Cerebral Amyloid Angiopathy: A Clinico-Radiological Study from South India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.280646
Source of Support: None, Conflict of Interest: None
Keywords: Cerebral amyloid angiopathy, cognitive involvement, intracerebral hemorrhage, MRIKey Messages: Our study on Clinico-radiological features of Cerebral amyloid angiopathy shows majority presenting with Intracerebral hemorrhage as the initial event and high prevalence of vascular risk factors and unrecognized cognitive symptoms at presentation.
Cerebral amyloid angiopathy (CAA), characterized by β-amyloid deposition in small and medium-sized arteries of the cerebral cortex and leptomeninges has been reported as a cause of intracerebral hemorrhage (ICH) for the last 40 years. Despite being reported as a major cause of spontaneous ICH in the Western population, it is often underreported from the Indian subcontinent, where spontaneous ICH accounts for over 11% of all strokes at the community level.
With the widespread availability of magnetic resonance imaging (MRI) with gradient and susceptibility sequences, probable/possible CAA can be diagnosed without autopsy or surgical biopsy using Boston criteria. Hence, more elderly are receiving a diagnosis of CAA in their lifetime. With the aging of the population and improved control of vascular risk factors, especially hypertension, physicians and neurologists from the developing countries such as India are increasingly likely to see this uncommon condition, which has important implications on planning risk factor management and antithrombotic use, especially dual antiplatelet and anticoagulation.
Hence, we decided to conduct this hospital-based study to understand the clinical and radiological profile of the patients with a diagnosis of probable/possible CAA and study their long-term outcome.
The study was conducted in …… Our study was a retrospective case series, where patients with a diagnosis of probable/possible CAA  were recruited after screening the medical records. The study recruitment period was 10 years (January 2006 to December 2015). The study was approved by the Institutional Ethics Committee. Demographic and clinical details were extracted using a structured proforma. We collected details on symptomatology, mean delay in diagnosis, vascular risk factors and prior antiplatelet and statin usage from medical records. Cognitive involvement was assessed using the mini-mental status exam (MMSE) and detailed neuropsychological battery for lobar functions wherever possible. MMSE <22/30 for those with <9 years of education and <26/30 for those with ≥9 years of formal education were taken as cutoffs for cognitive involvement, as per normative data published from our part of India previously. In cases where the formal cognitive assessment was not possible, we looked for any preceding self or spouse-reported cognitive symptoms. All the patients had undergone computerized axial tomography (CT)/MRI, which was reviewed from the patient archival and communication system (PACS) by a study neurologist (S.E.S.) and neuroradiologist (B.T.). All the subjects who had undergone MRI had T1, T2, FLAIR, GRE, susceptibility-weighted imaging, and diffusion-weighted imaging sequences along with vessel imaging—in the form of TOF–MR (time-of-flight–magnetic resonance) angiography of intra- and extracranial vessels. We looked for the presence of macro/microbleeds, their number and distribution, presence or absence of cortical subarachnoid hemorrhage (cSAH) and cortical superficial siderosis (cSS), defined as per standard definitions. We did not include cSS if it was contiguous with any ICH. The distribution of cSS and acute cSAH was classified as focal (restricted to ≤3 sulci) or disseminated (≥4 sulci). Also, we looked for the presence of brain atrophy, acute and chronic infarcts, and white matter changes, which was graded as per Fazekas grading. Follow-up of all patients was obtained from medical records or telephonic interview. In the follow-up, we looked for cognitive progression and any acute neurological events—strokes-hemorrhagic or ischemic/seizures or transient neurological events (TNE).
SPSS software, version 16 (SPSS Inc, Chicago, IL, USA) was used for statistical analysis. Baseline data were expressed in means and percentages. Fischer's exact test and Chi-square test were used to test the strength of association between variables. We looked for the association between each presenting symptom with vascular risk factors, age, gender, and imaging markers such as CMBs, white matter changes, cSAH, and cSS.
We had 28 subjects satisfying Boston criteria for probable/possible CAA during the study period who were included in the final analysis. The mean age of our cohort was 70.17 ± 8.85 years (range 55–87 years) with men outnumbering women (M: F ratio was 24:4). The majority of our patients were admitted with ICH (13/28, 46.4%). The diagnosis at admission is shown in [Table 1].
