Cerebral Hemorrhage With Biopsy-Proved Amyloid Angiopathy William H. Yong, MD; Marie E. Robert, MD; Diana Lenard Secor, MS; Theodore J. Kleikamp, MD; Harry V. Vinters, MD \s=b\ Clinical, radiological, and immunohistochemical findings in brain biopsy specimens from six patients with cerebral amyloid angiopathy\p=m-\associated intracerebral hemorrhage were reviewed. Acute clinical presentations included headache, nausea and vomiting, loss of consciousness, and focal neurological deficits such as hemiplegia and blindness. Transient schemic attacks experienced by one patient and referable to one hemisphere did not indicate impending hemorrhage in that region. Computed tomographic scans revealed acute, irregular, superficial, lobar hemorrhage with occasional ring enhancement. Immunohistochemical studies were performed on biopsy specimens using primary antibodies against portions of the Alzheimer A4 (\g=b\-)peptide or \g=g\ x=req-\ trace peptide (the vascular amyloid protein in f~^ erebral amyloid angiopathy (CAA) is characterized by the accumula¬ tion of amyloid in the media and adventitia of medium and small arteries, arterioles, and occasionally veins.1,2 It is responsible for up to 10% of primary, nontraumatic brain hemorrhages or 1% of all strokes.3 Cerebral amyloid angio¬ pathy commonly occurs in otherwise healthy and nondemented persons with¬ out cerebral hemorrhage or infarcts.2 The amyloid in most patients with CAA Accepted for publication August 1, 1991. From the Department of Pathology (Drs Yong, Robert, and Vinters and Ms Secor) and Brain Research Institute (Dr Vinters), UCLA School of Medicine; and Department of Neurology, St Vincent Hospital and Medical Center, Portland, Ore (Dr Kleikamp). Reprint requests to Department of Pathology (Neuropathology), CHS 18-170, UCLA School of Medicine, Los Angeles, CA 90024-1732 (Dr Vinters). patients with hereditary cerebral hemorrhage with amyloidosis\p=m-\Icelandictype). In all patients, anti-A4 and anti\p=m-\\g=g\-trace labeled cerebral microvessels. Immunoreactive senile plaques were few compared with the numbers of stained microvessels. Reactive astro- cytes in some patients were labeled by both antiserum samples, suggesting uptake or production of these proteins by the astrocytes. This study demonstrates the heterogeneous clinical and radiological features of cerebral amyloid angiopathy\p=m-\relatedbrain hemorrhage and the value of anti-A4 and anti\p=m-\\g=g\-trace immunohistochemical study of biopsy material from patients with suspected cerebral amyloid angiopathy\p=m-\relatedintraparenchymal bleeding. {Arch Neurol. 1992;49:51-58) is virtually identical biochemically to that found in patients with Alzheimer's disease (AD) senile or neuritic plaques, both lesions containing the Alzheimer A4 or ß-peptide. '~6 Cerebral amyloid an¬ giopathy occurs in over 90% of brains from patients with AD or senile demen¬ tia of the Alzheimer type (SDAT).6 Thus, understanding the pathophysiological nature of CAA may provide in¬ sights into AD and SD AT7 Cerebral amyloid angiopathy appears to increase in extent and severity with age.2 There are associations between CAA and syn¬ dromes of familial cerebral hemor¬ rhage,8,9 Down's syndrome,1" isolated cerebral vasculitis,1112 leukoencephalopathy,13" and cerebrovascular malfor¬ mations.1516 Traditionally, CAA has been diagnosed by staining tissue sec¬ tions with Congo red and viewing them under polarized light. Recently, im- Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 munohistochemical staining with pri¬ mary monoclonal or polyclonal antibod¬ ies to the Alzheimer A4 peptide and to a peptide related to y-trace or cystatin C (a protease inhibitor that composes the microvascular amyloid in hereditary Icelandic cerebral hemorrhage [HCHWA-I]) has been shown to localize amyloid in microvessel walls in brain tissue from appropriate patients.1718 In this study, we examine the efficacy of immunohistochemical methods in de¬ tecting CAA in brain biopsy specimens from six patients with peripheral lobar cerebral hemorrhages. Clinical and ra¬ diological features of the patients are emphasized. SUBJECTS AND METHODS The charts of six patients who ranged in age from 54 to 86 years were reviewed with respect to clinical presentation, age, sex, his¬ tory of hypertension and smoking or other risk factors for vascular disease, dementia, neurological and psychiatric symptoms, im¬ aging studies, surgical procedures, treat¬ ment, and outcome. Two of the patients (cases 1 and 4) were seen and treated at St Vincent Hospital and Medical Center, Port¬ land, Ore, and four (cases 2, 3, 5, and 6) were followed up at the UCLA Medical Center. Initial diagnosis of CAA in biopsy material (available from all patients) was by Congo red staining of paraffin sections and examina¬ tion with polarization microscopy. Immuno¬ histochemical (immunoperoxidase) staining of tissues was carried out utilizing primary polyclonal antibodies raised against synthet¬ ic peptides representing a 28-amino-acid fragment from the Alzheimer A4 (ß-) pep¬ tide19,20 or an 18-amino-acid fragment from a y-trace-like protein initially isolated from ce¬ rebral microvessels of patients with HCHWA-I.17,21,22 The immunocytochemical method involved formic acid pretreatment of brain tissue sections to improve immunola- ' beling" and a standard avidin-biotin-peroxidase technique for visualization of immunoreactivity. Control sections from all patients were treated with serum samples preabsorbed with the immunogen to which the antibody was raised as a measure of antibody labeling specificity. Sections of AD brain with widespread parenchymal and microvascular A4 amyloid were immunolabeled with anti-A4, while paraffin-embedded brain tis¬ sue from patients with HCHWA-I was stained with anti-y-trace antibody to serve positive controls and ensure consistent levels of staining in each run. Table 1. -Clinical Features of Spontaneous Cerebral Hemorrhage in Six Patients With CAA* Location of Patient No./Age, HTN/ Dementia y/Sex Smoker at Onset 1/54/M ?