695
Journal of Alzheimer’s Disease 44 (2015) 695–704 DOI 10.3233/JAD-141770 IOS Press
Association of Vascular Factors and Amnestic Mild Cognitive Impairment: A Comprehensive Approach Ignacio Casado Naranjoa,∗ , Juan Carlos Portilla Cuencaa , Beatriz Duque de San Juana , Alfonso Falc´on Garc´ıaa , Ra´ul Romero Sevillaa , Ana Serrano Cabreraa , Carmen C´amara Hij´onb , Silvia Romero Chalab , Jos´e Manuel Fuentesc and Jos´e Mar´ıa Ram´ırez Morenod a Department
of Neurology, Hospital San Pedro de Alc´antara, C´aceres, Spain of Immunology, Hospital San Pedro de Alc´antara, C´aceres, Spain c CIBERNED. Department of Biochemistry and Molecular Biology and Genetics, Universidad de Extremadura, C´aceres, Spain d Department of Neurology, Hospital Universitario Infanta Cristina, Badajoz, Spain b Department
Accepted 22 September 2014
Abstract. Background and objective: Current evidence shows that numerous classic vascular risk factors (VRF) contribute to mild cognitive impairment (MCI), but the effects of emerging VRFs are less well-known. Using a comprehensive approach, we assessed the frequency and strength of association between MCI and classic VRFs, subclinical markers of atherosclerosis (cystatin C, lipoprotein(a), high-sensitivity C-reactive protein, and intima-media thickness) and white matter hyperintensities (WMH). Methods: In this case-control study of consecutive MCI patients and cognitively normal controls, subjects underwent clinical and neuropsychological examinations, laboratory analyses, a carotid duplex scan, and a brain magnetic resonance imaging scan. Results: The study included 105 patients with amnestic MCI (aMCI): 24 with single domain amnestic MCI, 81 with multiple domain amnestic MCI, and 76 controls. Compared to controls, patients with aMCI were significantly older and had higher rates of arterial hypertension, atrial fibrillation, and depression. They also had a larger intima-media thickness and higher load of WMHs, both periventricular (WMHpv) and subcortical (WMHsc). In the adjusted analysis, all variables except WMHsc displayed a significant association with aMCI. Body mass index exerted a protective effect. Conclusions: Our findings suggest a direct association between aMCI and age, hypertension, atrial fibrillation depression, intima-media thickness, and WMHpv. Body mass index has a protective effect on this MCI subtype. Keywords: Atherosclerosis, carotid intima-media thickness, mild cognitive impairment, vascular risk factors, white matter hyperintensities
INTRODUCTION Mild cognitive impairment (MCI) is a state characterized by cognitive decline that is more pronounced than would be expected for an individual’s age and ∗ Correspondence to: Ignacio Casado Naranjo, MD, Department of Neurology, Hospital San Pedro de Alc´antara, Avda Pablo Naranjo s/n, 10001 C´aceres, Spain. Tel.: +34 92621532; Fax: +34 927256202; E-mail: [email protected].
education level, but which creates little to no interference with activities of daily life [1]. According to this definition, MCI has neither a specific outcome nor a specific etiology [2]. MCI should be regarded as a syndrome whose clinical presentation is heterogeneous; its different clinical subtypes correspond to whether the cognitive deficit is single domain amnestic (sd-aMCI), multiple domain amnestic (md-aMCI), single domain non-amnestic
ISSN 1387-2877/15/$27.50 © 2015 – IOS Press and the authors. All rights reserved
696
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
(sd-naMCI), or multiple domain non-amnestic (mdnaMCI). These profiles may be related to different pathophysiological processes [3, 4]. Since subjects with MCI may be at risk for developing dementia, the syndrome is considered a public health problem that affects 15% of all patients aged 65 and older [5]. There is renewed interest in the effects of risk factors and vascular pathology on cognitive impairment and dementia [6]. Since controlling and modifying these effects is now possible, it would also theoretically be possible to reduce the impact of those effects on cognition [7]. Although the evidence showing an association between classic vascular risk factors (VRFs) and the spectrum of cognitive impairment ranging from MCI to dementia is considered robust [8, 9], the association between other emerging VRFs and cognitive impairment requires further study [2]. Using a comprehensive approach, we assessed the frequency and strength of association between MCI and classic VRFs, subclinical markers of atherosclerosis [cystatin C, lipoprotein(a) (LpA), high-sensitivity C-reactive protein (hs-CRP), and intima-media thickness (IMT)] and white matter hyperintensities (WMH). METHODS Study design and subjects Case-control study. We selected consecutive patients aged ≥ 65 years who had been evaluated in a memory clinic and diagnosed with an MCI syndrome [10]. Controls with no cognitive complaints and testing as cognitively normal (Mini-Mental State Examination [11], MMSE ≥ 27) were selected simultaneously. We excluded both patients and controls with a history of transient ischemic attack or stroke, ischemic heart disease or peripheral artery disease, moderate or severe head trauma, other neurologic conditions associated with cognitive impairment, history of alcohol or substance abuse, malignant illness, inflammatory or autoimmune disease, major depression (score ≥ 10 Geriatric Depression Scale, short version) [12], and those fitted with a pacemaker. Lastly, after diagnosis of MCI subtypes we excluded patients with non-amnestic MCI (na-MCI) in order to homogenize the sample. All subjects underwent a complete clinical evaluation, including neuropsychological and functional screening, carotid duplex-doppler ultrasonography, and cerebral magnetic resonance image (MRI).
