Early neurological and cognitive impairments in subclinical cerebrovascular disease

The subclinical cerebrovascular disease (SCVD) is an important public health problem with demonstrated prognostic significance for stroke, future cognitive decline, and progression to dementia. The earliest possible detection of the silent presence of SCVD in adults at age at risk with normal functioning is very important for both clinical doctors and scientists.

Seventy-seven adult volunteers, recruited during the years 2005-2007, with mean age 58.7 (standard deviation 5.9) years, were assessed by four subtests from the Cambridge Neuropsychological Test Automated Battery (CANTAB)-Eclipse cognitive assessment system. We used a questionnaire survey for the presence of cerebrovascular risk factors (CVRFs) such as arterial hypertension, smoking and dyslipidemia, among others, as well as instrumental (Doppler examination) and neurological magnetic resonance imaging (MRI) procedures. Descriptive statistics, comparison (t-test, Chi-square) and univariate methods were used as followed by multifactor logistic regression and receiver operating characteristics analyses.

The risk factor questionnaire revealed nonspecific symptoms in 44 (67.7%) of the subjects. In 42 (64.6%) of all 65 subjects, we found at least one of the conventional CVRFs. Abnormal findings from the extra- and trans-cranial Doppler examination were established in 38 (58.5%) of all studied volunteers. Thirty-four subjects had brain MRI (52.3%), and abnormal findings were found in 12 (35.3%) of them. Two of the four subtests of CANTAB tool appeared to be potentially promising predictors of the outcome, as found at the univariate analysis (spatial working memory 1 [SWM1] total errors; intra-extra dimensional set 1 [IED1] total errors [adjusted]; IED2 total trials [adjusted]). We established that the best accuracy of 82.5% was achieved by a multifactor interaction logistic regression model, with the role CVRF and combined CANTAB predictor "IED total ratio (errors/trials) × SWM1 total errors" (P = 0.006).

Our results have contributed to the hypothesis that it is possible to identify, by noninvasive methods, subjects at age at risk who have mild degree of cognitive impairment and to establish the significant relationship of this impairment with existing CVRFs, nonspecific symptoms and subclinical abnormal brain Doppler/MRI findings. We created a combined neuropsychological predictor that was able to clearly distinguish between the presence and absence of abnormal Doppler/MRI findings. This pilot prognostic model showed a relatively high accuracy of >80%; therefore, the predictors may serve as biomarkers for SCVD in subjects at age at risk (51-65 years).