/ + No Clinical Presentation Loss of con¬ sciousness (h/oTIAs, R hemiparesis 1 y?) Subsequent Symptoms Seizures, dementia, depression, blindness, obtundation as 2/58/F Confusion, diffi¬ Memory loss, culty walking, irritability, urinary incon¬ depression, tinence, apha¬ paranoia sia, HA, N/V (h/o depres¬ sion, 38 y) L hemiparesis, Seizures, apha¬ HA, N/V, pro¬ sia, L hemi¬ gressive leth¬ paresis, L argy (h/o paresthesias REPORT OF CASES Table 1 summarizes the clinical features of all patients, and Table 2 summarizes the histologieal features of the cerebral biopsy spec¬ imens; relevant details of both are presented below. CASE 1.—A 54-year-old white man with a questionable history of mild hypertension and a two- to three-pack per day smoking history presented with loss of consciousness in 1984. In the previous year, he had com¬ plaints of numbness in the right hand and occasional episodes of dropping objects from that hand. At that time, a conrouted tomo¬ graphic (CT) scan of the head was negative, but electroencephalography (EEG) showed excessive synchronous 2- to 3-Hz slow-wave activity. He was started on a regimen of phenytoin sodium, 400 mg/d. He remained asymptomatic for 4 months and then had an episode of aphasia followed by a tonoclonic seizure. Magnetic resonance imaging (MRI) immediately before this episode demonstrat¬ ed multiple regions in the periventricular and centrum semiovale regions bilaterally, con¬ sistent with plaques of multiple sclerosis. Neuropsychological testing was consistent with organic dementia and depression. In 1987, his symptoms worsened, and MRI re¬ vealed additional small foci of probable demyelination involving the brain stem. The j)atient then suffered acute confusion and blindness. Computed tomography revealed a large hemorrhage in the right occipital and posterior parietal regions with mass effect. Cerebral arteriography failed to demon¬ strate an arteriovenous malformation (AVM) but showed mild atheromatous change of the cerebral vessels. He showed mild clinical improvement and in mid-1987 underwent a right temporal lobe biopsy that revealed CAA. Postoperatively, the patient was obtunded, and a CT scan demonstrated edema but no new hemorrhage. Dexamethasone sodium phosphate was administered, with clinical improvement. In late 1987, a CT scan re¬ vealed an additional left-sided occijjital lucency. Magnetic resonance imaging in April 1988 revealed multiple serpiginous channels with high signal intensity in the temporal region, suggestive of an AVM. In retrospect, and since an angiogram 1 year earlier had not shown AVM, this area may have represented acute hemorrhage in a gyral pattern. His condition improved slightly until October 1988, when his vision and ability to carry out activities of daily living deteriorated. Com¬ puted tomographic and MRI scans in Novem¬ ber 1988 demonstrated a right-sided parieto- 3/60/F Hemorrhages (Onset/ Surgical Procedures Subsequent) L, R periventric- R temporal ular/R occipibiopsy toparietal R tem¬ (X2), poral, L occip¬ ital, R hemi¬ sphere L frontal biopsy L, R frontal/ none R frontal/ R frontal evacu¬ ation of he¬ none matoma TIAs with R hemiparesis 7 y, L hemi¬ Yes paresis 6 mo) R frontal, L HA, confusion, Confusion, L ataxia, L hohemiparesis, L cerebellar/L monymous paresthesia, occipital, R cerebellar hemianopsia, memory loss, L hemiplegia irritability (h/o carotid and peripheral R frontal biopsy vascular dis¬ No 5/82/F ease) HA, slurred No change speech, R hemiparesis, L occipital/ L occipital evac¬ uation of he¬ none matoma obtundation 6/86/M No L hemiparesis, R parieto-occipi- R parieto-occipiMalaise, N/V, L homonyataxia, obtun¬ totemporal/ totemporal mous hemian¬ none evacuation of dation, L hematoma hemiparesis opsia *CAA indicates cerebral amyoid angiopathy; HTN, hypertension; TIA, transient ischemie attack; h/o, history of; HA, headache; and N/V, nausea and vomiting. Ages are age at onset of first hemorrhage. The periventricular lucencies in patient 1 probably represented infarcted and demyelinated areas. Table 2.—Immunohistochemical and Microscopic Features of Hemorrhagic Brain Tissue Associated With CAA* Staining of Staining of Cerebral Microvessels Case No. Congo Red Senile Plaques Anti-7Anti-A4 Trace Staining of Astrocytes Anti-7- Anti-7- Anti-A4 Trace Anti-A4 Trace CNA CNA CNA CNA CNA CNA *CAA indicates cerebral amyoid angiopathy; CNA, cannot assess. With the minimal amount of brain tissue sometimes available, ¡mmunolabeled sections did not always contain reactive astrocytes or senile plaques, and It was therefore not possible to comment on their immunolabeling properties. occipital hemorrhage (Fig 1, top and center). Magnetic resonance imaging in April 1989 showed maturation of the parieto-oceijrital hemorrhage. Following 1 month of increas¬ ing dementia in May 1989, the patient be¬ came acutely unresponsive, and a CT scan demonstrated an extensive, multilobar right Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 hemispheric hemorrhage with substantial subarachnoid spread (Fig 1, bottom). An EEG obtained during obtundation revealed marked abnormalities in the right hemi¬ sphere consisting of focal slowing over the right frontotemporal regions superimposed on generalized bilateral slowing. He suffered Fig 2. —Patient 2. Left, Enhanced computed tomographic scan of a low-density, ring-enhancing, superficial left frontal lobe lesion. Right, Computed tomographic scan with contrast of a simulta¬ neous hemorrhage in the right frontal lobe. An area of ring enhancement is visible with a central region of mixed density. recurrent grand mal seizures refractory to phenytoin and died 2 months later. Autopsy (restricted to the brain) revealed mild atherosclerosis except for a large basilar artery plaque. The frontal lobes bilaterally Fig 1. —Patient 1. Top, This T2-weighted mag¬ netic resonance imaging (MRI) image (repeti¬ tion/echo times, 2800/80 ms) demonstrates a serpiginous lobar lesion of high signal intensity, representing hemorrhage in the right occipital