All subjects signed an informed consent form and the study protocol was approved by our hospital’s ethics committee. Study evaluations Demographic data: age, gender, educational level classified as no studies to primary studies (≤ 6 years) or as secondary to university studies (> 6 years). Anthropometric data: weight (kg), height (cm), body mass index (BMI, kg/m2 ), and waist circumference. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) was measured twice in both arms using a certified digital sphygmomanometer (EDAN M8) after the patient had been resting in a seated position for 5 minutes and recorded the mean values from both measurements. Patient’s physical exercise routine was categorized as ≤2 days weekly or >2 days weekly. Vascular risk factors definitions and treatment Hypertension (HT): SBP ≥ 140 mm Hg and/or DBP ≥ 90 mm Hg and/or antihypertensive treatment); diabetes: clinical history, fasting blood glucose level > 126 mg/dL, or antidiabetic treatment; hyperlipidemia: total plasma cholesterol ≥ 200 mg/dL or treatment with hypolipidemic agents); atrial fibrillation (AF): clinical history or detection; sleep apnea: clinical history; tobacco use: active smoker versus non-smoker or smoke-free for > 6 months; alcohol consumption: > 2 drinks per day; and current medications: antihypertensive, hypolipidemic, or antidepressant agents. Laboratory test Fasting blood samples were obtained to perform a complete blood count and check metabolic parameters using standard kits (Roche Lab). LpA, cystatin C, and hs-CRP were measured using nephelometry (Siemens BN II unit). Only those subjects with MCI underwent APOE genotyping by means of genomic hybridization (Alzheimer StripAssay, ViennaLab). Neuropsychological and functional evaluation Neuropsychological and functional evaluation included performing the MMSE [11] to test each subject’s general cognitive state (this test has been adapted and validated to Spanish population [13]). A certified neuropsychologist who was blinded to clinical data performed a neuropsychological evaluation with specific tests for the following cognitive domains: episodic verbal memory (Free and Cued Selective Reminding Test [14]); attention and executive functions (Symbol
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
Digit Modalities Test [15]; language (letter fluency: words containing P or excluding E; animal fluency); and visual-spatial ability and constructional praxis (Rey-Osterrieth complex figure (ROCF) copy and recall exercises [16]). Test administration details and normative data for the Spanish population, adjusted by age and educational level, are available for each test [17–19]. Depression was assessed using the 15-item Geriatric Depression Scale [12]. Functional capacity was measured using the Functional Activities Questionnaire [20].
697
6 mm for all sequences. A neuroradiologist blinded to each subject’s clinical data checked for periventricular and subcortical WMH (WMHpv, WMHsc) and recorded results using the Age-Related White Matter Changes (ARWMC) scale [22]. To facilitate data analysis, scores of 2 and 3 on the ARWMC scale were grouped in a single category. Statistical analysis
We used a Philips ultrasound unit model HD 11 with a high-resolution L12-3 transducer for vascular studies with a frequency range of 3 to 12 MHz. Studies examined both carotid arteries in longitudinal and transverse sections; selected images were saved to a magnetic storage device. IMT of the common carotid arteries over ≈1.5 cm proximal to the flow divider was measured according to standard recommendations [21], by a single neurologist certified in neurosonology who was unaware of each subject’s clinical characteristics.
