Cerebral Amyloid Angiopathy-Related Inflammation: A Case of Unexpected Cognitive Recovery

Cognitive impairment in the older population is becoming a worldwide epidemic dominated by neurodegenerative diseases, such as Alzheimer disease (AD). While it is estimated that AD will affect 14 million people in the United States by 2050¹, one-third of all adults will experience a stroke, dementia, or both in their lifetime, likely resulting in post-acute or long-term nursing care.² From postmortem studies, it is estimated that 34% of dementia cases demonstrate significant vascular disease³ and 64% of patients who experience a stroke will demonstrate clinical signs of dementia.⁴ 

Sporadic cerebral amyloid angiopathy (CAA) is an underdiagnosed small vessel disease of the brain characterized by the progressive deposition of amyloid β (Aβ) protein in the walls of small-to-medium sized arteries, arterioles, and capillaries in the cerebral cortex and overlying leptomeninges.^5,6 According to population based autopsy studies, the prevalence of CAA is 20% to 40% in elderly healthy controls and 50% to 60% in dementia patients.^7-11 Advancing age and Apolipoprotein E allele are the two known risk factors for sporadic CAA.^5,12 The vascular amyloid in sporadic CAA is mostly composed of the more soluble, 40 amino acid fragment, in contrast with amyloid plaques found in AD which are predominantly composed of the 42 amino acid residue fragment.^13-15 This suggests different pathophysiologic mechanisms for these two disease entities. Furthermore, there seems to be an independent contribution of CAA to cognitive dysfunction. Based on a clinical-neuropathological study, CAA was significantly associated with dementia (odds ratio, 9.3), even after controlling for age and AD-related neuropathologies (such as neuritic and diffuse plaques).⁹ Few studies have been able to identify a distinct clinical presentation of patients with sporadic CAA without AD. However, a postmortem study analysis demonstrated impaired processing speed and executive dysfunction in CAA as compared to the typical AD profile where episodic memory is also impaired.¹⁶ As such, we conceptually think about two distinct CAA patient populations in our practice: patients who have CAA concurrent with AD and patients who have sporadic CAA. 

Clinically, sporadic CAA is often asymptomatic and may only present in the workup of intracerebral hemorrhage, transient ischemic attacks, and/or cognitive decline.^17-19 As stated above, severe CAA is associated with an accelerated rate of decline in global cognition and increases the likelihood of dementia. In AD, severe CAA is present in about 25% of patients.²⁰ CAA may potentially worsen the cognitive dysfunction in AD due to ongoing microvascular and macrovascular damage. Based on one prospective study in patients with CAA, the incidence of new onset dementia 4.5 years after one intracerebral hemorrhage (ICH) was 29%.²¹

Importantly, there is a subset of CAA known as cerebral amyloid angiopathy-related inflammation (CAA-ri), which includes vascular inflammation in amyloid-affected vessels. This type of CAA is more commonly associated with headaches, focal neurologic symptoms, seizures, and recurrent intracranial microbleeds.^18,19,22 An important treatment consideration is immunosuppressive therapy often without the need for an invasive brain biopsy.^22-24 

We present a case of probable CAA-ri with fluctuating cognitive impairment and suspected CAA on magnetic resonance imaging (MRI). The patient did not receive steroid therapy despite a remarkable cognitive recovery following recurrent intracranial hemorrhages. The goal of this case report is to review the clinical presentation of CAA-ri and present management recommendations for post-acute and long-term care (LTC) providers.  

An 82-year-old white woman with a history of mild cognitive impairment (MCI), diabetes mellitus, microscopic colitis, heart failure, atrial fibrillation, and arthritis presented to a skilled nursing facility (SNF) for rehabilitation following a right frontal lobe hemorrhagic stroke.  

Her social history was significant for a 40 pack-year smoking history (quit 25 years prior to presentation), 14 years of education, and a reported intelligent quotient  of 128 (approximately 98th percentile) after the age of 20 years.  