region (November 9, 1988). Center, A sagittal T'-weighted MRI image (600/20 ms) of the right occipital lesion (same date) demonstrates an area of high signal Intensity (indicating a chron¬ ic or subacute hemorrhage) surrounding a cen¬ tral region of lesser Intensity. Bottom, Com¬ puted tomographic scan without contrast of a multilobar right cerebral hemorrhage (May 17, 1989) demonstrating delineation of superior gyri probably secondary to subarachnoid Involvement. demonstrated recent hemorrhage. A recent 3-cm encapsulated hematoma was located at the left frontotemporal junction. There was bilateral cortical loss in the temporal regions, with a resorbed hemorrhagic lesion on the left. The right parieto-occipital region re¬ vealed a resorbed hemorrhagic cortical le¬ sion that extended to the ventricles. A re¬ sorbed 2-cm hemorrhagic lesion was noted to the left of the splenium of the corpus callo¬ sum. The ventricular system was dilated, especially the temporal horns of the lateral ventricles. A "degenerative vasculopathy" with extensive amyloid deposition in cere¬ bral cortical leptomeningeal microvessels was noted. Bilateral infarcts were noted, as were multiple old and acute microinfarcts of the cerebellar cortex. Senile plaques were not as prominent as amyloidotic microvessels in the neocortex. CASE 2.—A 58-year-old normotensive black woman with a 44-pack per year history of smoking presented in 1984 with a 2- to 3week history of progressive confusion, diffi¬ culty walking, and incontinence. Beginning at the age of 20 years, she had experienced episodes of depression requiring prolonged hospitalization. Treatments had included electroconvulsive therapy and antipsychotic medications. She subsequently received the diagnosis of bipolar disorder with psychotic features and was also thought to have a mixed personality disorder. In the 2 years before the most recent admission, she had undergone increasing numbers of hospitalizations for major depressive episodes. Two weeks before admission, gait and speech dif¬ ficulties developed. Two days before admis¬ sion, the patient had nausea, vomiting, and urinary incontinence. On the day before hos- Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 pitalization, a headache developed. A CT scan of the brain with contrast revealed ringenhancing lesions, one in each frontal lobe, that were thought to be consistent with ma¬ lignant neoplasia (Fig 2). Results of a leftsided frontal craniectomy with ultrasoundguided needle biopsy revealed CAA, gliosis, and hemorrhage, but no evidence of malig¬ nant neoplasia. A follow-up CT scan revealed resolving lesions consistent with prior hema¬ tomas. She continued to have episodes of depression and experienced increased mem¬ ory loss and irritability. Three years postoperatively, an EEG demonstrated intermit¬ tent, generalized theta wave and polymorphic delta wave activity consistent with encephalopathy (toxic, metabolic, or de¬ generative) in the left anterior region. The EEG changes were not thought to be attrib¬ utable to resolving hematomas. CASE 3.—A 60-year-old normotensive white woman with a 25-pack per year smok¬ ing history presented with a 12-hour history of left-sided hemiparesis in 1988. She had a 9year history of transient ischemie attacks (TIAs). During the first 7 years, she had had three TIAs characterized by right-sided fa¬ cial and extremity paresthesias and weak¬ ness. The first two TIAs had also included an expressive aphasia. Computed tomographic scanning of the brain was repeatedly normal, as was a carotid artery duplex scan. Six months before admission, she suffered a TIA characterized by left-sided facial and ex¬ tremity paresthesias and weakness, and pos¬ sibly a seizure. Both CT and MRI scans were normal. On the day before admission, a head¬ ache and episodes of vomiting developed. In the 12 hours before admission, left-sided weakness developed as before, but she also demonstrated a rapidly progressing lethargy that culminated in a somnolent though painarousable state. A CT scan of the brain re¬ vealed a right frontal lobe hematoma (Fig 3), which was evacuated. The biopsy material revealed CAA. Follow-up CT scans revealed no further hemorrhage. She suffered sei¬ zures postoperatively and was treated with phenytoin and acetazolamide sodium. Six months postoperatively, she had continuing complaints of slurred speech and left-sided numbness and weakness. CASE 4.—A 69-year-old white male smok¬ er (two packs per day of unknown duration) with a history of controlled hypertension and carotid and peripheral artery disease was admitted for mental status changes and diffi¬ culties with activities of daily living in 1989. Four days before admission, headaches de¬ veloped. On the day before admission, he drove into a parked car. No clear history of head trauma was obtained. Subsequently, slight confusion and difficulty with ambulation and dressing developed. On examina¬ tion, he had a left-sided homonymous hemianopsia, a left-sided hemiplegia, and ataxia. A right frontal lobe biopsy specimen was believed initially to be consistent with a ster¬ ile abscess or cerebritis. Subsequent Congo red staining demonstrated CAA. Postopera¬ tively, he experienced fluctuating mental status changes with periods of confusion and disorientation and periodic episodes of left upper extremity weakness and numbness. Three months postoperatively, he suffered a severe episode of acute confusion. He was given colchicine (0.5 to 1 mg/d), which ap¬ peared helpful in reducing his neurological deficits for several months. By 1 year postop¬ eratively, he had increasing memory loss and irritability marked by rage attacks. CASE 5.