Quantitative variables were expressed as mean and standard deviation or median and interquartile range depending on whether they showed a normal distribution. Categorical variables were given as absolute and relative frequencies. The association between qualitative variables was studied using the chi-square test, and the t-test was used for quantitative variables. Prognostic factors associated with MCI were identified by completing two different multivariate binary logistic regression analyses: one examining risk factors and the other examining results from IMT and neuroimaging studies. The independent variables included were those that showed an association in the bivariate analysis with a significance level of less than 0.20. The variables remaining in the model were those that continued to show an association with the same significance level (P < 0.20) or those with a high level of clinical significance. We used the ‘enter’ method to adjust the model. ROC curve was used to measure model discrimination and calibration was evaluated with the Hosmer-Lemeshow test. Backward stepwise multiple linear regression analysis was used to identify predictors of results on cognitive tests examining the different domains under study. The dependent variables included were the scores on cognitive tests employed to evaluate each domain. Independent or predictor variables were as follows: age, medical history (HT, hypercholesterolemia, diabetes, AF, smoking, BMI, BP), history of psychopathology (depression), and structural findings (IMT and neuroimaging). A p-value of 0.05 was considered statistically significant for all analyses. SPSS® version 15 was used for data analysis.
Cerebral MRI study
RESULTS
The scan was completed using a 1.0 Tesla system (Philips Gyroscan T10-NT) with a head coil. We recorded the following sequences: T2 and PD (TR 2459 ms, TE 20 or 90 ms); sagittal T2 (TR 4326 ms, TE 130 ms); coronal T1 (TR 539, TE 15) and coronal FLAIR (TR 11 000; TE 140). Slice thickness was
The study included 105 patients with MCI and 76 controls. The MCI group contained 24 individuals with sd-aMCI, and 81 with md-aMCI. The characteristics of the sample are shown in Table 1. Compared to controls, patients with aMCI were significantly older [77.5 (4.6) versus 75.3 (5.2), p = 0.003] and had higher SBP
Diagnosis of MCI subtypes MCI patients were classified into the following subtypes on the basis of the performance on test in the cognitive domain cited previously: sd-aMCI, if there was impairment in memory alone, md-aMCI, if there was impairment in memory and at least another cognitive domain, and na-MCI if the subjects had impairment in one or more non-memory domains. Impairment was defined by the presence of a test scoring 1.5 standard deviations below the average score of healthy control subjects after correction for age and education. Post-hoc analyses showed no statistically significant differences with regard to demographics, clinical, analytical, IMT measures or MRI characteristics between subjects with single and multiple domain amnestic MCI, and therefore we analyzed the data of subjects as a single group called amnestic MCI (aMCI). IMT measurement
698
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
readings [150 versus 135, p = 0.022]. They also showed a higher prevalence of HT (81.7% versus 68.5%, p = 0.042), AF (12.4% versus 1.3%, p = 0.006), and depression (18.1% versus 3.9%, p = 0.009). However, both weight [70.3 (11.0) versus 76.1 (11.4), p = 0.001 and BMI [28 (4) versus 30 (5), p = 0.021] were significantly lower in the patient group. Neurological and functional evaluation Patients with aMCI had significantly lower scores on the MMSE and lower test scores for all examined cognitive domains. Conversely Geriatric Depression Scale and Functional Activities Questionnaire scores were higher (Fig. 1). Metabolic variables, atherosclerosis markers, and WMH on brain MRI There were no differences between patients with aMCI and controls for either metabolic parameters or plasma atherosclerotic markers. However, the aMCI group presented a significantly larger IMT and an increased WMH load (Table 2). Predictors of aMCI Results from the unadjusted and adjusted logistic regression analysis are shown in Table 3. The following variables showed a significant association
with aMCI in the unadjusted analysis: age [OR 1.099, CI 95% (1.032–1.172) p = 0.003], HT [OR 2.354, CI 95% (1.185–4.677) p = 0.05], AF [OR 10.598, CI 95% (1.355–82.875) p = 0.002], depression [OR 5.376, CI 95% (1.530–18.895), p = 0.002], IMT [OR 32.041, CI 95% (6.700–153.218) p < 0.0001], WMHpv intensity, grade 1 versus grade 0 (reference) [OR 4.200, CI 95% (1.796–9.820), p = 0.001] and grade 2 versus grade 0 [OR 16.450, CI 95% (5.810–46.578), p < 0.0001], WMHsc intensity, grade 1 versus grade 0 [OR 3.551, CI 95% (1.717–7.345), p = 0.001] and grade 2 versus grade 0 [OR 6.539, CI 95% (2.443–17.502), p < 0.001]. In the adjusted analysis, the following variables maintained the association with aMCI: age [OR 1.089, CI 95% (1.018–1.166) p = 0.009], HT [OR 2.383, CI 95% (1.036–5.479) p = 0.039], AF [OR 8.762, CI 95% (1.068–71.916) p = 0.010], depression [OR 7.611, CI 95% (1.962–29.525) p = 0.002], IMT [OR 19.356, CI 95% (3.106–120.630) p = 0.002], WMHpv intensity, grade 1 versus grade 0 (reference) [OR 3.108, CI 95% (1.225–7.883) p = 0.017] and grade 2 versus grade 0 [OR 6.942, CI 95% (1.950–24.716), p = 0.003]. WMHsc intensity exhibited a decrease in significance. BMI displayed a protective association in the unadjusted model [OR 0.923, CI 95% (0.861–0.989) p = 0.021] and the adjusted model [OR 0.894, CI 95% (0.822–0.972) p = 0.007].
