Abstract
Background/Objective:
Vascular risk factors and neurovascular dysfunction may be closely related to cognitive impairment and dementia. In this study, we evaluated the association between hemodynamic markers and longitudinal cognitive changes in patients with mild cognitive impairment (MCI). Furthermore, we investigated whether hemodynamic markers could predict the risk of progression to Alzheimer’s disease (AD) in patients with MCI.
Methods:
A total of 68 subjects with amnestic MCI were recruited. Using transcranial Doppler (TCD) ultrasonography, cerebrovascular reactivity was evaluated with a breath-holding test (breath holding index; BHI) in addition to the mean flow velocity (MFV) and pulsatility index (PI) of the middle cerebral artery. We followed subjects for 24 months and each subject underwent neuropsychological testing and TCD ultrasonography, annually. According to the follow-up neuropsychological studies and clinical interviews at 12 months, we divided the patients with MCI into two groups: patients with stable cognitive performance and patients who progressed to AD.
Results:
Lower BHI and higher PI were observed in patients who progressed to AD. The changes of MMSE score over the first 12 months correlated with lower baseline MMSE score and changes of MFV and BHI. The changes of MMSE score over 24 months were closely related to higher baseline resistance index and PI values. Multivariate logistic regression showed that abnormal baseline BHI value could predict a conversion from MCI to AD.
Conclusions:
We confirmed there is a close association between hemodynamic changes represented by TCD markers and cognitive decline, supporting the clinical value of hemodynamic markers in predicting MCI patients who will progress to AD.
INTRODUCTION
Recent autopsy studies have reported that vascular pathologies have an important role in the manifestation of Alzheimer’s disease (AD) [1]. Furthermore, systematic reviews and meta-analyses have highlighted the importance of vascular risk factors on AD onset and progression [2–4]. Therefore, understanding the contribution of cerebrovascular dysfunction on the progression of cognitive decline in AD or mild cognitive impairment (MCI) is an important step in the development of new preventive therapies aimed at slowing the progression of the disease in patients with a risk of MCI as well as in already demented patients.
Transcranial Doppler (TCD) ultrasonography is a non-invasive imaging technique widely used for the investigation of cerebrovascular hemodynamics in the major cerebral arteries. Previously, we reported that AD patients had decreased mean flow velocity (MFV), increased pulsatility index (PI), and decreased cerebrovascular reactivity (CVR), compared to control subjects, similar to findings from previous studies [2, 5–7]. In addition, we found that Mini-Mental State Examination (MMSE) scores are closely related to PI and CVR, suggesting that there might be an association between impaired cerebral microvessel function and cognitive impairment [8]. However, these findings were from a cross-sectional study, so it is unclear whether the decrease in microvascular function reflects diminished demand caused by advanced neurodegeneration or whether cerebral small-vessel disease precedes and contributes to dementia and neurodegeneration.
In this study, we evaluated the association between hemodynamic markers and longitudinal cognitive changes in patients with MCI. Furthermore, we investigated whether hemodynamic markers could predict the risk of progression to AD in patients with MCI.
METHODS
Participants
Between May 2011 and December 2012, we consecutively enrolled patients who visited the Department of Neurology at Bucheon St. Mary’s Hospital. A total of 68 amnestic MCI patients who fulfilled the clinical diagnostic criteria by Petersen et al. [9], for MCI were recruited. APOE genotype was also determined. Height and weight were measured, and body mass index (BMI) was calculated. Vascular risk factor history was also obtained (e.g., hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke, and smoking). Global cognitive function was assessed using the Korean version of the Mini-Mental State Examination (K-MMSE) [10]. Patient functional performance was assessed with Clinical Dementia Rating (CDR) and CDR Sum of Boxes (CDR-SOB) scores. Patients with neurological or psychiatric illnesses such as schizophrenia, epilepsy, and encephalitis were excluded. Patients with physical illnesses that could interfere with the clinical study, such as hearing or vision loss, aphasia, malignancy, and hepatic or renal disorders were excluded. Blood tests for excluding medical diseases included a complete blood count, blood chemistry tests, vitamin B12/folate, syphilis serology, and thyroid function tests.
All participants underwent 3.0-T brain magnetic resonance imaging (MRI) (Intera; Philips Medical Systems, Best, The Netherlands) including fluid-attenuated inversion recovery imaging and T1/T2-weighted imaging. The slice thickness was 5 mm without an interslice gap. The three-dimensional time-of-flight method was used as the imaging protocol for MR angiography. Individuals with brain lesions that were related to cognition such as large artery infarctions or multiple lacunes, brain tumors, and normal pressure hydrocephalus were also excluded. Carotid artery stenosis was defined according to the North American Symptomatic Carotid Endarterectomy Trial method [11], and patients with carotid artery stenosis with greater than 50% lumen diameter reduction were excluded.
