheumatic valvular disease was the most common; this was noted in six patients. Of the other six patients, one had a congenital bicuspid aortic valve, two had mitral valve prolapse, and three had had previous episodes of infective endocarditis. Valve replacement operations had been performed on the diseased mitral or aortic valves in eight of these patients before the onset of the infective endocarditis. Of the remaining four patients, one had peripheral cavernous hemangioma with the KlippelWeber syndrome, two had been treated with large doses of steroid hormones for systemic lupus erythematosus and rapidly progressive glomerulonephritis, respectively, and one had no apparent predisposing diseases. Organisms causing infective endocarditis were identified in 14 cases by blood culture, and grampositive cocci were detected in the infected valves of the remaining two patients at autopsy (Table 1). Staphylococcus aureus, detected in six patients, was the most common organism. The interval from the onset of infective endocarditis to the onset of neurological complications varied from simultaneous onset to 3 months. Neurological complications occurred during the active stage of the infective endocarditis in all patients, and at autopsy the infected valves demonstrated active purulent inflammation in all patients. Eight patients had been anticoagulated, and two had been on hemodialysis until the hemorrhagic events were recognized. Another patient showed prolongation of prothrombin time despite having received no anticoagulation therapy. Two patients experienced ICH during or immediately after open heart surgery. Except for one patient who died after 44 days, all patients died within 16 days after the onset of neurological complications. Deterioration of neurological symptoms was observed after the occurrence of intracranial hemorrhagic events in many patients. Table 2 summarizes the neuroradiological and neuropathological findings in the 16 patients. ICH was found in all autopsy cases, and hemorrhagic infarct was found on brain CT examination in the surgery case (case 14). Parenchymal hematomas were found in 12 patients (cases 112), hemorrhagic infarcts in four (cases 1215), and primary subarachnoid hemorrhages in two (cases 1516). Figure 1 shows the macroscopic appearance of representative cases with parenchymal hematomas. The hematomas, often multiple, were usually located in cortical or subcortical regions and showed continuity with the subarachnoid space. Pathological study of the cerebral arterial trees supplying the regions with hematomas, including the circle of Willis and its major branches, did not demonstrate ruptured aneurysms. Occlusion of cerebral arteries by septic emboli was found in two patients. Pyogenic arteritis with fragments of septic emboli was found in eight patients. In these cases, the arterial walls were infiltrated with purulent inflammatory cells and showed rupture without aneurysmal dilatation as shown in Figure 2. These ruptures were often present in the small cortical arteries localized in the spaces of cortical sulci, and there was localized purulent meningitis around the ruptured arteries. Chronological analysis of neurological symptoms and brain CT findings of the 12 patients revealed that the massive hematomas observed at autopsy were the results of hemorrhagic transformation of the ischemic infarct in at least five patients. Four patients had demonstrated lowdensity areas without any areas of high density on brain CT examination shortly after the onset of neurological symptoms or signs, and one patient had a normal CT appearance. The regions that previously demonstrated low density on CT scan were transformed into massive parenchymal highdensity areas on followup CT scan, when the neurological symptoms had deteriorated (Figure 3). Anticoagulation therapy that was performed in patients with replaced valves or other predisposing disorders may have enhanced thrombolysis and reopening of the vessels and induced parenchymal hematomas rather than petechial or diffuse hemorrhages within the infarct. In fact, eight patients had been anticoagulated for the prophylaxis of the thromboembolism secondary to their predisposing cardiovascular disorders, and one additional patient showed impairment of coagulation. Hemorrhagic infarct, defined as petechial or diffuse hemorrhages (coalescent petechiae) within the infarct, was found in four patients. In case 12, the patient had a small parenchymal hematoma in the left frontal lobe associated with intraventricular hemorrhage and multiple hemorrhagic infarcts distributed in both occipital lobes; these were possibly caused by uncinate compression of the posterior cerebral arteries. A mycotic aneurysm was found in one of the branches of the left posterior cerebral artery, and its lumen was occluded with a septic embolus. The arterial wall was infiltrated by a purulent infiltrate and showed aneurysmal dilatation but was not ruptured. Other arterial branches supplying infarcted areas and hematoma were not occluded and showed no pathological changes. A hemorrhagic infarct in the right parietal lobe supplied by the