Table 1 Characteristics of study sample aMCI (n = 105) Demographics Age, y Male gender, n (%) Lower educational level (≤6), n (%) Physical activity (≥2 times/week), n (%) APOE 4, n (%) Clinical Weight, kg Body mass index Waist circumference, cm Systolic blood pressure, mm Hg Hypertension, n (%) Hyperlipidemia, n (%) Diabetes mellitus, n (%) Atrial fibrillation, n (%) Sleep apnea, n (%) Current smoker Current drinker Depression, n (%) Antihypertensives, n (%) Statins, n (%) Antidepressants, n (%) ∗ Not
available
Control (n = 76)
p
77.5 (4.6) 48 (45.7) 58 (55.2) 46 (43.8) 34 (32.3)
75.3 (5.2) 35 (41.6) 42 (55.3) 39 (51.3) NA∗
0.003 0.964 0.997 0.318
70.3 (10.9) 28.4 (4.0) 100 (10) 151 (20) 86 (81.9) 52 (49.5) 25 (23.8) 13 (12.4) 6 (5.7) 3 (2.9) 22 (21) 19 (18.1) 80 (76.2) 51 (48.6) 30 (28.6)
76.1 (11.4) 29.9 (4.7) 102 (9) 144 (18) 52 (69.3) 32 (42.1) 23 (30.3) 1 (1.3) 3 (3.9) 1 (1.3) 19 (25) 3 (4) 46 (60.5) 32 (42.1) 6 (7.9)
0.001 0.021 0.153 0.022 0.049 0.323 0.332 0.006 0.436 0.486 0.521 0.009 0.024 0.389 0.001
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
699
Fig. 1. Neuropsychological test scores. Neuropsychological test scores are corrected for age and educational level. MMSE, Mini-Mental State Examination; FCSRT-IR, Free and Cued Selective Reminding Test-Immediate Recall; FCSRT-DR, Free and Cued Selective Reminding Test-Delayed Recall; SDMT, Symbol Digit Modalities Test; ROCF, Rey-Osterrieth Complex Figure; GDS, Geriatric Depression Scale; FAQ, Functional Activities Questionnaire. Table 2 Metabolic or atherosclerotic markers and MRI findings
Glucose (mg/dL) HbA1c (%) Triglycerides (mg/dL) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) LpA (mg/dL) Cystatin C (mg/L) hsCRP (mg/L) Intima-media thickness (mm) White matter hyperintensities Periventricular 0, n (%) 1, n (%) 2, n (%) 3, n (%) Subcortical 0, n (%) 1, n (%) 2, n (%) 3, n (%) ∗ 68
aMCI (n = 105)
Control* (n = 76)
p
107.13 (25.60) 6.05 (0.91) 98.95 (62.83) 197.66 (34.53) 60.95 (16.67) 117.75 (31.11) 27.14 (24.55) 0.94 (0.31) 3.31 (3.82) 1.032 (0.27)
109.15 (25.61) 6.12 (0.72) 104.07 (39.21) 198.34 (28.30) 59.26 (13.45) 116.96 (27.57) 27.31 (22.39) 0.90 (0.20) 3.58 (6.19) 0.850 (0.22)
0.600 0.599 0.531 0.885 0.452 0.859 0.962 0.331 0.718
Journal of Alzheimer’s Disease 44 (2015) 695–704 DOI 10.3233/JAD-141770 IOS Press
Association of Vascular Factors and Amnestic Mild Cognitive Impairment: A Comprehensive Approach Ignacio Casado Naranjoa,∗ , Juan Carlos Portilla Cuencaa , Beatriz Duque de San Juana , Alfonso Falc´on Garc´ıaa , Ra´ul Romero Sevillaa , Ana Serrano Cabreraa , Carmen C´amara Hij´onb , Silvia Romero Chalab , Jos´e Manuel Fuentesc and Jos´e Mar´ıa Ram´ırez Morenod a Department
of Neurology, Hospital San Pedro de Alc´antara, C´aceres, Spain of Immunology, Hospital San Pedro de Alc´antara, C´aceres, Spain c CIBERNED. Department of Biochemistry and Molecular Biology and Genetics, Universidad de Extremadura, C´aceres, Spain d Department of Neurology, Hospital Universitario Infanta Cristina, Badajoz, Spain b Department
Accepted 22 September 2014
Abstract. Background and objective: Current evidence shows that numerous classic vascular risk factors (VRF) contribute to mild cognitive impairment (MCI), but the effects of emerging VRFs are less well-known. Using a comprehensive approach, we assessed the frequency and strength of association between MCI and classic VRFs, subclinical markers of atherosclerosis (cystatin C, lipoprotein(a), high-sensitivity C-reactive protein, and intima-media thickness) and white matter hyperintensities (WMH). Methods: In this case-control study of consecutive MCI patients and cognitively normal controls, subjects underwent clinical and neuropsychological examinations, laboratory analyses, a carotid duplex scan, and a brain magnetic resonance imaging scan. Results: The study included 105 patients with amnestic MCI (aMCI): 24 with single domain amnestic MCI, 81 with multiple domain amnestic MCI, and 76 controls. Compared to controls, patients