Standard protocol approvals, registrations, and patient consents
This study was approved by the Institutional Review Board of The Catholic University of Korea, Bucheon St. Mary’s Hospital and all participants provided informed consent.
TCD measurement
All patients underwent evaluation using TCD ultrasonography. TCD examination was performed using a 2-MHz Doppler probe (Viasys Healthcare, Model Sonara) through the temporal bone window, by a sonographer who was blinded to the clinical diagnosis. The participants were placed in the supine position, and the TCD probe was fixed on the temporal window. The proximal segments of the middle cerebral artery at depths of 55–65 mm were examined bilaterally, and the MFV was evaluated. The changes in the middle cerebral artery flow velocity have been found to be reliably correlated with changes with cerebral blood flow [12]. In addition, PI and RI (Resistance index) were calculated. PI was calculated by subtracting the end-diastolic velocity from the peak systolic velocity and then dividing by the MFV. Thus, the PI is analogous to pulse pressure and is recognized as a measure of distal flow resistance and vascular wall rigidity. The RI is a measure of peripheral flow resistance and was calculated by subtracting the end-diastolic velocity from the peak systolic velocity and then dividing by the peak systolic velocity.
After baseline assessments, CVR was evaluated using a breath-holding (BH) test. Participants were requested to hold their breath for at least 30 seconds to reach a maximal increase in flow velocity, and the MFV during the BH task (hypercapnia condition) was recorded. All TCD data were recorded for offline analysis. Measurements were repeated under basal conditions at the maximum increase in flow velocity during hypercapnia. CVR was calculated as a percentage of the baseline MFV and absolute changes were determined by subtracting the baseline values from the maximum MFV in the middle cerebral artery during the BH task, which was defined as the breath holding index (BHI) in this study.
Because there was a large dispersion of subjects in the TCD parameters, we categorized baseline TCD parameter as a single, binary variable, assigning the patient a 0 value to normal value, and 1 in the case of abnormal values. We defined abnormal, in the case of MFV, as 45 cm/s or lower, the PI as 1.2 or higher, the RI as 0.8 or higher, and the BHI as 0.69 or lower. These parameters were demonstrated to have a predictive value for cerebrovascular disease and cognitive decline in previous studies [13, 14].
Follow up
Every 12 months, each subject underwent a clinical evaluation and the same neuropsychological tests, including the K-MMSE, CDR, and CDR-SOB. All participants also underwent follow-up evaluations for hemodynamic markers including MFV, PI, RI, and BHI using TCD ultrasonography. A total of 7 patients with MCI were followed just for 12 months and 61 patients for 24 months. Of these patients followed up for 24 months, 44 patients underwent TCD measurements also. We calculated changes in the hemodynamic parameter values by subtracting the TCD parameter values at follow-up from the baseline values. There were no significant differences in baseline characteristics and hemodynamic markers between those who followed up during 12 months or 24 months. According to follow-up neuropsychological studies and clinical interviews at 12 months, we divided the patients with MCI into two groups: patients with stable cognitive performance (group 1) and patients who progressed to AD (group 2). Diagnosis of AD was made according to the criteria of the National Institute of Neurologic and Communicative Disorders and Stroke and the Alzheimer Disease and Related Disorders Association criteria for probable AD [15].
Statistical analysis
To compare the baseline demographic characteristics between patients who remained stable and patients who progressed to AD, we first conducted Kolmogorov-Smirnov tests to figure out the pattern distribution of the data and used t-tests and Chi-Square tests for the data showing the pattern of a normal distribution. When it comes to non-normal distribution of the data, the Mann-Whitney test was used for continuous variables and Fisher’s exact test was used for categorical variables.
Analysis of covariance (ANCOVA) was used to compare baseline TCD parameters including MFV, PI, RI, and BHI values between the two groups adjusting for age and baseline MMSE scores. The distribution of the dependent variables with a baseline PI and BHI was skewed, so a log-transformation (natural log) was used to normalize its distribution. To evaluate the association of MMSE score changes during 12 and 24 months and hemodynamic markers, we performed a Spearman rank correlation analysis. To determine the risk factors that could predict progression to AD in patients with MCI, we performed multivariate logistic regression models using age, baseline MMSE, CDR-SOB, and baseline TCD markers. Factors that showed significant differences between patients with and patients without progression to dementia were selected as candidate predictors (age, baseline K-MMSE scores, CDR-SOB scores, baseline MFV, PI, RI, and BHI) in the multivariate logistic regression model. The backwards stepwise logistic regression (p = 0.1) assisted model selection. All statistical analyses were performed using SPSS version 18 (SPSS, Chicago, IL, USA), and a p < 0.05 was considered significant.