3 Anticoagulation t t HD HD MVP, mitral valve HD, hemodialysis; 1 3 n n o I 8 I M Q. TABLE 1. Clinical Profiles of Patients With Infective Endocarditis and Neurological Complications Case Age/sex 32/F 29/M 46/M 61/M 54/M 30/M 26/M 56/M 52/M 68/F 57/M 53/F 67/F 57/M 46/M 48/M Predisposing cardiovascular or systemic diseases KlippelWeber syndrome Healed IEAVR and MVR RHD (MV) MVPMVR ASOFP bypassieavr SLEsteroid pulse therapy RHD (AV) RHD (MV)MVR RHD (MV) RHD (MV)MVR Healed IEAVR RHD (MV) MVR Unknown MVP Bicuspid AVAVR RPGNsteroid therapy Infective organisms Grampositive coccus* Candida parapsilosis Grampositive coccus* Staphylococcus aureus Streptococcus group C Neisseria Streptococcus alpha Corynebacterium Staphylococcus epidermis Streptococcus sanguis S. epidermis Time from IE onset to CNS symptoms 30 days Simultaneous 2 months 1 day 38 days 9 days 3 months 4 days 1 day lday 5 days 11 days 40 days 1 day 10 days 20 days Neurological symptoms and signs R hemiparesis, anisocoria, coma R hemiparesis, coma R hemiparesis, coma Dysarthria, R hemiparesis TIA (oculomotor palsy, syncope) L hemiplegia, coma Headache, L hemiplegia Headache, L hemiparesis Headache, R hemiplegia Anisocoria, quadriplegia Anisocoria, L hemiparesis L paresthesia, convulsion R Jacksonian seizure Stupor, L hemiparesis Loss of consciousness Loss of consciousness Time from CNS symptoms to death 5 days 9 days 44 days 11 days 4 days 4 days 9 hours 9 days 2 days 14 hours 3 hours 16 days 15 days Alive 3 days 5 days 'Organisms were detected in infected cardiac valves at autopsy although blood culture was negative. tpatients underwent open heart surgery immediately before hemorrhagic events. ^Prolongation of prothrombin time despite no anticoagulation therapy. IE, infective endocarditis; CNS, central nervous system;, yes;, no; AV, aortic valve; AVR, aortic valve replacement; MV, mitral valve; MVR, mitral valve replacement; prolapse; RHD, rheumatic heart disease; ASO, arteriosclerosis obliterans; FP bypass, femoropopliteal bypass; TIA, transient ischemic attack; SLE, systemic lupus erythematosus; RPGN, rapidly progressive glomerulonephritis.

4 . * * ACA, TABLE 2. Neuroradiological and Neuropathological Characteristics of Brain and Cerebral Arteries Case Brain CT findings at onset Pathology at autopsy Cerebral angiography findings 1 LDA (L parietal) 2 LDA (Bil temporoparietal) 3 LDA (L temporal) 4 LDA (L semiovale) 5 Normal 6 HD mass (R occipital, L frontal) 7 HD mass (R frontal) 8 HD mass (Bil frontal, Bil occip) 9 HD mass (L occipital) 10 HD mass (L parietal) 11 HD mass (R occipital), LDA (cerebellum) 12 HD in LDA (L frontal, Bil occipital) 13 HD in LDA (R parietal) 14 HD in LDA (R temporal) SAH Hematoma (L parietal) Hematoma (R temporal, L parietal) Hematoma (L temporal) Hematoma (L occipital, R parietal) Hematoma (L frontal) Hematoma (R occipital, L frontal) Hematoma (R frontal) Hematoma (Bil frontal, R temp, Bil occip) Hematoma (L occipital) Hematoma (L parietal) Hematoma (R occipital, R frontal) Hematoma (L frontal), HI (Bil occipital) HI (R parietal) SAH, HI (R occipital) SAH AN (), Oc (Bil MCA, BA) AN(), Oc() AN (), Oc () AN (), Oc () AN(LACA), Oc() AN, Oc (R MCA branch) AN (SCA), Oc () Mycotic aneurysm 'Rupture of arteries shown on histopathological analysis. CT, computed tomographic; LDA, lowdensity area; HD, high density;, present;, absent; Bil, bilateral; AN, aneurysm; Oc, occlusion; MCA, middle cerebral artery; BA, basilar artery; anterior cerebral artery; SCA, superior cerebellar artery; HI, hemorrhagic infarct; SAH, subarachnoid hemorrhage. _ * * Septic emboli Pyogenic arteritis

5 Masuda et al Intracranial Hemorrhage in Infective Endocarditis 847 FIGURE 1. Axial sections of brains with parenchymatous hematomas (cases 36 and 811). Right hemispheres are shown at the right in cases 39; this is reversed in cases 10 and 11. middle cerebral artery was demonstrated in case 13. Histological study of the cerebral arteries supplying the infarcted area did not show pathological evidence of mycotic aneurysm. However, there was a rupture of the small cortical arteries located in the spaces of cortical sulci in the infarcted area, although purulent inflammation destructive to the artery wall was not conspicuous. In case 14, the arterial specimen removed from the patient demonstrated rupture of a mycotic aneurysm in a branch of the right middle cerebral artery (Figure 4B). Although a pathological study of the brain was not performed in this surgical case, the brain CT scan revealed diffuse hemorrhage within the ischemic infarct, and cerebral angiography demonstrated an aneurysm and occlusion of the branches of the right middle cerebral artery (Figure 4A). Primary subarachnoid hemorrhage was found in two patients (Cases 15 and 16), and ruptured mycotic aneu FiGURE 2. Case 4. Photomicrograph of ruptured arterial wall showing inflammation in wall of one of the smalt arteries located in subarachnoid space, with extensive inflammation and disruption of internal elastic lamina. Elastica van Gieson stain, x66.