—An 82-year-old white woman was admitted for severe headache, slurred speech, and right-sided hemiparesis. Her medical history included intraductal carcino¬ ma in 1987, a history of myocardial infarc¬ tion, and congestive heart failure. Her family reported that she had had frequent head¬ aches for several years. She did not smoke and was not hypertensive. She was living independently at home, where she cared for her husband. Three months before admis¬ sion, she fell at home and reported minor aches and pains, which resolved. In Decem¬ ber 1990, she suffered a severe headache, slurred speech, and a right-sided hemipare¬ sis. A head CT scan demonstrated a large left-sided occipital subdural and intraparenchymal hematoma with a 1.0-cm shift of midline structures to the right. A subdural and intraparenchymal hematoma measuring 6.0x4.0x2.0 cm was evacuated, and CAA was demonstrated in the biopsy material. Postoperatively, neurological status im¬ proved only minimally. A repeated head CT scan demonstrated some increased cerebral edema and continued right shift. Recovery of neurological function seemed a remote possi¬ bility at the time of manuscript preparation of this article. CASE 6.—In January 1991, an 86-year-old Japanese man with a vague history of cardiac disease had an episode of malaise, nausea, and vomiting. He became ataxic and con¬ fused. On examination, he had left-sided he¬ miparesis and was obtunded. A CT scan without contrast revealed a hemorrhage in the right parietal and temporal lobes with extension into the ventricles and a 1-cm midline shift to the left. He underwent a right- Fig 3. —Patient 3. Computed tomographic scan without contrast of a superficial frontal lobe hemorrhage. There is acute hemorrhage with surrounding edema (low density) and midline shift. Fig 4. —Patient 3. Cerebral biopsy specimen shows abundant acute hemorrhage and thick¬ ened ectatic amyloid-laden microvessels (ar¬ rows) (hematoxylin-eosin, sided parieto-occipital craniectomy, and the hematoma was evacuated. Microscopic ex¬ amination showed CAA. Postoperatively, he was alert and oriented but had a left-sided homonymous hemianopsia and left-sided he¬ miparesis. A CT scan revealed a largely evacuated hematoma and decreased midline shift. On postoperative day 5, pneumonia de¬ veloped. Treatment with antibiotics was fol¬ lowed by return to a fully oriented state after 1 day. He was subsequently transferred to a rehabilitation facility. RESULTS Summary of Biopsy Findings Brain tissue from all patients demon¬ strated the characteristic light micro¬ scopic features2,6 of CAA, with thick¬ ened, hyaline, eosinophilic vessel walls (Fig 4). Congo red staining confirmed the finding in all cases (Table 2). Bielschowsky's (silver) staining in selected cases showed a relative lack of senile (neuritic) plaques compared with large numbers of amyloidotic microvessels. Quantisation of this finding was difficult in view of the relatively small size of some of the biopsy specimens. All specimens revealed spongiosis and inflammatory cells. A frontal lobe biopsy specimen of patient 3 was re¬ markable for markedly ectatic amyloi¬ dotic vessels (Fig 4). A right frontal lobe biopsy specimen from patient 4 re¬ vealed focally prominent histiocytes and clusters of mononuclear cells, pri¬ marily in a perivascular location. Mito¬ ses seen (including granular forms) were among the reactive cells. Some larger arteries demonstrated fibrous in¬ timai hyperplasia and others showed or- Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 200). ganizing thrombi. Excessive numbers of thick-walled blood vessels initially suggested the presence of a vascular malformation. Both anti-A4 and anti--y-trace consis¬ tently labeled leptomeningeal and cere¬ bral microvessels in all patients (Fig 5). The anti-A4 labeled the few plaques ob¬ served, while - -trace did not. The antiA4 and anti-7-trace inconsistently la¬ beled reactive astrocytes (Table 2). In patient 2, gliosis was prominent, and strong immunoreactivity for both peptides was observed in numerous astro¬ cytes. In patient 4, reactive astrocytes were only lightly stained by anti-A4, while anti-7-trace labeling showed strong immunoreactivity. Conversely, in patient 1, anti-7-trace revealed mini¬ mal staining of astrocytes, but anti-A4 reacted strongly. The astrocytes from patient 6 showed only light staining by anti-A4. Rarely, strongly immunostained ves¬ sels (whether reactive for anti-A4 or -trace) were surrounded by acute hem¬ orrhage (Fig 6, left). Less commonly, sites of apparent extravasation of blood from immunoreactive vessels were en¬ countered (Fig 6, right). Large meningeal arteries, when present within the biopsy material, sometimes showed im¬ munoreactive material largely confined to their adventitial component (Fig 7). COMMENT Clinical Features The age of the patients with CAAassociated hemorrhage is comparable Fig 5. —Patient 1. Left, A section of brain biopsy specimen immunostained for Alzheimer A4 peptide demonstrates prominent staining of media and adventitia (arrow) of a small artery. Note the virtual absence of parenchymal (senile plaque) immunoreactivity. Center, Another section Immunostained for -y-trace peptide shows intense labeling of the vessel wall. Note rare astrocytes in the surrounding neuropil that show light cytoplasmic immunostaining (arrow). Right, Leptomeningeal vessels, immunostained for -y-trace. Note the patchy nature of peptide deposition within the vessel wall. The smaller of the two labeled vessels shows segments of its wall that do not demonstrate ¡mmunolabeling (arrow) ( 550). with that of previous reports of CAA, all indicating that the incidence of this CAA-related complication increases in those over 50 years of age.2,3'28,24 Four of the patients in this study are near or below the mean ages (male, 71 years; female, 73 years) of those with CAAassociated hemorrhages.3 No sex pre¬ dominance has been consistently docu¬ mented,3 and there was none in this small sample. The presentation of CAA may include nonspecific symptoms such as headache, nausea, and vomiting in addition to focal neurological deficits.25 Recurrent hemorrhages often occur.3 One report describes a 58-year-old woman who suffered a total of eight strokes.26 In our study, patient 1 ap¬ pears by imaging studies to have had seven hemorrhagic or ischemie events, though not all presented clinically as