We found that many subjects had been symptomatic for long (mean 32 ± 29.3 months), ranging between 1 and 84 months from the first symptom till the diagnosis was made. Cognitive symptoms were the first symptom in a significant percentage of subjects, even though presentation as frank dementia was less common. Less than half of the patients (12/28, 42.9%) received a diagnosis of CAA at first presentation itself, despite having MRI done. Our subjects had a high prevalence of vascular risk factors also, which is shown in [Table 2].
Our 26 subjects had MRI (1.5 T Siemens Inc.) and two had only CT head for review. Imaging findings are summarized in [Table 3] and distribution of microbleeds is given in [Table 4].
In total, 50% of our patients were discharged with moderate to severe disability (modified Rankin scale 3 or above). We had at least a 12-month follow-up of 22 subjects and the mean follow-up duration was 33 ± 29 months, range 1–96 months. Recurrent neurological events were seen in eight of the subjects, with three having seizures, two each having TNE and ICH and one subject having an ischemic stroke. However, one-third of the patients (9/28) had cognitive progression during the follow-up.
Univariate analysis to test associations showed that those presenting as intracerebral bleed has no association with age, gender, presence of hypertension, prior antiplatelet or statin use, or neuroimaging characteristics such as the presence of CMB, cortical SAH, or leukoaraiosis. Those with cognitive symptoms as presentation also had no positive association with clinical parameters. However, we could find a weak correlation with severe leukoaraiosis (Fazekas grade 2 and 3; P = 0.04). The presence of infarcts or bleeds, macro- or microbleeds or SAH failed to show an association with the cognitive presentation. Those who had cognitive progression on follow-up did not show any statistically significant association with any of the clinical or imaging parameters. Seizures during the clinical course had a significant association with the presence of CS (P = 0.01); however, they failed to show an association with clinical variables or the presence of macro/microbleeds. Also, those with TNE during illness had a significant association with cSAH (P = 0.003).
With aging of the population and better primary and secondary prevention programs for lifestyle diseases, developing countries such as India are more likely to see neurodegenerative diseases and its myriad presentations in the coming years. There is a paucity of literature on CAA from the Indian subcontinent, with seven reports of nine patients so far.,,,,,, CAA, an important cause of nontraumatic ICH and cognitive decline in the elderly in the West is probably under recognized by physicians and neurologists, which prompted us to conduct this study.
Ours were a predominantly male cohort in the seventh decade of life. Aging is considered as the strongest risk factor for developing CAA with pathological studies indicating an increase in an amyloid deposition from seventh to ninth decade of life. We could not find any specific gender predilection for CAA in existing literature, even though most of the reports from India were of male patients.
Even though ICH was the main event that made our patients present to the hospital, the majority were having prior cognitive symptoms. TNE was seen in the one-third of patients. Of the nine reported patients from India, ICH was seen in the majority (7/9), and three had dementia.,,,,,, Surprisingly, none had seizures or ischemic symptoms at presentation. Asian literature on CAA indicates a lower incidence of spontaneous ICH related to CAA, which was explained by a higher incidence of hypertensive ICH in the Oriental population. Also, more cognitive involvement was found in the elderly secondary to CAA, especially in those with AD pathology. In comparison, Western data show that CAA is a major cause of spontaneous ICH in elderly, accounting for 5%–20% of all ICHs. Studies have also shown that the presence of CAA pathology increases the odds for developing dementia nine-fold, after adjusting for age and Alzheimer's pathology, and cognitive involvement has been reported to have association with CMB, micoinfarcts, and white mater changes as well. Rapidly progressive dementia secondary to CAA-related inflammation has been rarely reported, with only one biopsy-proven case so far from our part of the country.
After ICH, the most common acute clinical presentation described in CAA is TNE, also known as an amyloid spell. Marching type of sensory phenomenon is described, often difficult to distinguish from a sensory seizure, which was seen in most of our patients. Also, therapeutic response to antiepileptics in many of these patients suggests common pathogenesis for TNE and epilepsy in CAA patients. Even though CAA is reported as a cause for the late-onset seizure, the strength of the association is still not clear.
We failed to find any association of risk factors with ICH or cognitive decline in our cohort. Even though CAA risk may not be increased by conventional vascular risk factors, hypertension may contribute to CAA-related ICH in the Western population. However, Asian reports have been contradictory with authors from Japan and China reporting a lower prevalence of hypertension in CAA-related ICH (41%–45%) as compared with all causes for ICH (75%–80%)., Even though the prevalence of hypertension in our cohort was comparable to those with spontaneous ICH unrelated to CAA, it is difficult to explain the lack of positive association, which may have been confounded by the smaller sample size of those without hypertension. Statin use, which has been shown to increase the risk of spontaneous ICH, has till now not shown any consistent association with CAA-related ICH or presence of CMBs. Unlike statins, prior antiplatelet use has shown an association with CAA-related ICH and the presence of CMBs on imaging. However, we could not find a significant statistical association here also.