6 848 Stroke Vol 23, No 6 June 1992 FIGURE 3. Case 1. Patient experienced hemorrhagic transformation of ischemic infarct to parenchymal hematoma. (Right hemispheres are shown at right.) Panel A: Brain computed tomographic (CT) scan 4 days after onset of right hemiparesis. Panel B: Followup brain CTscan 12 hours after open heart surgery. Panel C: Axial section of brain at autopsy.

7 Masuda et al Intracranial Hemorrhage in Infective Endocarditis 849 FIGURE 4. Case 14. Panel A: Right internal carotid arteriogram demonstrating aneurysm and occlusions of some middle cerebral artery branches. Panel B: Photomicrograph of ruptured mycotic aneurysm in same patient. Elastica van Gieson, X.50. rysms were found in both patients. Cerebral angiography showed that the aneurysm in case 16 was at the origin of the left superior cerebellar artery. In the other patient, the aneurysm was located in one of the branches of the right posterior cerebral artery. Histopathologic section of the aneurysms showed that septic emboli containing calcified material had occluded the lumen and that purulent inflammation extended to the arterial wall, possibly leading to its rupture. Case 15 also demonstrated rupture of extensive pyogenic arteritis of the small arteries located in the cortical sulci of the right occipital lobe. This rupture was considered to be responsible for the subarachnoid hemorrhage and hemorrhagic infarct seen in the right occipital lobe. Discussion Despite remarkable advances in treatment for patients with infective endocarditis, the incidence and morbidity rate of neurological complications have remained high, and these complications still cause controversial problems in areas such as indications for open heart surgery 15 and anticoagulation Of the various forms of neurological complications, it is the mecha

8 850 Stroke Vol 23, No 6 June 1992 nisms of ICH that must be clarified most urgently, since ICH often leads to a poor outcome. Therapeutic anticoagulation increases the risk of intracranial hemorrhage, and may induce massive parenchymal hematoma in cardioembolic stroke. 16 " 19 Cardiac surgery after acute embolic stroke increases the risk of secondary bleeding into the ischemic region; the mechanisms underlying this bleeding are discussed elsewhere. 15 Of the 16 patients examined in this report, four had hemorrhagic infarcts and an additional five patients experienced antecedent ischemic episodes that led subsequently to parenchymal hematomas. Thus, the hemorrhagic events are not the primary event in many patients with ICH complicating infective endocarditis; embolic episodes with secondary hemorrhagic transformation of the infarcts are the most frequent mechanism leading to ICH among those dying with infective endocarditis. The vascular lesions responsible for ICH complicating infective endocarditis are not fully understood. Cerebral mycotic aneurysms may be responsible in most instances Cerebral mycotic aneurysms have been noted in 210% of patients with infective endocarditis; these mycotic aneurysms can occur in any vessel, including the circle of Willis, and tend to involve the more distal portions of the middle cerebral artery, near the surface of the brain; aneurysms involve the secondary and tertiary branches, especially those located in the region of the Sylvian fissure. 5 ' Such aneurysms are rarely located in intraparenchymal small arteries. Most patients with ICH described previously and in our study have cortical or subcortical hematomas; this location of the hematomas is not consistent with the topography of mycotic aneurysms described by others. We observed mycotic aneurysms at autopsy in only five patients. Three were ruptured and were associated with subarachnoid hemorrhage in two patients and hemorrhagic infarction in one patient. Rupture of mycotic aneurysms was not confirmed in any patients with parenchymal hematomas, although the hemorrhagic sites were carefully studied in stepwise histological sections. Furthermore, cerebral angiography was performed on seven patients, but no aneurysms were demonstrated in four patients despite the existence of ICH. Thus, it is unreasonable to consider rupture of mycotic aneurysms as a frequent cause of parenchymal hemorrhagic events. Pyogenic arteritis was found in 10 patients; in five of these, rupture of the arterial wall was demonstrated without any aneurysmal dilatation. Ruptured pyogenic