strokes. Of particular interest is patient 3, who experienced three TIAs refer¬ able to the left cerebral hemisphere over the course of 7 years and subse¬ quently had a TIA referable to the right hemisphere. Ultimately, she experi¬ enced hemorrhage in the hemisphere most recently affected by TIAs. It is unclear why the hemisphere that first produced symptoms was not the one in which clinically significant bleeding first occurred. Possibilities include two distinct types of CAA lesion (one more likely to hemorrhage than the other), or different rates of evolution in CAA-re- Fig 6. —Left, Anti-Alzheimer A4 peptide-immunostained section from patient 6 shows labeled parenchymal and meningeal vessels but virtually no senile plaques. An occasional parenchymal vessel (arrow) is surrounded by acute hemorrhage (x95). Right, A mlcrovessel from patient 5 shows -y-trace immunoreactivity (arrow) along one portion of its wall, while there is acute extravasation of blood at its opposite end. Rare labeled astrocytes (arrowhead) are present (AntiA4 and anti-gamma trace immunostaining performed as described in Vinters et al,17 625). lated lesions. Transient ischemie at¬ tacks referable to CAA have previously been described,27 and this patient's his¬ tory reinforces this rare origin of TIAs, Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 though the mechanism of such TIAs is not apparent. Microinfarcts or microhe- morrhages have been suggested as the cause.8 Radiological Features Fig 7. —Patient 5. A large meningeal artery (anti--y-trace stain) shows prominent adventitial peptide (arrows, left) that is not visible in a parallel section (right) stained with adsorbed antibody (see "Subjects and Methods" section) (left, 625; right, 245). While Torack2* has reported cerebral bleeding after and others29 have noted frequent hemorrhage during neurosurgical procedures, interventions ranging from biopsy to hematoma evacuation did not exacerbate intraoperative or postoperative hemorrhage in our pa¬ tients. A study of 11 patients with CAA also suggested that neurosurgery may be safely performed.30 Hypertension is not believed to be a major risk factor for CAA-associated hemorrhages, and only one of the six patients in this study (pa¬ tient 4) had a clear hypertensive histo¬ ry. Four of our patients were normotensive, with diastolic systemic blood pressures consistently not exceeding 90 mm Hg. There is no statistically sig¬ nificant difference in the incidence of infarcts or hemorrhages in hyperten¬ sive and nonhypertensive patients who have CAA associated with AD/SDAT,31 nor is there a correlation between ele¬ vated blood pressure and the presence of CAA.32 A history of heavy smoking, was pre¬ sent in at least four of six patients. Giv¬ en the advanced age of patients 5 and 6 and their fulminant presentation, an ab¬ sence of smoking history may not reflect their life experience. In patient 6, smok¬ ing history was not documented at all. Since smoke toxins are known to dam¬ age blood vessels,3334 it is interesting to speculate on a possible association be¬ tween smoking and symptomatic forms of this vasculopathy. In a study of pa¬ tients with "probable AD" by National Institutes of Health criteria, 37% of fe- males and 98% of males had used tobac¬ co, but only 10% of those females and 14% of those males were still smoking at the time of AD onset.35 The mean inter¬ val between smoking cessation and dis¬ ease onset was 18.3 years for women and 16.6 years for men. Average daily consumption before cessation of smok¬ ing was 18.2 and 18.8 cigarettes per day, respectively, and mean use was 32 years for females and 44 years for males. Since almost 100% of patients with AD have CAA, smoking may play a role in this vasculopathy. Unlike the majority of those clinically diagnosed with AD, four of our patients had not stopped smok¬ ing. Caution must be taken in drawing parallels between the AD smoking study and our patients, since none of the patients in our study was given the diag¬ nosis of AD by clinical criteria, and no independent assessment of CAA was made in the AD smoking study. A more extensive and rigorous study on a larger sample will be necessary to demon¬ strate conclusively an association be¬ tween smoking and CAA. Patient 2 demonstrated a long history of depression that appeared to be exac¬ erbated in the period preceding her CAA-associated hemorrhage, though whether this represents a truly organic effect is unclear. It is extremely unlike¬ ly, however, that her initial depression was due to CAA, since it began at the age of 20 years, when nonfamilial CAA is not known to occur.3 Of note is the information that her daughter also suf¬ fers from depression. Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 The radiological features of CAA may mimic other forms of brain abnormality that enter into the clinical differential diagnosis. While it has been suggested that CAA should be considered in the differential diagnosis when multiple subcortical lobar lesions appear in the normotensive elderly person, in prac¬ tice it may be difficult to distinguish from other lesions without a biopsy.36 When examined by CT scan without contrast, CAA hemorrhages tend to be superficial, lobar, dense when acute and hypodense when subacute or chronic, and somewhat irregular, and at times to show "fingers" or extensions, even into the subarachnoid space.36"3' Hemor¬ rhages may be multiple, bilateral, and recurrent. Resolution can result in leu¬ komalacia and localized ventricular dila¬ tation.39 Extension from the cortical and subcortical areas into the subarachnoid space appears common, though it does not usually delineate the gyri, as seen in Fig 1, bottom. However, pure subar¬ achnoid hemorrhage secondary to CAA is rare.37,4" Hemorrhage favors the frontoparietal cortex and white matter, though CAA may predominate in the parieto-occipital regions.24" While most CT scan studies of CAA-related hemor¬ rhages have been done without con¬ trast, the few done with contrast indi¬ cate that ring enhancement