with aMCI were significantly older and had higher rates of arterial hypertension, atrial fibrillation, and depression. They also had a larger intima-media thickness and higher load of WMHs, both periventricular (WMHpv) and subcortical (WMHsc). In the adjusted analysis, all variables except WMHsc displayed a significant association with aMCI. Body mass index exerted a protective effect. Conclusions: Our findings suggest a direct association between aMCI and age, hypertension, atrial fibrillation depression, intima-media thickness, and WMHpv. Body mass index has a protective effect on this MCI subtype. Keywords: Atherosclerosis, carotid intima-media thickness, mild cognitive impairment, vascular risk factors, white matter hyperintensities
INTRODUCTION Mild cognitive impairment (MCI) is a state characterized by cognitive decline that is more pronounced than would be expected for an individual’s age and ∗ Correspondence to: Ignacio Casado Naranjo, MD, Department of Neurology, Hospital San Pedro de Alc´antara, Avda Pablo Naranjo s/n, 10001 C´aceres, Spain. Tel.: +34 92621532; Fax: +34 927256202; E-mail: [email protected].
education level, but which creates little to no interference with activities of daily life [1]. According to this definition, MCI has neither a specific outcome nor a specific etiology [2]. MCI should be regarded as a syndrome whose clinical presentation is heterogeneous; its different clinical subtypes correspond to whether the cognitive deficit is single domain amnestic (sd-aMCI), multiple domain amnestic (md-aMCI), single domain non-amnestic
ISSN 1387-2877/15/$27.50 © 2015 – IOS Press and the authors. All rights reserved
696
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
(sd-naMCI), or multiple domain non-amnestic (mdnaMCI). These profiles may be related to different pathophysiological processes [3, 4]. Since subjects with MCI may be at risk for developing dementia, the syndrome is considered a public health problem that affects 15% of all patients aged 65 and older [5]. There is renewed interest in the effects of risk factors and vascular pathology on cognitive impairment and dementia [6]. Since controlling and modifying these effects is now possible, it would also theoretically be possible to reduce the impact of those effects on cognition [7]. Although the evidence showing an association between classic vascular risk factors (VRFs) and the spectrum of cognitive impairment ranging from MCI to dementia is considered robust [8, 9], the association between other emerging VRFs and cognitive impairment requires further study [2]. Using a comprehensive approach, we assessed the frequency and strength of association between MCI and classic VRFs, subclinical markers of atherosclerosis [cystatin C, lipoprotein(a) (LpA), high-sensitivity C-reactive protein (hs-CRP), and intima-media thickness (IMT)] and white matter hyperintensities (WMH). METHODS Study design and subjects Case-control study. We selected consecutive patients aged ≥ 65 years who had been evaluated in a memory clinic and diagnosed with an MCI syndrome [10]. Controls with no cognitive complaints and testing as cognitively normal (Mini-Mental State Examination [11], MMSE ≥ 27) were selected simultaneously. We excluded both patients and controls with a history of transient ischemic attack or stroke, ischemic heart disease or peripheral artery disease, moderate or severe head trauma, other neurologic conditions associated with cognitive impairment, history of alcohol or substance abuse, malignant illness, inflammatory or autoimmune disease, major depression (score ≥ 10 Geriatric Depression Scale, short version) [12], and those fitted with a pacemaker. Lastly, after diagnosis of MCI subtypes we excluded patients with non-amnestic MCI (na-MCI) in order to homogenize the sample. All subjects underwent a complete clinical evaluation, including neuropsychological and functional screening, carotid duplex-doppler ultrasonography, and cerebral magnetic resonance image (MRI).