Ten years prior to presentation, the patient had been given the diagnosis of AD and been prescribed cholinesterase inhibitors with resultant night terrors requiring discontinuation. She had a significant history of recurrent hemorrhagic strokes resulting in hospitalizations about 1 to 2 times per year over a course of 6 years prior to presentation. The clinical symptoms during the episodes were largely nonspecific, but common symptoms included headaches and episodes of confusion. Her medical history was significant for microscopic colitis, arthritis, and MCI for the previous 14 years, during which she had been prescribed various cholinesterase inhibitors. 

Five years prior to presentation, the patient underwent comprehensive neuropsychologic evaluation demonstrating impaired performance in abstract thinking and executive functioning, but with intact verbal and visual memory. Her performance on list learning and verbal delayed memory was superior (>95th percentile). During the 10 years prior to presentation, she had exhibited a clinical course of headaches, seizures, cognitive decline, and focal neurological complaints with an extraordinarily swift recovery following each episode.  

One month prior to presentation, she had been found unresponsive and subsequently diagnosed with multiple left and right frontal lobe intracerebral hemorrhages. At the time of presentation, she was receiving carvedilol, levetiracetam, furosemide, levothyroxine, losartan, memantine, esomeprazole, celecoxib, and trazodone.  

Head computed tomography imaging (Figure 1, Row A) following the episode of unresponsiveness 1 month prior to our evaluation demonstrated increased white matter hypodensities suggestive of vasogenic edema with a 5 mm hyperdensity in the right frontal lobe consistent with a focal parenchymal hemorrhage. Brain MRI (Figure 1, Rows B–D) obtained a week prior to the hospital admission demonstrated innumerable foci of old microhemorrhages throughout both cerebral hemispheres evident in susceptibility weighted imaging (SWI), suggesting CAA. These images showed worsening of microbleeds compared to an MRI 6 years prior to presentation.  

Upon physical examination, the patient was afebrile with an irregularly irregular heart rate of 60, blood pressure 111/63 mm Hg, and oxygen saturation of 96% on room air. She was minimally cooperative during exam; she expressed decreased concentration, but oriented to person and place. She demonstrated 0/3 item recall and could only name two animals in one minute. She had decreased coordination of upper extremities and difficulty following simple commands but without gross motor weakness. A review of laboratory data, including complete blood count, basic metabolic panel, and coagulation studies was unremarkable. Prior laboratory data revealed a sedimentation rate of 5 mm/hr and an antineutrophil cytoplasmic antibody screen that was negative.  

During her 104-day interrupted stay at the SNF,  her Montreal Cognitive Assessment (MoCA) scores ranged from 11/30 on presentation to 27/30 six weeks later. Prior cognitive screening assessment scores were not available for comparison purposes. Her rehabilitation course was significant for multiple falls, episodes of confusion, and multiple hospitalizations resulting in continued microbleeds with resultant step-wise deterioration in her cognition and functional status. She was eventually discharged from our SNF. To the best of our knowledge, she was readmitted to an outside hospital following a fall with altered mental status and discharged home. She subsequently died.

 While there has been substantial improvement in clinical outcomes associated with ischemic stroke, hemorrhagic stroke mortality continues to approach 50%.²⁵ We present a challenging case encountered in post-acute care that illustrates the broad differential of dementia and the clinical management challenges that arise in post-acute and LTC medicine. Sporadic CAA is characterized by progressive deposition of amyloid β proteins in the walls of cerebral blood vessels. CAA-related inflammation (CAA-ri) is a rare subset of CAA characterized by an inflammatory vasculitis resulting from the deposition of amyloid β protein. It is hypothesized that the deposition causes an autoimmune response to vascular amyloid β that initiates an inflammatory cascade, leading to ischemia through obstruction of the vessel lumen, increased local coagulation, and alterations in vasomotor tone.²⁶ This theory is further supported by the presence of anti-β amyloid antibodies in the cerebrospinal fluid (CSF) of patients with CAA-ri.²⁷ 

A common cause of lobar hemorrhage is CAA, which clinically can manifest as a waxing and waning mental status in a rare entity known as CAA-ri. The patient presented in this case report had probable CAA-ri and was given an incorrect diagnosis of AD at the age of 72 years. Although she had a clinical course and imaging markers consistent with CAA-ri, she did not receive immunosuppressive therapy.  