may occur.36,38 Filloux and Townsend38 noted that ring enhancement persisted up to 2 weeks after the first CT scan. Thus, waiting for a hemorrhagic CAA lesion to resolve as a means of distinguishing it from tumor may not be optimal. In par¬ ticular, multiple, peripherally placed ring-enhancing lesions that would be consistent with metastatic tumor in an elderly patient may require biopsy to rule out the latter possibility. Magnetic resonance imaging may also be superior to CT scan in the detection of subacute or remote hemorrhage. Patient 1 demonstrated multiple periventricular white-matter lesions (PWMLs) by MRI that were at first thought to represent demyelination sec¬ ondary to multiple sclerosis. However, autopsy revealed old infarcts and hem¬ orrhage related to CAA. Similar MRI findings of leukoencephalopathy in the periventricular regions have been noted in spontaneous and Dutch variants of CAA.41,42 However, patchy white mat¬ ter lesions may be detected by MRI in 20% to 30% of asymptomatic subjects and are associated with increased age and vascular risk factors such as hyper¬ tension, diabetes mellitus, coronary ar¬ tery disease, and TIAs.434·5 While PWMLs appear to be common in heredi¬ tary cerebral hemorrhage with amyloi- dosis-Dutch type (HCHWA-D),46 they appear less commonly in spontaneous CAA.14* The occurrence of PWMLs has been attributed to hypoperfusion of the deep white matter due to stenosis of long perforating arterioles.14 Periven- tricular white-matter lesions in HCHWA-D were not related to demen¬ tia according to Haan et al,46 since de¬ mentia was approximately equally prevalent in patients with and without PWML. Microscopic and Immunocytochemical Features in Relation to Pathogenesis of CAA Senile plaque cores in AD and SDAT, and amyloid in CAA, are composed of a 4200-d molecular weight protein that is only minimally different.47 This 4200-d Alzheimer A4 or ß-peptide is cleaved from a 695-amino acid precursor en¬ coded by a gene on chromosome 21.48~50 In our study, the immunohistochemical methods used involve polyclonal anti¬ bodies against the C-terminal end of the Alzheimer A4 precursor molecule. The site of origin of the A4 precursor mole¬ cule is controversial.6 Some patients demonstrate exclusive anti-A4 staining of cortical arteriolar or capillary walls, and others have strong perivascular and/or parenchymal immunolabeling. 17,2° In this study, arterioles, capillaries, and reactive astrocytes were labeled by both anti-A4 and anti-7-trace. Colocalization of both types of antibodies has also been observed in microvessels of patients with AD.01 The finding of prom¬ inent colocalization within vessel walls of A4 and 7-trace peptides in this group of six patients with cerebral hemor¬ rhage supports the view that 7-trace peptide may be a significant cofactor in the pathogenesis of bleeding in CAA.1718"51 Our patients were also of in¬ terest insofar as they usually showed predominantly or exclusively microvascular deposition of brain amyloid, in this regard resembling patients with HCHWA-D, though examination of more extensive brain tissue may show that significant amounts of plaque amy¬ loid were also present. The role of astrocytes in cerebral amyloidogenesis is unclear. The reac¬ tivity of astroglial cells with anti-A4 and anti-7-trace antibodies raises the possi¬ bility that these cells are phagocytosing peptides from the parenchyma, mi¬ crovessels, or perivascular regions. Ini¬ tial production of A4 or cystatin C may be by other cells. Anti-cystatin C serum has labeled neurons in neocortical biop¬ sy specimens of patients with cerebral tumors52 and in postmortem hypothalami.'3 In vitro, human astrocytes have demonstrated phagocytic potential.54 However, in animal models, adult astro- cytes do not appear to phagocytose col¬ loidal carbon, though there appears to be evidence that neonatal astrocytes have that capability.55"56 Alternatively, the astroglial cells themselves may be producing the amyloid peptides. Microglial and astroglial cells are capable of constitutive secretion of lysozyme and cystatin C in vitro.57 Interleukin 1 produced by microglial cells stimulates astrogliosis.58,59 Possibly, the stimulus for astrocytes to produce glial fibers may also trigger the increased produc¬ tion of both A4 precursor protein and cystatin C. Since not all astrocytes were immunolabeled, there may be either lo¬ cal differences in the quantities of amy¬ loid protein available for phagocytosis or varying levels of stimulus for the pro¬ duction of amyloid. In nonhereditary CAA, chronic in¬ flammation may be one starting point for amyloid formation. The late age at onset of CAA is consistent with a pro¬ longed exposure to some inflammatory agent. Since abnormalities of cerebral microvessels as seen in AVMs or angiitis11'1215·16 are associated with CAA, an inflammatory agent, in a similar man¬ ner, may change the properties of the microvessel walls over time to permit qi amyloidosis. Smoke may be such an agent, since it is known to cause damage to-fetal and adult vessels in animal mod¬ els and to adult vessels in humans.83,34,60'61 The blood-borne toxin may stimulate ameboid microglial cells (which are be¬ lieved to be the primary source of inter¬ leukin 1 in the brain62), macrophages, and possibly other cells to secrete that factor. Interleukin 1 produced by mac¬ rophages in the brain or in extracranial sites may in turn stimulate production of A4 precursor protein in their respec¬ tive locations. As suggested above, as¬ trocytes may be one source of the pep¬ tides. Under the influence of interleukin 1 or perhaps additional factors, periph¬ eral mononuclear or localized cell wall proteases may cleave A4 precursor pro¬ tein to form A4. Localized proteolytic enzymes have been suggested to ex¬ plain the selective distribution of CAA lesions.68 By postulating a factor similar to amyloid enhancing factor, a glyco- peptide that accelerates amyloid depo¬ sition in animal models of systemic amy¬ loidosis,64 and differing local levels of