CAA-related macrobleeds, regardless of their size, exhibit a few common characteristics—distinctive cortico–subcortical distribution, multiplicity, and generally spares deep white mater (WM), basal ganglia, and brainstem, correlating well with the anatomic distribution of β-amyloid containing vessels. Less commonly cerebellum can be involved. We also found a similar pattern, with the majority having multiple macrobleeds and some with intraventricular extension. Presence of macrobleeds in the deep WM, basal ganglia, infratentorial–brainstem location, classical for hypertensive ICH has been reported by many authors in CAA also. Association of lobar ICH with CAA has been consistently reported in the literature, unlike other locations, which failed to show a positive correlation. In our study, we had a very high prevalence of hypertension in our cohort, comparable to those with hypertensive ICH, which might have been the reason why we failed to find an association with CAA. We hypothesize that many of our subjects might have had dual pathology, which needs further biopsy studies from our part of the world for confirmation.
Regarding associations of CMB, most studied ones are ICH and dementia. In a recent study on long-term follow-up of the CAA patients with CMBs alone without macrobleeds, authors found a small but substantial risk for ICH in their follow-up. Some have found CMBs increasing the risk of recurrent ICH in those with lobar ICH. We could not look into the association of CMB-only subjects with future ICH risk due to very small sample numbers in that subgroup. Those with CMBs have been reported to have worse performance in vascular dementia/cognitive decline. However, our analysis failed to find an association between lobar ICH and cognitive presentation/progression and CMBs, which might have been due to the high prevalence of hypertension and preexisting WM changes that can independently influence cognition and ICH risk.
Silent infarctions are well reported in the brains of patients with advanced CAA. Studies using MR-diffusion sequences have shown a high prevalence of silent infarcts in CAA, often associated with high CMB burden, recent ICH, and severe WM changes, suggesting a dynamic interplay of hemorrhagic and ischemic components of CAA. Even though presentation as the ischemic stroke was less common in our cohort also, we found that half of the patients had evidence of silent ischemia on imaging. The subcortical location of infarction in the majority may suggest an added role of hypertension in its pathogenesis.
The majority of our patients had moderate to severe leukoaraiosis, even though we could not find a posterior predilection, as reported by a few authors. We found a significant association for WM changes with cognitive impairment, which has been reported previously also. However, we failed to find an association with hypertension or lobar ICH as the majority had long-standing vascular risk factors.
Cortical SAH and cSS are well-described imaging findings in sporadic CAA. The largest cohort of isolated cSAH found that CAA was the most common etiology in patients over 60 years of age. cSS in our cohort had a strong association with TNE, which has been reported in several large series as well. Seizures are infrequently reported in CAA, often associated with CAA-related inflammation. Our finding of a strong association of seizures with cortical SAH has never been reported previously. Pathogenesis of TNE remains unclear and plausible explanations include focal seizure-like activity or migraine aura-like spreading depression. Common pathogenesis for TNE and seizure may be a logical explanation for our finding. Also, there are reports of patients with TNE responding to antiepileptics, which indirectly supports our hypothesis.
Our study is not without its limitations. Only two patients underwent emergency hematoma evacuation and we did not have histopathology confirmation in any of our patients. However, we made the diagnosis of probable/possible CAA as per Boston criteria, which is well validated in literature. We could get follow-up of 22 subjects only. Also, cognitive involvement must have been underreported in our study due to the retrospective nature of data collection. Also, the number of subjects with isolated CMBs or cSAH/cSS without lobar bleeds was less, to see the effect of these surrogate markers on symptom presentation, progression, and recurrent events.
Despite these limitations, we believe our study brings to light several less known clinical and imaging findings in CAA, which has been infrequently been reported from Asia, especially the Indian subcontinent. Seizures, even though rarely reported in CAA may be more common than previously thought of and need to be recognized early as various treatment options are available. cSAH patients showed a strong correlation with seizures, whereas a significant percentage of those with TNE had cSS on imaging. Leukoaraiosis of the severe grade was consistently found in patients with cognitive involvement at presentation, which was independent of age and hypertension.
Financial support and sponsorship
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest
The authors have no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4]