The Study of Risk Factors and The Predict Model of Cerebral Microbleeds

Abstract

Background: To determine the association of cardiovascular risk factors with the presence and anatomic location of CMBs and construct a factor-based evaluating model to predict a high CMBs burden.

Methods: We assessed the relation of age, male, various cardiovascular risk factors, medication use, stroke histories and white matter hyperintensities (WMH) to the presence and location of CMBs with univariate analysis and multiple logistic regression. Finally, we added risk factors for a high CMBs burden to a factor-based evaluating model score.

Results: 485 patients were included in our study. CMBs were more prevalent with advanced age, male sex, more cardiovascular risk factors and WMH. Alcohol use, hemorrhagic stroke history and the degree of deep white matter hyperintensity (DWMH) were independent predictors for a high CMBs burden (≥10). We finally structured a prediction model-HPSAD3 that consisted of hypertension, alcohol use, hemorrhagic stroke history and WMH to predict a high CMBs burden. The model-HPSAD3 has a higher positive predict value (77.08%) and negative predict value (75.89%) to predict a high CMBs burden when the cut-off score is 4.

Conclusions: Hypertension, alcohol use, hemorrhagic stroke history and WMH were added into the model- HPSAD3, and there was a higher possibility of patients with CMBs ≥10 when the score of HPSAD3 ≥4.

Keywords: Cerebral microbleeds; HPSAD3; high CMBs burden; prediction model; risk factors; white matter hyperintensities.