Results
Characteristic of the participants
A total of 68 patients were enrolled in this study. During the 12 months of follow-up, 48 (70.58%) patients had stable cognitive performance (group 1), while 20 (29.41%) progressed from MCI to AD (group 2).
Table 1 presents basic demographics and neuropsychological test results. The mean age was 71.27±8.69 years in group 1 and 72.45±7.50 years in group 2. There were no statistically significant differences in gender, education years, and vascular risk factors including hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke history, and smoking between the two patient groups. Among the statistically significant differences the baseline MMSE score was lower [group 1:24.06±4.64, group 2:20.55±4.33, <0.005] and the CDR-SOB score [group 1:1.54±0.94, group 2:2.29±1.22, p = 0.009] was higher in MCI patients who converted to AD. The means of the baseline TCD parameters are shown in Table 2. The mean PI values were significantly higher in patients who progressed to AD [group 1:1.33±0.75, group 2:1.43±0.12, p = 0.037], and the mean RI values were also higher in patients who progressed to AD [group 1:0.67±0.01, group 2:0.70±0.11:0.011]. The mean BHI values were significantly lower in patients who progressed to AD [group 1:0.57±0.24, group 2:0.50±0.04, p = 0.040]. The mean MFV values were similar between the two groups.
Baseline demographics and neuropsychological test results of MCI participants
Group 1, patients who had stable cognitive function; Group 2, patients who progressed to AD; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; CDR, Clinical Dementia Rating; SOB, Sum of Boxes; GDepS, Geriatric Depression Scale; BMI, body mass index; HTN, hypertension; DM, diabetes mellitus; IHD, ischemic heart disease. Values are presented as the mean±standard deviation or number. Statistical analysis was performed by using Chi-square tests or Fisher’s exact test for categorical variables and t-tests or Mann-Whitney test for continuous variables.
Baseline TCD parameters in MCI participants
Group 1, patients who had stable cognitive function; Group 2, patients who progressed to AD; MCI, mild cognitive impairment; Group 1, patients who had stable cognitive performance; Group 2, patients who progressed to Alzheimer’s dementia; MFV, mean flow velocity; PI, pulsatility index; RI, resistance index; BHI, breathing holding index. Values are presented as the mean±standard deviation. TCD parameters (baseline PI, BHI) were subject to logarithmic transformation before proceeding to ANOVA with adjustments made for the covariates of age and baseline MMSE score.
Table 3 shows the association between MMSE score changes and baseline TCD parameters including MFV, PI, RI, and BHI. The changes of MMSE scores over the first 12 months correlated with lower baseline MMSE scores [β: – 0.263, p = 0.018], and correlated with changes of MFV [β: 0.217, p = 0.042] and BHI [β: 0.235, p = 0.031]. The changes of MMSE scores over 24 months were closely related to higher baseline RI [β: 0.237, p = 0.033] and PI [β: 0.221, p = 0.043] values.
The association of MMSE score changes with hemodynamic markers in MCI participants
MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; CDR, Clinical Dementia Rating; SOB, Sum of Boxes; MFV, mean flow velocity; BHI, breathing holding index; PI, pulsatility index, Δ, changes of value; ρ, Spearman’s rank correlation coefficient. Statistical analysis was performed by the Spearman rank correlation analysis.
We additionally evaluated risk factors that would predict patients with MCI who progressed to AD. Multivariate logistic regression after adjustment for age, baseline MMSE, and CDR-SOB scores showed that abnormal baseline abnormal BHI value could predict a conversion from MCI to AD. Risk, defined in odds ratios, of cognitive deterioration in MCI patients with an abnormal BHI value was estimated to be 4.29 times [95% CI 1.038–17.729, p = 0.044] higher than in those with a normal BHI. Abnormal baseline PI values showed a risk of conversion to dementia of 3.824 times higher if compared with a normal PI value [95% CI 0.993–14.720, p = 0.051], although this difference was statistically insignificant (Table 4).
Multivariate logistic regression model considering the evolution from amnestic mild cognitive impairment to Alzheimer’s dementia at 12 months
MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; CDR, Clinical Dementia Rating; SOB, Sum of Boxes; MFV, mean flow velocity; PI, pulsatility index; BHI, breathing holding index. Multivariate logistic regression model was constructed using backwards stepwise selection (with entry criteria of p≤0.1 and stay criteria of p≤0.05).
Discussion
In this study, we found that MCI patients who progressed to AD showed higher baseline PI and RI values and lower baseline BHI values compared to MCI patients who remained stable without progression to AD. And pathologic baseline PI and BHI values were associated with higher risk of developing dementia in MCI patients. Furthermore, the cognitive decline was closely associated with hemodynamic markers also.