All subjects signed an informed consent form and the study protocol was approved by our hospital’s ethics committee. Study evaluations Demographic data: age, gender, educational level classified as no studies to primary studies (≤ 6 years) or as secondary to university studies (> 6 years). Anthropometric data: weight (kg), height (cm), body mass index (BMI, kg/m2 ), and waist circumference. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) was measured twice in both arms using a certified digital sphygmomanometer (EDAN M8) after the patient had been resting in a seated position for 5 minutes and recorded the mean values from both measurements. Patient’s physical exercise routine was categorized as ≤2 days weekly or >2 days weekly. Vascular risk factors definitions and treatment Hypertension (HT): SBP ≥ 140 mm Hg and/or DBP ≥ 90 mm Hg and/or antihypertensive treatment); diabetes: clinical history, fasting blood glucose level > 126 mg/dL, or antidiabetic treatment; hyperlipidemia: total plasma cholesterol ≥ 200 mg/dL or treatment with hypolipidemic agents); atrial fibrillation (AF): clinical history or detection; sleep apnea: clinical history; tobacco use: active smoker versus non-smoker or smoke-free for > 6 months; alcohol consumption: > 2 drinks per day; and current medications: antihypertensive, hypolipidemic, or antidepressant agents. Laboratory test Fasting blood samples were obtained to perform a complete blood count and check metabolic parameters using standard kits (Roche Lab). LpA, cystatin C, and hs-CRP were measured using nephelometry (Siemens BN II unit). Only those subjects with MCI underwent APOE genotyping by means of genomic hybridization (Alzheimer StripAssay, ViennaLab). Neuropsychological and functional evaluation Neuropsychological and functional evaluation included performing the MMSE [11] to test each subject’s general cognitive state (this test has been adapted and validated to Spanish population [13]). A certified neuropsychologist who was blinded to clinical data performed a neuropsychological evaluation with specific tests for the following cognitive domains: episodic verbal memory (Free and Cued Selective Reminding Test [14]); attention and executive functions (Symbol
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
Digit Modalities Test [15]; language (letter fluency: words containing P or excluding E; animal fluency); and visual-spatial ability and constructional praxis (Rey-Osterrieth complex figure (ROCF) copy and recall exercises [16]). Test administration details and normative data for the Spanish population, adjusted by age and educational level, are available for each test [17–19]. Depression was assessed using the 15-item Geriatric Depression Scale [12]. Functional capacity was measured using the Functional Activities Questionnaire [20].
697
6 mm for all sequences. A neuroradiologist blinded to each subject’s clinical data checked for periventricular and subcortical WMH (WMHpv, WMHsc) and recorded results using the Age-Related White Matter Changes (ARWMC) scale [22]. To facilitate data analysis, scores of 2 and 3 on the ARWMC scale were grouped in a single category. Statistical analysis
We used a Philips ultrasound unit model HD 11 with a high-resolution L12-3 transducer for vascular studies with a frequency range of 3 to 12 MHz. Studies examined both carotid arteries in longitudinal and transverse sections; selected images were saved to a magnetic storage device. IMT of the common carotid arteries over ≈1.5 cm proximal to the flow divider was measured according to standard recommendations [21], by a single neurologist certified in neurosonology who was unaware of each subject’s clinical characteristics.