it, one could explain varying evolutional rates of CAA and related lesions. Finally, this study reemphasizes the importance of carefully examining evac¬ uated nontraumatic cerebral hemato¬ mas, especially from elderly patients, in an attempt to establish a causative mechanism for the bleeding.6'' Such ex¬ amination is enhanced by the use of immunohistochemical techniques (such as Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 we have described) in tissue from appro¬ priately selected patients. This research was supported in part by US Pub¬ lic Health Service grant R29 NS26312 and P30 AG10123 (Dr Vinters). A portion of this work was submitted in fulfill¬ ment of senior thesis requirements of the UCLA School of Medicine (Dr Yong). Illustrations were prepared by Carol Appleton and Stephen Kaufman, MS. References 1. Robert ME, Yong WH, Secor DL, Vinters HV. Diagnosing cerebral amyloid angiopathy in material from cerebral hemorrhages: the value of immunocytochemistry. Lab Invest. 1991;64:102A. Abstract. 2. Vinters HV, Gilbert JJ. Cerebral amyloid angiopathy: incidence and complications in the aging brain, II: the distribution of amyloid vascular changes. Stroke. 1983;14:924-928. 3. Vinters HV. Cerebral amyloid angiopathy: a critical review. Stroke. 1987;18:311-324. 4. Masters CL, Multhaup G, Simms G, Pottgiesser J, Martins RN, Breyreuther K. Neuronal origin of a cerebral amyloid: neurofibrillary tangles of Alzheimer's disease contain the same protein as the amyloid of plaque cores and blood vessels. EMBOJ. 1985;4:2757-2763. 5. Glenner GG, Wong CW. Alzheimer's disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun. 1984;120:885-890. 6. Vinters HV, Miller BL, Pardridge WM. Brain amyloid and Alzheimer disease. Ann Intern Med. 1988;109:41-54. 7. Glenner GG. On causative theories in Alzheimer's disease. Hum Pathol. 1985;16:433-435. 8. Gudmundsson G, Hallgr\l=i'\mssonJ, J\l=o'\nasson TA, Bjarnason O. Hereditary cerebral haemorrhage with amyloidosis. Brain. 1972;95:387-404. 9. Wattendorff AR, Bots GTAM, Went LN, Endtz LJ. Familial cerebral amyloid angiopathy presenting as recurrent cerebral haemorrhage. J Neurol Sci. 1982;55:121-135. 10. Glenner GG, Wong CW. Alzheimer's disease and Down's syndrome: sharing of a unique cerebro\x=req-\ vascular amyloid fibril protein. Biochem Biophys Res Commun. 1984;122:1131-1135. 11. Reid AH, Maloney AFJ. Giant cell arteritis and arteriolitis associated with amyloid angiopathy in an elderly Mongol. Acta Neuropathol. 1974; 27:131-137. 12. Gray F, Vinters HV, Le Noan H, Salama J, Delaporte P, Poirier J. Cerebral amyloid angiopathy and granulomatous angiitis: immunohistochemical study using antibodies to the Alzheimer A4 peptide. Hum Pathol. 1990;21:1290-1293. 13. Heffner RR Jr, Porro RS, Olson ME, Earle KM. A demyelinating disorder associated with cerebrovascular amyloid angiopathy. Arch Neurol. 1976;33:501-506. 14. Gray F, Dubas F, Roullet E, Escourolle R. Leukoencephalopathy in diffuse hemorrhagic cerebral amyloid angiopathy. Ann Neurol. 1985;18:54\x=req-\ 59. 15. Peterson EW, Schulz DM. Amyloid in vessels of a vascular malformation in brain. Arch Pathol. 1961;72:480-483. 16. Hart MN, Merz P, Bennett-Gray J, et al. \g=b\ x=req-\ Amyloid protein of Alzheimer's disease is found in cerebral and spinal cord vascular malformations. Am J Pathol. 1988;132:167-172. 17. Vinters HV, Secor DL, Pardridge WM, Gray F. Immunohistochemical study of cerebral amyloid angiopathy, III: widespread Alzheimer A4 peptide in cerebral microvessel walls colocalizes with gamma trace in patients with leukoencephalopathy. Ann Neurol. 1990;28:34-42. 18. Maruyama K, Ikeda SI, Ishihara T, Allsop D, Yanagisawa N. Immunohistochemical characterization of cerebrovascular amyloid in 46 autopsied cases using antibodies to beta protein and cys- tatin C. Stroke. 1990;21:397-403. 19. Vinters HV, Pardridge WM, Yang J. Immunohistochemical study of cerebral amyloid angiopathy: use of an antiserum to a synthetic 28\x=req-\ amino-acid peptide fragment of the Alzheimer's disease amyloid precursor. Hum Pathol. 1988;19:214-222. 20. Vinters HV, Pardridge WM, Secor DL, Ishii N. Immunohistochemical study of cerebral amyloid angiopathy, II: enhancement of immunostaining using formic acid pretreatment of tissue sections. Am J Pathol. 1988;133:150-162. 21. Ghiso J, Pons-Estel B, Frangione B. Hereditary cerebral amyloid angiopathy: the amyloid fibrils contain a protein which is a variant of cystatin C, an inhibitor of lysosomal cysteine proteases. Biochem Biophys Res Commun. 1986;136:548-554. 22. Ghiso J, Jensson 0, Frangione B. Amyloid fibrils in hereditary cerebral hemorrhage with amyloidosis of Icelandic type is a variant of gammatrace basic protein (cystatin C). Proc Natl Acad Sci USA. 1986;83:2974-2978. 23. Tomonaga M. Cerebral amyloid angiopathy in the elderly. J Am Geriatr Soc. 1981;29:151-157. 24. Hauw JJ, Vignolo P, Duyckaerts C, et al. \l=E'\tudeneuropathologique de 12 centenaires: la fr\l=e'\quence de la d\l=e'\mences\l=e'\nilede type Alzheimer n'est pas particuli\l=e`\rement\l=e'\lev\l=e'\edans ce groupe de personnes tr\l=e`\s\l=a^\g\l=e'\es.Rev Neurol. 1986;142:107\x=req-\ 115. 25. Ishii N, Nishihara Y, Horie A. Amyloid angiopathy and lobar cerebral hemorrhage. J Neurol Neurosurg Psychiatry. 1984;47:1203-1210. HA. Surface ultrastructure of human arteries with special reference to the effects of smoking. Acta Pathol Microbiol Scand Section A. 1979;87:201\x=req-\ 209. 35. Barclay L, Kheyfets S. Tobacco use in Alzheimer's disease. Prog Clin Biol Res. 1989;317:189\x=req-\ 194. 36. Wagle WA, Smith TW, Weiner M. Intracerebral hemorrhage caused by cerebral amyloid angiopathy: radiographic-pathologic correlation. AJNR. 1984;5:171-176. 37. Ohshima T, Endo T, Nukui H, Ikeda S, Allsop D, Onaya T. Cerebral amyloid angiopathy as a cause of subarachnoid hemorrhage. Stroke. 1990;21:480-483. 38. Filloux FM, Townsend JJ. Congophilic angiopathy with intracerebral hemorrhage. West J Med. 1985;143:498-502. 