Use Lay Language. CAA-ri : a Series of 28 Patients

We retrospectively reviewed the clinical, radiological and biological data of 28 consecutive patients with a diagnosis of CAA-ri between 2004 and 2016, in 6 different neurological centers from Occitanie in the south of France (Montpellier, Toulouse, Nimes, Narbonne, Bezier, Perpignan). Inclusion criteria were the following: 1) pathological diagnosis assessed by both CNS Aß vascular deposition and transmural of perivascular inflammation or 2) a probable or possible CAA-ri according to the recently criteria proposed by Auriel et al. (Auriel, JAMA 2016). CAA diagnosis before onset was done retrospectively according to Boston Criteria (Knudsen, Neurology, 2001) (Smith, Cardiovascular disease and stroke, 2003).

1. Clinical data For each patient, the following demographic and clinical data were systematically recorded: age, gender, neurological medical history, independence, neurological symptoms, initial evolution at discharge, treatment type and duration and follow-up data. Onset of the disease was defined as acute (<48h), sub acute (>48h and <4weeks) or chronic (>4weeks) depending on the time elapsed between the manifestation of the first symptoms and time of referral to the treating center. Response to treatment was assessed according to the treating neurologist global opinion. Relapse was defined as a recurrence of symptoms and new/enlarging lesions on cerebral MRI. Disability at the end of follow-up was defined using the modified Rankin scale (mRS) through a review of the last medical record by two neurologists (SC, XA). Outcome was dichotomized into good outcome (mRS 0 to 2) and poor outcome (3 to 6).
2. Laboratory data Laboratory data included ApoE genotype in n=20 patients and cerebrospinal fluid (CSF) analysis (n=25 patients). In addition to the standard test, CSF analysis included Alzheimer's disease (AD) biomarkers study: results were compared with series of patients with AD, CAA and control patients (Charamidou, JAD 2016) (Renard, JAD, 2016). Cerebral biopsy was performed in 6 patients.
3. Neuroimaging data MRI examinations were performed either at 1.5 or 3 Tesla for all the patients upon routine clinical care. Sequences included T1-weighted pre and post contrast, T2, Flair, diffusion, T2-Gradient Echo sequencing (T2-GRE). SWI was performed in only 5 patients and was thus not included in the analysis. Imagings of all patients were reviewed by two neurologists (SC, XA) and a neuroradiologist (NM). On the first MRI performed at diagnosis, confluent T2/FLAIR white matter hyperintensity (WMH) Figure 1A, size was characterized using a qualitative score as follow: small (<50% of one lobe), medium (>50% of one lobe and less than an entire lobe) and large (the volume of confluent WMH uni or multifocal exceeding one lobe), and mass effect was classified as mild (sulcal effacement), moderate (mass effect on the lateral ventricle) or severe (subfalcine herniation). The presence and number of lobar intracerebral macrohemorrhage (ICH), cerebral microbleeds (CMBs) and cortical superficial siderosis (cSS) were evaluated on T2-GRE. The distribution and severity of cSS was classified as focal (restricted to ≤3 sulci) or disseminated (≥4 sulci) (Roongpiboonsopit, Neurology, 2016). Localization and number of Dwi-positive cerebral infarcts as well as the presence of contrast enhancement were noticed. Almost all the patients (96%) had at least one follow-up MRI enabling analysis of the evolution of the main lesion, occurrence of infarcts, ICH, the raise CMBs' number and the contrast enhancement Figure 1. This series comprises only routinely collected clinical, biological and radiological data and was judged not to require formal ethical committee approval.
4. Statistical analysis Data were summarized for continuous variables as mean ±SD (Max, Min) and number (%) for categorical variables. Statistical comparisons were performed by Fisher exact test for categorical variables and non-parametric Wilcoxon rank sum test for continuous variables. Comparison of continuous variables between groups was performed using non-parametric. Survival was estimated with the Kaplan-Meier method. The relationship between categorical parameters and outcome measures was assessed by non-parametric log-rank test. Finally we use an ANOVA test to compare biomarkers between CAA-ri, Alzheimer's Disease (AD) and controls. For all analyses, a two-tailed p-value <0.05 was used as the criterion for statistical significance. Statistical analyses were performed using R version 3.0.1 (The R Foundation for Statistical Computing). To test the association between MCB's number and clinical outcome we categorized the number of microbleeds in 3 categories: 0 to 99 CMB, 100 to 199 and more than 200 per patient.