It used to be that clinicians could only diagnose inflammation related to cerebral amyloid angiopathy by sticking a needle in the brain and taking a biopsy. Seem unpleasant? Clinicians don’t like it much either. Now, scientists led by Steven Greenberg at Massachusetts General Hospital, Boston, have devised non-invasive diagnostic criteria for this relatively rare condition. They are based on clinical symptoms and magnetic resonance imaging (MRI). In patients who had pathologically confirmed CAA-related inflammation (CAA-ri), the criteria identified the disorder with high specificity and sensitivity, the scientists report in the December 28 JAMA Neurology online. Since this condition responds well to immunosuppressive therapy, these guidelines will allow doctors to more easily treat affected patients.

“These criteria diagnosed people with CAA-ri without false positives, which is what we really wanted to avoid,” said Greenberg. “We don’t want to treat unaffected individuals if we can help it.” 

Lopsided Lesion:

Swelling in the brain due to CAA-related inflammation is patchy and asymmetrical (left), as opposed to changes (right) seen in normal aging. [© 2015 American Medical Association. All rights reserved.]

“Differentiating patients with CAA from those with the inflammatory component is critical,” wrote Costantino Iadecola of Weill Cornell Medical College, New York, who was not involved in the study (see full comment below). Iadecola said he believes further validation is needed in a larger group of patients, but added, “the study is an important first step enabling physicians to swiftly identify CAA-ri and initiate treatment.” 

CAA-ri is closely related to ARIA-E, the inflammation sometimes seen in AD patients treated with antibodies to Aβ (Werring and Sperling, 2013; Jul 2011 conference news). In CAA, Aβ builds up on the walls of arteries in the brain. Some patients show no symptoms, but others suffer intracranial bleeds or dementia. In addition, a small fraction of patients get CAA-ri, which can cause swelling, headaches, seizures, and stroke. Most of these symptoms resolve with immunosuppressive drugs such as high-dose corticosteroids. These drugs can cause psychosis and hyperglycemia, so doctors prescribe them judiciously. For the patients most likely to benefit, clinicians currently resort to brain biopsies, which are unpopular with patients. Can they get diagnosed another way?

Previously proposed diagnostic criteria based on clinical and radiological characteristics specified that patients would have a headache that persisted for days to weeks, seizures, and/or focal neurological deficits, and be less alert or exhibit behavioral changes. MRI should show evidence of patchy swelling, multiple microbleeds, and intracranial hemorrhage. All this must occur in the absence of cancer, infection, or other known causes (Chung et al., 2011). 

Greenberg and colleagues refined these criteria. They propose that the headaches and behavioral problems could occur over months to years, and specify that swelling should be asymmetrical and extend outward toward the surface of the brain (Linn et al., 2010). They also include cortical superficial siderosis among the types of bleeding. In their experience, this type of iron deposition underneath the pia mater occurs more frequently in people with CAA-ri.

In the current study, first author Eitan Auriel tested these criteria in patients who died with CAA and were autopsied. Combing through the medical records of Massachusetts General Hospital and the University of Milano-Bicocca, Monza, Italy, the researchers came up with 17 patients with autopsy-confirmed CAA-ri and 37 with non-inflammatory CAA over18 years. Fourteen of the 17 CAA-ri patients met the proposed criteria for probable disease, while only one person with non-inflammatory CAA did. That patient may actually have had CAA-ri, wrote the authors, as his records showed he had recovered on a short course of heavy-duty immunosuppressive drugs. Overall, the criteria had a specificity of 97 percent and sensitivity of 82 percent for detecting CAA-ri.  