arteritis was found in small cortical arteries (ranging from 200 /xm to 1 mm in diameter) located deep in the spaces of cortical sulci, and unruptured pyogenic arteritis were also seen in intraparenchymal small arteries in some cases. Such a distribution pattern seems to be consistent with the topography of intracerebral hemorrhage reported in the present study and previously in the literature. Pyogenic arteritis was usually associated with purulent meningitis in the surrounding subarachnoid space of the cortical sulci and with intravascular fragments of embolic materials. It was not clear whether these fragments were the result of primary septic embolization or of fragmentation of septic emboli that lodged in the proximal arteries. Nevertheless, if the rupture of pyogenic arteritis plays an important role in the pathogenesis of ICH, the fact that neurological complications occur during the active stage of infective endocarditis becomes easier to understand. Our present study demonstrated histopathological evidence of the rupture of pyogenic arteritis without aneurysmal dilatation and extended the observations of Hart et al. 1 3 Secondary hemorrhagic transformation of an ischemic infarct, possibly caused by septic embolization, was the most frequent mechanism of ICH in this series. The rupture of inflamed arteries during uncontrolled infection may be responsible for hemorrhagic transformations encountered in patients with infective endocarditis. Anticoagulation therapy and open heart surgery may enhance such hemorrhagic transformations. Thus, antimicrobial treatment should be performed extensively during the acute phase of infective endocarditis; the resumption of anticoagulation therapy and indications for open heart surgery should be carefully evaluated. Acknowledgments The authors wish to thank Dr. Masami Imakita and Dr. Hatsue Ueda for permission to study their autopsy cases. References 1. Hart RG, KaganHallet K, Joerns SE: Mechanisms of intracranial hemorrhage in infective endocarditis. Stroke 1987;18: Hart RG, Foster JW, Luther MF, Kanter MC: Stroke and infective endocarditis. Stroke 1990;21: Kanter MC, Hart RG: Neurologic complications of infective endocarditis. Neurology 1991;41: Jones HR Jr, Siekert RG, Geraci JE: Neurologic manifestations of bacterial endocarditis. Ann Intern Med 1969;71: Jones HR Jr, Siekert RG: Neurological manifestations of infective endocarditis: Review of clinical and therapeutical challenges. Brain 1989;112: Pruitt AA, Rubin RH, Karchmer AW, Duncan GW: Neurologic complications of bacterial endocarditis. Medicine 1978;57: Salgado AV, Furlan AJ, Keys TF: Mycotic aneurysm, subarachnoid hemorrhage, and indications for cerebral angiography in infective endocarditis. Stroke 1987;18: Salgado AV, Furlan AJ, Keys TF, Nichols TR, Beck GJ: Neurologic complications of endocarditis: A 12year experience. Neurology 1989;39: Frazee JG, Cahan LD, Winter J: Bacterial intracranial aneurysms. JNeurosurg 1980;53: Bullock R, Van Dellen JR, Van Den Heever CM: Intracranial mycotic aneurysms: A review of 9 cases. S Afr Med J 1981;60: Bullock R, Van Dellen JR: Rupture of bacterial intracranial aneurysms following replacement of cardiac valve. Surg Neurol 1982;17: Bamford J, Hodges J, Warlow C: Late rupture of a mycotic aneurysm after "cure" of bacterial endocarditis. J Neurol 1986;233: Brust JCM, Taylor Dickinson PC, Hughes JEM, Holtzman RNN: The diagnosis and treatment of cerebral mycotic aneurysms. Ann Neurol 1990;27: Yock DH Jr: Septic saddle embolus causing basilar artery rupture without mycotic aneurysm. Am J Neuroradiol 1984;4: Maruyama M, Kuriyama Y, Sawada T, Yamaguchi T, Fujita T, Omae T: Brain damage after open heart surgery in patients with acute cardioembolic stroke. Stroke 1989;20: Lieberman A, Hass WK, Pinto R, Isom WO, Kupersmith M, Bear G, Chase R: Intracranial hemorrhage and infarction in anticoagulated patients with prosthetic heart valves. Stroke 1978;9: Cerebral embolism study group: Cardioembolic stroke, early anticoagulation, and brain hemorrhage. Arch Intern Med 1987;147: Okada Y, Yamaguchi T, Minematsu K, Miyashita T, Sawada T, Sadoshima S, Fujishima M, Omae T: Hemorrhagic transformation in cerebral embolism. Stroke 1989;20: Paschalis C, Pugsley W, Harrison MJG: Rate of cerebral embolic events in relation to antibiotic and anticoagulant therapy in patients with bacterial endocarditis. Eur Neurol 1990;30:8789