Premium
P1‐330: THE RELATION BETWEEN MRI FINDINGS AND COGNITIVE IMPAIRMENT IN CHRONIC KIDNEY DISEASE (CKD) PATIENTS AND THE MODIFYING EFFECT OF ALBUMINURIA
Author(s) -
Murray Anne,
Burns Christine M.,
Vemuri Prashanthi,
Davey Cynthia,
Tupper David,
Lundt Emily S.,
Wille Samantha M.,
Steinert Kaely,
Jack Clifford R.,
Knopman David S.
Publication year - 2019
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2019.06.885
Subject(s) - albuminuria , medicine , renal function , kidney disease , cognition , brain size , creatinine , cohort , logistic regression , biomarker , cardiology , magnetic resonance imaging , radiology , psychiatry , biochemistry , chemistry
hemorrhagic complications is reported. The increased prevalence of cerebral microbleeds (CMB) and cortical superficial siderosis (cSS) reflecting amyloid burden is one of the explanatory mechanisms. They are not included as an influencing factor in the scores available, based on clinical variables but not neuroimaging markers of the future hemorrhagic risk. We investigated the association of four clinical scores to predict major bleeding in patients under oral anticoagulation: ATRIA, OBRI, HAS-BLED and HEMORR2HAGES, with CMB and cSS. Methods: We retrospectively applied the scores on 600 patients from Geneva University Hospitals memory clinic. Clinical information available at the time of brain MRI was used to generate all scores, blinded to brain MRI findings. Group comparisons were performed using the Fischer’s exact test or unpaired t-test, as appropriate. We first investigated the prediction strength by multiple logistic regression models. In a second step, the area under the receiver operating characteristic curve (AUC) was used as an individual measure of the overall predictive accuracy for each score regarding CMB and cSS. Finally, we compared the performances of the four scores to predict the outcomes using a test of equality of the AUCs. Results: Patients with CMB had an increased prevalence of hypertension (65.7% vs. 53.4%, p1⁄40.003) and heart failure (19.9% vs. 10%, p1⁄40.001), as well as an increased risk of major bleeding, according to all four scores assessed. In the logistic regression models, only HAS-BLED and HEMORR2HAGES were associated with both 5 CMB (OR 1.25, 95%CI 1.051.49, p1⁄40.012) and cSS (OR 1.54, 95%CI 1.26-1.89, p1⁄40.0001), the latter score with a higher prediction strength. This was supported by the ROC curve analysis, HEMORR2HAGES score presenting the best sensitivity/specificity report for 5 CMB (AUC 0.643, 95I%IC 0.57-0.72, p1⁄40.019) and cSS (AUC 0.729, 95%IC 0.63-0.83, p1⁄4 0.0071) prediction. Conclusions: Bleeding risk scores are associated with small, but clinically relevant hemorrhagic lesions, as the presence of 5 CMB and cSS. Among the four scores, HEMORR2HAGES presents the best sensitivity/specificity report, but the accuracy for prediction on a per-patient level deserves further studies.