PI reflects the resistance of the microvascular bed distal to the site of measurement, and is considered a popular index of microangiopathic changes in cerebral blood vessels [16]. CVR, defined as BHI in this study, is a cerebrovascular autoregulation parameter describing the ability of cerebral arterioles to vasodilate when cerebral perfusion pressure is low [17]. Stefani et al. suggested that increased PI and impaired CVR could result in an insufficient blood supply to the brain which promotes selective brain capillary degeneration and a neuron-glia energy crisis, that contribute to the pathobiology of AD [18, 19]. Neuroimaging studies with PET and SPECT also showed that reduced vasoreactivity was associated with increased cognitive impairment in dementia patients [20, 21]. Our results are in agreement with these previous findings, and we hypothesize that the patients with increased vascular resistance and impaired cerebrovascular reactivity were vulnerable to the progression of AD pathology due to insufficient blood supply to the brain. Many previous studies showed that MFV was decreased in AD patients. In this study, the baseline MFV value did not differ depending on the progression in MCI patients, but the MCI patients who were converted to AD showed greater decrease in MFV values. This means that MFV is more likely to be seen in a state of hypoperfusion due to increased vascular resistance and impaired cerebrovascular reactivity.
Multiple studies have revealed a strong association between vascular pathology and AD, but the details of the mechanisms by which the altered hemodynamics produces cognitive dysfunction have yet to be clearly elucidated. Several studies showed that impairment of cerebral autoregulation was closely associated with Aβ deposition and progression in AD [22–24]. The most plausible hypothesis for the mechanism underlying this association is that cerebrovascular pathology could aggravate amyloid and tau pathologies in patients with pre-existing genetic and lifestyle-related risk factors for AD [25]. de la Torre et al. reported that critical cerebral hypoperfusion could have a pathogenic role in AD due to progressive circulatory insufficiency that can destabilize neurons, synapses, and neurotransmission leading to a neurodegenerative process characterized by the formation of Aβ senile plaques, neurofibrillary tangles, and amyloid angiopathy [7, 27]. On the other hands, Iadecola et al. suggested that Aβ could compromise the ability of endothelial cells to produce vasodilatory factors, impairing cerebral auto-regulatory mechanisms that maintain an adequate blood flow during hypotension [28, 29]. Zhang X. also suggested a direct link between brain hypoperfusion and amyloidosis by showing that sustained cerebral hypoperfusion elicited hypoxia-inducible factor-1a secretion, which binds to the beta secretase cleavage enzyme-1 promotor resulting in increased Aβ production [30].
While there have been several studies that showed hemodynamic changes such as increased vascular resistance and decreased cerebral vasoreactivity in AD patients [5, 31], it is unclear whether these results are secondary to the decreased metabolic demand in AD patients, or whether such hemodynamic changes directly affect cognitive decline in AD patients. From our longitudinal study, the MCI patients with baseline increased vascular resistance and decreased cerebrovascular reactivity were more likely to progress to dementia and the hemodynamic changes were closely correlated with the changes of cognitive function. These results suggest that the hemodynamic changes were not only the consequences of neuronal loss, but also preceding factors to influence the progression to AD, at least in MCI, although this is uncertain for the cognitively normal [4].
Our study has several limitations. First, the number of patients enrolled in this study was relatively small and follow-up duration was too short to see the conversion from MCI to AD. There might be patients who will progress to dementia during the long-term follow-up. However, we think that it is meaningful, that our study showed, that hemodynamic markers could identify high-risk patients who will progress to dementia even within a relatively short period of time of two years. We saw that hemodynamic changes have already preceded the progression to dementia for our patients in two years. Further studies with larger sample sizes and long-term follow-up can clarify and extend these results. Second, we have inconsistent results according to the association of cognitive decline and all hemodynamic markers we tested at 12 months and 24 months. We hypothesized that baseline cognitive function and changes in blood flow velocity are closely related to changes in cognitive function in the early stages in MCI patients. Over time, however, their relevance deteriorated, arterial stiffness such as PI and RI might be more closely related to changes in cognitive function in MCI patients. Third, MCI patients who progressed to dementia had lower baseline MMSE scores and higher baseline CDR SOB scores, so we may have included patients with early dementia whose daily living capacity had already begun to deteriorate.
In conclusion, our data suggest that there is a close association between hemodynamic changes represented by TCD markers and cognitive decline, supporting the clinical value of hemodynamic markers for predicting progression to AD in MCI patients. Therefore, evaluation of hemodynamic parameters in MCI patients could be a worthwhile clinical recommendation. A long-term follow-up study with larger sample sizes would be required to strengthen our results and recommendations.
DISCLOSURE STATEMENT
Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/18-0026r4).