Quantitative variables were expressed as mean and standard deviation or median and interquartile range depending on whether they showed a normal distribution. Categorical variables were given as absolute and relative frequencies. The association between qualitative variables was studied using the chi-square test, and the t-test was used for quantitative variables. Prognostic factors associated with MCI were identified by completing two different multivariate binary logistic regression analyses: one examining risk factors and the other examining results from IMT and neuroimaging studies. The independent variables included were those that showed an association in the bivariate analysis with a significance level of less than 0.20. The variables remaining in the model were those that continued to show an association with the same significance level (P < 0.20) or those with a high level of clinical significance. We used the ‘enter’ method to adjust the model. ROC curve was used to measure model discrimination and calibration was evaluated with the Hosmer-Lemeshow test. Backward stepwise multiple linear regression analysis was used to identify predictors of results on cognitive tests examining the different domains under study. The dependent variables included were the scores on cognitive tests employed to evaluate each domain. Independent or predictor variables were as follows: age, medical history (HT, hypercholesterolemia, diabetes, AF, smoking, BMI, BP), history of psychopathology (depression), and structural findings (IMT and neuroimaging). A p-value of 0.05 was considered statistically significant for all analyses. SPSS® version 15 was used for data analysis.
Cerebral MRI study
RESULTS
The scan was completed using a 1.0 Tesla system (Philips Gyroscan T10-NT) with a head coil. We recorded the following sequences: T2 and PD (TR 2459 ms, TE 20 or 90 ms); sagittal T2 (TR 4326 ms, TE 130 ms); coronal T1 (TR 539, TE 15) and coronal FLAIR (TR 11 000; TE 140). Slice thickness was
The study included 105 patients with MCI and 76 controls. The MCI group contained 24 individuals with sd-aMCI, and 81 with md-aMCI. The characteristics of the sample are shown in Table 1. Compared to controls, patients with aMCI were significantly older [77.5 (4.6) versus 75.3 (5.2), p = 0.003] and had higher SBP
Diagnosis of MCI subtypes MCI patients were classified into the following subtypes on the basis of the performance on test in the cognitive domain cited previously: sd-aMCI, if there was impairment in memory alone, md-aMCI, if there was impairment in memory and at least another cognitive domain, and na-MCI if the subjects had impairment in one or more non-memory domains. Impairment was defined by the presence of a test scoring 1.5 standard deviations below the average score of healthy control subjects after correction for age and education. Post-hoc analyses showed no statistically significant differences with regard to demographics, clinical, analytical, IMT measures or MRI characteristics between subjects with single and multiple domain amnestic MCI, and therefore we analyzed the data of subjects as a single group called amnestic MCI (aMCI). IMT measurement
698
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
readings [150 versus 135, p = 0.022]. They also showed a higher prevalence of HT (81.7% versus 68.5%, p = 0.042), AF (12.4% versus 1.3%, p = 0.006), and depression (18.1% versus 3.9%, p = 0.009). However, both weight [70.3 (11.0) versus 76.1 (11.4), p = 0.001 and BMI [28 (4) versus 30 (5), p = 0.021] were significantly lower in the patient group. Neurological and functional evaluation Patients with aMCI had significantly lower scores on the MMSE and lower test scores for all examined cognitive domains. Conversely Geriatric Depression Scale and Functional Activities Questionnaire scores were higher (Fig. 1). Metabolic variables, atherosclerosis markers, and WMH on brain MRI There were no differences between patients with aMCI and controls for either metabolic parameters or plasma atherosclerotic markers. However, the aMCI group presented a significantly larger IMT and an increased WMH load (Table 2). Predictors of aMCI Results from the unadjusted and adjusted logistic regression analysis are shown in Table 3. The following variables showed a significant association
with aMCI in the unadjusted analysis: age [OR 1.099, CI 95% (1.032–1.172) p = 0.003], HT [OR 2.354, CI 95% (1.185–4.677) p = 0.05], AF [OR 10.598, CI 95% (1.355–82.875) p = 0.002], depression [OR 5.376, CI 95% (1.530–18.895), p = 0.002], IMT [OR 32.041, CI 95% (6.700–153.218) p < 0.0001], WMHpv intensity, grade 1 versus grade 0 (reference) [OR 4.200, CI 95% (1.796–9.820), p = 0.001] and grade 2 versus grade 0 [OR 16.450, CI 95% (5.810–46.578), p < 0.0001], WMHsc intensity, grade 1 versus grade 0 [OR 3.551, CI 95% (1.717–7.345), p = 0.001] and grade 2 versus grade 0 [OR 6.539, CI 95% (2.443–17.502), p < 0.001]. In the adjusted analysis, the following variables maintained the association with aMCI: age [OR 1.089, CI 95% (1.018–1.166) p = 0.009], HT [OR 2.383, CI 95% (1.036–5.479) p = 0.039], AF [OR 8.762, CI 95% (1.068–71.916) p = 0.010], depression [OR 7.611, CI 95% (1.962–29.525) p = 0.002], IMT [OR 19.356, CI 95% (3.106–120.630) p = 0.002], WMHpv intensity, grade 1 versus grade 0 (reference) [OR 3.108, CI 95% (1.225–7.883) p = 0.017] and grade 2 versus grade 0 [OR 6.942, CI 95% (1.950–24.716), p = 0.003]. WMHsc intensity exhibited a decrease in significance. BMI displayed a protective association in the unadjusted model [OR 0.923, CI 95% (0.861–0.989) p = 0.021] and the adjusted model [OR 0.894, CI 95% (0.822–0.972) p = 0.007].