CAA vs AD Clinical Presentation 
One of the most important issues in caring for residents with CAA is to recognize the unique history of cognitive decline and how this differs from traditional neurodegenerative disorders encountered in post-acute and LTC settings. Most dementia syndromes encountered in the LTC care setting are AD and/or vascular dementia. AD generally presents as a slow cognitive trajectory over years and it predominantly affects episodic memory but can be of a mixed presentation if associated with CAA. As aforementioned, severe CAA concurrent with AD (which occurs in about 25% of AD patients) may worsen the cognitive decline typically associated with AD. Sporadic CAA without AD typically presents as a spontaneous lobar ICH in elderly patients or discovered incidentally in the workup of dementia.  

While sporadic CAA may present with lobar intracerebral hemorrhage, it can also present with transient focal neurological symptoms and signs (TFNSS), also referred to as “amyloid spells.” In case reports, these are reported as recurrent, stereotyped paraesthesias often lasting several minutes but can include focal seizures with or without Todd’s paralysis, focal deficits, or visual symptoms typically associated with migraines.¹⁷ Often, these episodes are misdiagnosed as transient ischemic attacks (TIA), leading to risky antithrombotic and/or antiplatelet therapy which may put patients at higher bleeding risk. While there is anecdotal data to support the use of antiepileptic (levetiracetam) and antimigraine (topiramate) medications to control the frequency and severity of TFNSS, randomized controlled data is not available.  

On the contrary, CAA-ri presents with less obvious neurologic changes (eg, headaches, seizures, subacute cognitive decline, delirium, and/or behavioral changes), a more pronounced waxing/waning course, and usually in younger patients (aged 67 vs 77 years) compared with CAA.^17-19 In 2011, Chung and colleagues proposed a set of diagnostic criteria for probable CAA-ri to improve diagnostic accuracy (Table 1).¹⁹ The patient presented in this case report meets all six criteria required for the diagnosis of probable CAA-ri. She presented with a clinical course of headaches, “seizures,” and focal neurologic complaints. These episodes were transient, and she had swift recovery after each episode. 

Additionally, a presentation of transient focal neurologic deficits is often not symptomatic of neurodegenerative disorders, including AD. Although they may occur, seizure disorders are not a common presentation in patients with neurodegenerative disorders. As such, when these are encountered in the evaluation of a patient with cognitive concerns, the diagnosis of CAA or CAA-ri should be considered. 

Even after a thorough clinical evaluation of mild cognitive decline, it is sometimes necessary to refer a patient for comprehensive neuropsychological evaluation to assist in making a correct diagnosis. In contrast to AD, which is mainly an amnestic syndrome, the cognitive profile in patients with CAA is generally nonamnestic, with deficits in executive function and processing speed, as aforementioned. The cognitive profile of the presented patient revealed impairment in abstract thinking and executive function with superior performance (scores of >95%) in tests of verbal list learning and delayed recall consistent with a diagnosis of single-domain nonamnestic MCI. It is important to highlight that this pattern is highly atypical for AD. Her prolonged clinical course at the stage of MCI (more than 14 years) also suggests that neurodegenerative disorders are likely not occurring.  

CAA-ri may present with behavioral and psychiatric symptoms, including delirium, depression, and personality changes. When presented with a clinical and radiographic presentation suggestive of CAA, it is important to consider a broad differential as CAA may be a coincidental finding in many patients. A short list to consider includes possible central nervous system (CNS) infections (eg, partially treated bacterial or fungal meningitis, neurosyphilis, or progressive multifocal leukoencephalopathy), autoimmune-related conditions (eg, neurosarcoidosis, acute disseminated encephalomyelitis), malignant processes (eg, primary CNS lymphoma, carcinomatous meningitis, or gliomatosis cerebri), or malignant hypertension (eg, posterior reversible encephalopathy syndrome). Each of these diagnoses present with unique laboratory characteristics. It is recommended that all patients undergo routine laboratory studies, including inflammatory markers. If the diagnosis is not clear, it is reasonable to proceed with examination of the cerebral spinal fluid (CSF), as it is helpful to exclude infectious and neoplastic processes. CSF studies of CAA-ri typically demonstrate a mild-to-moderate lymphocytic pleocytosis and a slight-to-moderate elevation in protein levels. While anti-β amyloid antibodies may be detected, they are not useful in the clinical diagnosis of CAA-ri.  