“The authors put forward a strong case for using clinical and MRI criteria for diagnosing CAA-ri,” said Roy Weller, University of Southampton, U.K. (see full comment below). “If the diagnosis can be made with a high degree of certainty, this would avoid an invasive cerebral biopsy, which would in itself add to the discomfort of the patient and increase the risk of complications.” Greenberg noted that MRIs are less than clear-cut for some patients, who will still require a biopsy. Even as Greenberg continues to collect data on the effectiveness of these guidelines and encourages other groups to help validate them, he claims the current data is sufficient to put the criteria into clinical practice now. He did acknowledge that this retrospective study of a small group of autopsy patients likely reflects the most severe cases, and agreed that it remains to be seen how well these criteria will generalize to all CAA-ri patients.

These new criteria may help physicians managing ARIA-E in AD immunotherapy trials, once any such therapy is approved.  “If AD immunotherapy is effective, then understanding how to predict, detect, and treat ARIA will be a major focus of putting that therapy into practice,” Greenberg told Alzforum. Iadecola agreed. “If CAA-ri is the basis of ARIA, immunosuppression may allow affected patients to continue immunotherapy.” He noted that ARIA is less serious than CAA-ri, so not all cases of ARIA may require treatment. “Nevertheless, having the opportunity to prevent or treat ARIA would expand the number of patients who could benefit from immunotherapy,” he said.—Gwyneth Dickey Zakaib


  1. We feel that this paper by Steve Greenberg and colleagues is a very important contribution for two main reasons.

    1. Cerebral amyloid angiopathy with related inflammation (CAA-ri) is a severe complication of cerebral amyloid angiopathy; it is characterized by progressive cognitive decline, seizures, and headaches. Many patients with CAA-ri respond to steroids, so making the diagnosis is a very important step. On the positive side, once the diagnosis is made, the patients can be treated and many will recover. Giving steroids to patients without CAA-ri, however, may result in complications from the steroid therapy with no therapeutic benefit. What the present paper does is take a group of biopsy-proven CAA-ri patients and show that the clinical and MRI findings are different from the group of patients with CAA and no related inflammation. If the diagnosis can be made with a high degree of certainty on clinical and MRI criteria alone, as suggested by this paper, this would avoid an invasive cerebral biopsy which would, in itself, add to the discomfort of the patient and increase the risk of complications. The authors put forward a strong case for using clinical and MRI criteria for diagnosing CAA-ri.

    2. Another important contribution is the comparison of imaging in CAA-ri with the results of the bapineuzumab trials that show clinical and imaging abnormalities similar to CAA-ri. MRI suggests that in both cases there is an increase in brain tissue fluid and this supports the hypothesis that there is impairment of drainage of interstitial fluid from the brain along the walls of arteries in both conditions (Weller et al., 2015). In patients treated with bapineuzumab, it appears that amyloid removed from brain tissue by the immunotherapy passes into the perivascular drainage pathways in artery walls and blocks those pathways. This results in a failure of fluid drainage from the underlying white matter and the corresponding changes on MRI. In the CAA-ri cases, it appears that the inflammatory reaction associated with amyloid in the vessel walls also impairs fluid drainage from the white matter. Thus, in addition to establishing firm diagnostic criteria for CAA-ri, the present paper brings into focus the physiological effects on the brain of CAA and CAA-ri that help us understand the pathological effects of increasing amyloid in vessel walls and the effects of inflammation. It would be interesting to see the MRI following steroid therapy in the CAA-ri patients to observe the timing and completeness of resolution of white-matter fluid retention. This would give some indication of how restoration of fluid flow correlates with resolution of pathological lesions in CAA-ri.

    The mechanisms behind the development of CAA-ri are not known. That in CAA-ri there are lymphocytes in the arterial wall rather than simply perivascular macrophages or lymphocytic perivascular cuffs suggests that it is possible that immune complexes may form in the walls of arteries, blocking the perivascular drainage of interstitial fluid and Aβ (Carare et al., 2013; Teeling et al., 2012). The presence of auto-antibodies against Aβ in the CSF of this group of patients further strengthens the hypothesis that immune complexes may form in the brains of some CAA patients, impeding the perivascular clearance of solutes along basement membranes and worsening the clinical presentation (Piazza et al., 2013). This valuable clinical study should form a platform for experimental studies into the pathophysiology of perivascular lymphatic drainage of solutes and how it interacts with the innate and adaptive immune responses.