Table 1 Characteristics of study sample aMCI (n = 105) Demographics Age, y Male gender, n (%) Lower educational level (≤6), n (%) Physical activity (≥2 times/week), n (%) APOE 4, n (%) Clinical Weight, kg Body mass index Waist circumference, cm Systolic blood pressure, mm Hg Hypertension, n (%) Hyperlipidemia, n (%) Diabetes mellitus, n (%) Atrial fibrillation, n (%) Sleep apnea, n (%) Current smoker Current drinker Depression, n (%) Antihypertensives, n (%) Statins, n (%) Antidepressants, n (%) ∗ Not
available
Control (n = 76)
p
77.5 (4.6) 48 (45.7) 58 (55.2) 46 (43.8) 34 (32.3)
75.3 (5.2) 35 (41.6) 42 (55.3) 39 (51.3) NA∗
0.003 0.964 0.997 0.318
70.3 (10.9) 28.4 (4.0) 100 (10) 151 (20) 86 (81.9) 52 (49.5) 25 (23.8) 13 (12.4) 6 (5.7) 3 (2.9) 22 (21) 19 (18.1) 80 (76.2) 51 (48.6) 30 (28.6)
76.1 (11.4) 29.9 (4.7) 102 (9) 144 (18) 52 (69.3) 32 (42.1) 23 (30.3) 1 (1.3) 3 (3.9) 1 (1.3) 19 (25) 3 (4) 46 (60.5) 32 (42.1) 6 (7.9)
0.001 0.021 0.153 0.022 0.049 0.323 0.332 0.006 0.436 0.486 0.521 0.009 0.024 0.389 0.001
I Casado-Naranjo et al. / Vascular Factors and Amnestic Mild Cognitive Impairment
699
Fig. 1. Neuropsychological test scores. Neuropsychological test scores are corrected for age and educational level. MMSE, Mini-Mental State Examination; FCSRT-IR, Free and Cued Selective Reminding Test-Immediate Recall; FCSRT-DR, Free and Cued Selective Reminding Test-Delayed Recall; SDMT, Symbol Digit Modalities Test; ROCF, Rey-Osterrieth Complex Figure; GDS, Geriatric Depression Scale; FAQ, Functional Activities Questionnaire. Table 2 Metabolic or atherosclerotic markers and MRI findings
Glucose (mg/dL) HbA1c (%) Triglycerides (mg/dL) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) LpA (mg/dL) Cystatin C (mg/L) hsCRP (mg/L) Intima-media thickness (mm) White matter hyperintensities Periventricular 0, n (%) 1, n (%) 2, n (%) 3, n (%) Subcortical 0, n (%) 1, n (%) 2, n (%) 3, n (%) ∗ 68
aMCI (n = 105)
Control* (n = 76)
p
107.13 (25.60) 6.05 (0.91) 98.95 (62.83) 197.66 (34.53) 60.95 (16.67) 117.75 (31.11) 27.14 (24.55) 0.94 (0.31) 3.31 (3.82) 1.032 (0.27)
109.15 (25.61) 6.12 (0.72) 104.07 (39.21) 198.34 (28.30) 59.26 (13.45) 116.96 (27.57) 27.31 (22.39) 0.90 (0.20) 3.58 (6.19) 0.850 (0.22)
0.600 0.599 0.531 0.885 0.452 0.859 0.962 0.331 0.718