39. Brown RT, Coates RK, Gilbert JJ. Radio- graphic-pathologic correlation in cerebral amyloid angiopathy: a review of 12 patients. J Can Assoc Radiol. 1985;36:308-311. 40. Gilbert JJ, Vinters HV. Cerebral amyloid angiopathy: incidence and complications in the aging brain, I: cerebral hemorrhage. Stroke. 1983;14:915-923. 41. Loes DJ, Biller J, Yuh WTC, et al. Leukoencephalopathy in cerebral amyloid angiopathy: MR imaging in four cases. AJNR. 1990;11:485-488. 42. Haan J, Roos RAC, Algra PR, Lanser JBK, Bots GTAM, Vegter-van der Vlis M. Hereditary cerebral haemorrhage with amyloidosis-Dutch 26. Finelli PF, Kessimian N, Bernstein PW. Cerebral amyloid angiopathy manifesting as recurrent intracerebral hemorrhage. Arch Neurol. type: magnetic resonance imaging findings in 7 cases. Brain. 1990;113:1251-1267. 43. Bradley WG Jr, Waluch V, Brant-Zawadzki M, Yadley RA, Wycoff RR. Patchy periventricular 28. Torack RM. Congophilic angiopathy complicated by surgery and massive hemorrhage: a light and electron microscopic study. Am J Pathol. Imaging. 1984;1:35-41. 44. Gerard G, Weisberg LA. MRI periventricular lesions in adults. Neurology. 1986;36:998-1001. 45. Awad IA, Spetzler RF, Hodak JA, Awad CA, Carey R. Incidental subcortical lesions identified on magnetic resonance imaging in the elderly, 1984;41:330-333. 27. Smith DB, Hitchcock M, Philpott PJ. Cerebral amyloid angiopathy presenting as transient ischemic attacks: case report. J Neurosurg. 1985;63:963-964. 1975;81:349-365. 29. Roosen N, Martin JJ, De La Porte C, Van Vyve M. Intracerebral hemorrhage due to cerebral angiopathy: case report. J Neurosurg. 1985;63:965\x=req-\ 969. 30. Greene GM, Godersky JC, Biller J, Hart MN, Adams HP Jr. Surgical experience with cerebral amyloid angiopathy. Stroke. 1990;21:1545\x=req-\ 1549. 31. Ferreiro JA, Ansbacher LE, Vinters HV. Stroke related to cerebral amyloid angiopathy: the significance of systemic vascular disease. J Neurol. 1989;236:267-272. 32. Masuda J, Tanaka K, Ueda K, Omae T. Autopsy study of incidence and distribution of cerebral amyloid angiopathy in Hisayama, Japan. Stroke. 1988;19:205-210. 33. Kaufmann RC, Amankwah KS, Weberg AD. The effect of maternal smoke exposure on the ultrastructure of fetal peripheral blood vessels in the mouse. J Perinat Med. 1986;14:309-317. 34. Bylock A, Bondjers G, Jansson I, Hansson white matter lesions in the elderly: a common observation during NMR imaging. Noninvasive Med I: correlation with age and cerebrovascular risk factors. Stroke. 1986;17:1084-1089. 46. Haan J, Algra PR, Roos RAC. Hereditary cerebral hemorrhage with amyloidosis-Dutch type: clinical and computed tomographic analysis of 24 cases. Arch Neurol. 1990;47:649-653. 47. Prelli F, Castano E, Glenner GG, Frangione B. Differences between vascular and plaque core amyloid in Alzheimer's disease. J Neurochem. 1988;51:648-651. 48. Kang J, Lemaire HG, Unterbeck A, et al. The precursor of Alzheimer's disease amyloid A4 protein resembles a cell-surface receptor. Nature. 1987;325:733-736. 49. Tanzi RE, Gusella JF, Watkins PC, et al. Amyloid beta protein gene: cDNA, mRNA distribution, and genetic linkage near the Alzheimer locus. Science. 1987;325:880-884. 50. Tanzi RE, St George-Hyslop PH, Gusella JF. Molecular genetic approaches to Alzheimer's disease. Trends Neurosci. 1989;12:152-158. 51. Vinters HV, Nishimura GS, Secor DL, Par- Downloaded From: http://archneur.jamanetwork.com/ by a University of Iowa User on 06/21/2015 dridge WM. Immunoreactive A4 and gamma-trace peptide colocalization in amyloidotic arteriolar lesions in brains of patients with Alzheimer's disease. Am J Pathol. 1990;137:233-240. 52. L\l=o"\fbergH, Grubb AO, Brun A. Human brain cortical neurons contain gamma-trace: rapid isolation, immunohistochemical and physicochemical characterization of human gamma-trace. Biomed Res. 1981;2:298-306. 53. Bernstein HG, J\l=a"\rvinenM, P\l=o"\ll\l=a"\nenR, Schirpke H, Kn\l=o"\felB,Rinne R. CystatinCcontaining neurons in human postmortem hypothalamus. Neurosci Lett. 1988;88:131-134. 54. Watabe K, Osborne D, Kim SU. Phagocytic activity of human adult astrocytes and oligodendrocytes in culture. J Neuropathol Exp Neurol. 1989;48:499-506. 55. Al-Ali SY, Al-Zuhair AGH, Dawod B. Ultra- structural study of phagocytic activities of young astrocytes in injured neonatal rat brain following intracerebral injection of colloidal carbon. Glia. 1988;1:211-218. 56. Ling EA, Penney D, Leblond CP. Use of carbon labeling to demonstrate the role of blood monocytes as precursors of the 'ameboid cells' present in the corpus callosum of postnatal rats. JComp Neurol. 1980;193:631-657. 57. Zucker-Franklin D, Warfel A, Grusky G, Frangione B, Teitel D. Novel monocyte-like properties of microglial/astroglial cells: constitutive secretion of lysozyme and cystatin-C. Lab Invest. 1987;57:176-185. 58. Giulian D, Young DG, Woodward J, Brown DC, Lachman LB. Interleukin-1 is an astroglial growth factor in the developing brain. J Neurosci. 1988;8:709-714. 59. Giulian D. Ameboid microglia as effectors of inflammation in the central nervous system. J Neurosci Res. 1987;18:155-171. 60. Boutet M, Bazin M, Turcotte H, Lagac\l=e'\R. Effects of cigarette smoke on rat thoracic aorta. Artery. 1980;7:56-72. 61. Kjeldsen K, Astrup P, Wanstrup J. Ultrastructural intimal changes in the rabbit aorta after a moderate carbon monoxide exposure. Atherosclerosis. 1972;16:67-82. 62. Giulian D, Baker TJ, Shih LCN, Lachman LB. Interleukin 1 of the central nervous system is produced by ameboid microglia. J Exp Med. 1986;164:594-604. 63. Glenner GG. Amyloid \g=b\protein and the basis for Alzheimer's disease. Prog Clin Biol Res. 1989;317:857-868. 64. Brandwein SR, Sipe JD, Skinner M, Cohen AS. Effect of colchicine on experimental amyloidosis in two CBA/J mouse models: chronic inflammatory stimulation and administration of amyloid-enhancing factor during acute inflammation. Lab Invest. 1985;52:319-325. 65. Hinton DR, Dolan E, Sima AAF. The value of histopathological examination of surgically removed blood clot in determining the etiology of spontaneous intracerebral hemorrhage. Stroke. 1984;15:517-520.