    . White matter changes in dementia: role of impaired drainage of interstitial fluid. Brain Pathol. 2015 Jan;25(1):63-78. PubMed.

    . Immune complex formation impairs the elimination of solutes from the brain: implications for immunotherapy in Alzheimer's disease. Acta Neuropathol Commun. 2013 Aug 9;1(1):48. PubMed.

    . Intracerebral immune complex formation induces inflammation in the brain that depends on Fc receptor interaction. Acta Neuropathol. 2012 Oct;124(4):479-90. Epub 2012 May 18 PubMed.

    . Anti-amyloid β autoantibodies in cerebral amyloid angiopathy-related inflammation: Implications for Amyloid-Modifying Therapies. Ann Neurol. 2013 Feb 11; PubMed.

  2. This paper by Auriel et al. addresses an important clinical problem. CAA with an inflammatory component is an aggressive and devastating disease, possibly autoimmune-based, but can be ameliorated by treatment with immunosuppression. Therefore, differentiating patients with CAA and no neuroinflammation from those with the inflammatory component is critical. Although diagnosis can be made with a brain biopsy, this is an invasive procedure with significant risks for the patient.

    Auriel et al. developed criteria that permit the diagnosis of CAA-ri with a high degree of sensitivity and specificity without a brain biopsy. Although further validation is needed in a larger group of patients, the study is an important first step in enabling physicians to swiftly identify CAA-ri and initiate treatment.

    Inasmuch as CAA-ri resembles the ARIA syndrome observed in patients undergoing immunotherapy for AD, the findings may be relevant also to AD patients who, due to this complication, cannot fully benefit from immunotherapy. It’s a bit of a stretch, but if CAA-ri is at the basis of ARIA, immunosuppression may be beneficial, allowing the affected patients to continue immunotherapy. On the flip side, ARIA is not as serious as CAA-ri, so treatment may not be warranted. Nevertheless, having the opportunity to prevent or treat ARIA would expand the number of patients who could benefit from immunotherapy.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Paris: Renamed ARIA, Vasogenic Edema Common to Anti-Amyloid Therapy

Paper Citations

  1. . Inflammatory cerebral amyloid angiopathy and amyloid-modifying therapies: variations on the same ARIA?. Ann Neurol. 2013 Apr;73(4):439-41. PubMed.
  2. . Cerebral amyloid angiopathy related inflammation: three case reports and a review. J Neurol Neurosurg Psychiatry. 2011 Jan;82(1):20-6. PubMed.
  3. . Prevalence of superficial siderosis in patients with cerebral amyloid angiopathy. Neurology. 2010 Apr 27;74(17):1346-50. PubMed.

Further Reading


  1. . An Individual with Cerebral Amyloid Angiopathy-Related Inflammation Who Displayed Involuntary Movements. J Am Geriatr Soc. 2015 Dec;63(12):2644-2645. PubMed.
  2. . Cerebrospinal Fluid Anti-Amyloid-β Autoantibodies and Amyloid PET in Cerebral Amyloid Angiopathy-Related Inflammation. J Alzheimers Dis. 2015;50(1):1-7. PubMed.
  3. . Reduction of microbleeds by immunosuppression in a patient with Aβ-related vascular inflammation. Neurol Neuroimmunol Neuroinflamm. 2015 Dec;2(6):e165. Epub 2015 Oct 15 PubMed.
  4. . Cerebral amyloid angiopathy-related inflammation: A potentially reversible cause of dementia with characteristic imaging findings. J Neuroradiol. 2016 Feb;43(1):11-7. Epub 2015 Oct 23 PubMed.

Primary Papers

  1. . Validation of Clinicoradiological Criteria for the Diagnosis of Cerebral Amyloid Angiopathy-Related Inflammation. JAMA Neurol. 2016 Feb;73(2):197-202. PubMed.