The most important diagnostic modality remains MRI with susceptibility-weighted imaging or gradient echo sequences. The patient presented had MRI evidence of numerous cortical and subcortical cerebral microbleeds, consistent with the diagnosis of CAA. Additionally, there was evidence of a few lobar macrohemorrhages such as would be detected in hemorrhagic stroke. To distinguish CAA from CAA-ri, a comprehensive retrospective clinical review at the Mayo Clinic demonstrated that leptomeningeal enhancement was more strongly associated with CAA-ri (sensitivity 70.4% and specificity 92.6%) and intracerebral hemorrhage was more strongly associated with CAA (sensitivity 62.9% and specificity 92.6%).²⁸   

Treatment Considerations 
Given the potential for response to immunosuppression, it is advisable to consider treatment. The most common treatment is corticosteroids, although treatment with cyclophosphamide, methotrexate, and mycophenolate mofetil has been reported.¹⁹ The optimal duration of therapy remains to be determined but should be based on clinical and radiologic response to immunosuppressive therapy. In one study, seven of 12 patients (58%) with CAA-ri demonstrated significant improvement within 1 to 2 weeks, while another three of 12 patients demonstrated an initial response offset by recurrence, resulting in a partial response rate of 83%.²⁹ In another cohort, five of six patients (83%) with CAA-ri and follow-up information demonstrated clinical and radiographic improvement after immunosuppressive therapy.¹⁸ 

Blood pressure management in older adults is controversial. However, in persons who have experienced a stroke or TIA, most guidelines suggest a systolic blood pressure less than 140 mm Hg.³⁰ In patients with CAA, the Progress Trial demonstrated that active treatment of blood pressure reduced the risk of CAA-related ICH by 77%.³¹ The blood pressure achieved in the treatment arm was 138/82 mm Hg from a baseline blood pressure of 147/86 mm Hg, suggesting that a target blood pressure of less than 140/90 mm Hg is reasonable and should be achieved in patients who have CAA.  

The risks associated with statin use following ICH remain controversial but appear to be different based on the underlying pathophysiology. In lobar ICH, statin therapy increased the annual probability of ICH recurrence from 14% to 22%, thus shifting the risk-benefit analysis for cardiovascular protection to an overall increased risk with use of statins.³² In deep ICH, most commonly caused by hypertensive vascular disease, the harm attributed to statin therapy was not as clear. Given this data, it is advisable to avoid statin therapy in older adults with a history of repeated CAA-related lobar hemorrhagic strokes.  

The clinical course of CAA-ri is variable due to the underlying pathophysiology. However, there is typically a favorable response to immunosuppressive therapy within a few weeks. As such, in select patients who meet the criteria for probable CAA-ri, as illustrated in Table 1, and who do not have absolute contraindications to immunosuppressive therapy, a trial of immunosuppressive therapy should be considered without a brain biopsy. It is important to note the disadvantage of diagnostic uncertainty of early immunosuppressive therapy without a brain biopsy. Therefore, those who fail to respond to high-dose corticosteroid therapy within 3 weeks should have the diagnosis confirmed with a brain biopsy (Table 2).  

CAA is a common, yet underdiagnosed condition characterized by the deposition of amyloid proteins in cerebral blood vessels. A less well-known entity is CAA-ri, which is potentially reversible with early immunosuppressive therapy. Given the potential for response to immunosuppression, it is advisable to consider treatment with corticosteroid therapy. Diagnosis should be confirmed via brain biopsy for patients who do not respond to corticosteroid therapy within 3 weeks. 

Authors: Christian Bergman, MD¹ • Hamid R Okhravi, MD²

Affiliations:
¹Division of Geriatric Medicine, Virginia Commonwealth University, Richmond, VA
²Glennan Center for Geriatrics and Gerontology, Eastern Virginia Medical School, Norfolk, VA

Disclosure: The authors report no relevant financial relationships.

Address correspondence to:
Hamid R. Okhravi, MD
825 Fairfax Avenue, Suite 201
Hofheimer Hall Norfolk, VA 23507
Phone: (757) 446-7040
Email: OkhravHR@evms.edu