Cognitive functions and white matter lesions on magnetic resonance images in a sample of normal Iranian population with cardiovascular risk factors

Abstract

Purpose

Due to a suggestive three-way relationship between brain structural disorders, microvascular lesions, and cognitive impairments, we aimed to examine the association of the volume and number of white matter hyperintensity lesions and lacunar infarcts with cognitive impairment among patients with cardiovascular risk factors in a sample of the Iranian population.

Materials and methods

This study was conducted on a total number of 156 normal subjects aged 30–74 years with cardiovascular risk factors. We used the Framingham general cardiovascular risk factors prediction model to calculate the likelihood of each risk factor. The total number of lacunar infarcts and the volume of white matter hyperintensity lesion were calculated in brain magnetic resonance imaging. Cognition status was assessed using the Montreal Cognitive Assessment questionnaire.

Results

An adverse association was revealed between Montreal Cognitive Assessment score and different cardiovascular risk profiles including the Framingham body mass index score (p < 0.001) and the Framingham lipid score (p < 0.001). The total volume of white matter hyperintensity was negatively associated with total Montreal Cognitive Assessment cognition score (p < 0.001). Our study also showed an adverse association between total number of lacunar infarcts and total Montreal Cognitive Assessment cognition score (p = 0.038) and with some cognition components including memory (p = 0.013), attention (p = 0.037), abstraction (p = 0.046), and orientation (p = 0.002).

Conclusion

Periventricular lesions are associated with impaired memory, language, and visuoconstruction while subcortical lesions are associated with impairment in naming, attention, language, and abstraction functions in normal subjects with cardiovascular risk factors but without cardiovascular or cerebrovascular disorders.

Introduction

Vascular cognitive impairment is related to a range of ischemic and vascular disorders of the brain leading to cognitive impairment and dementia.¹ There is a significant drop in the incidence of vascular cognitive impairment over the past few years, probably because of successful treatment of cardiovascular risk factors and vascular factors.^2,3 The interesting fact that dementia shares many risk factors with coronary artery disorders, implies that these two entities might be connected in pathogenesis and mechanism.^4–6 Accordingly, several vascular factors such as transient ischemic attacks (TIAs), strokes, heart failure, and aortic stiffness contribute to the spectrum of vascular cognitive impairment ranging from mild cognitive dysfunction and early memory loss to dementia especially in the aging population.^7,8 The bidirectional connections between cardiovascular disorders and cognitive dysfunction are demonstrated by early executive brain dysfunction in those with congestive heart failure (CHF) and decreased cardiac output, and also by the fact that treatment of heart failure by cardiac transplantation, ventricular resynchronization therapy, and physical training is associated with reversed cognitive impairment.^9–11

The silent cerebrovascular lesions are classified as white matter hyperintensity (WMH) and the lacunar infarcts which are all diagnosed with magnetic resonance imaging (MRI).^12,13 Although these lesions are frequent in the elderly, their clinical significance has been the matter of research in the previous decade.^5,7,9,14 Several lines of evidence have demonstrated that the volume of these cerebrovascular lesions detected in MRI is associated with increased risk of ischemic strokes,¹⁵ renal dysfunction,¹⁶ gait imbalance and motor disturbance,^17,18 and cognitive dysfunction and dementia.^5,7,8,12,14 It has been well documented that microvascular lesions are a key reason for brain stroke. Moreover, these lesions are the major contributor to cognitive impairments that have major vascular components.^19–22 Therefore, there seems to be a three-way relationship between brain structural disorders, microvascular lesions, and cognitive impairments.

Cognitive dysfunction is multifactorial and age-related brain atrophy exacerbates the problem with neurological function, but identifying the etiological factors could help to build effective preventive strategies and also therapeutic modalities. Although the relation between the volume of these lesions and cognitive function has been determined, especially in elderly,^6,13,14 the clinical significance of these lesions in those with cardiovascular risk factors has yet to be identified. Thus, the aim of the current study was to determine the relationship between the volume of microvascular lesions and the cognitive function of patients with cardiovascular risk factors in a sample of the Iranian population.

Materials and methods

Study population

This cross-sectional study was conducted during a six-month period from September 2014–February 2015 in cardiology and neurology clinics of Shohada Tajrish Hospital, a tertiary healthcare center affiliated with the Shahid Beheshti University of Medical Sciences, Tehran, Iran. We included a consecutive sample of Iranian patients with cardiovascular risk factors referred to our outpatient clinics. We included those aged between 30–74 years who were able to read and write and those who had cardiovascular risk factors defined as diabetes mellitus (DM), hypertension, dyslipidemia, smoking, and body mass index (BMI) of more than 25 kg/m². Diabetes was defined as a physician diagnosis, a fasting blood glucose level ≥ 126 mg/dl or use of diabetes medication; hyperlipidemia defined by physician diagnosis, use of medication, serum cholesterol concentration > 220 mg/dl, low-density lipoprotein (LDL) cholesterol > 130 mg/dl, or serum triglyceride concentration > 150 mg/dl; hypertension was defined as the current use of antihypertensive medication or an average blood pressure > 140/90 mm Hg on two separate recordings; and current smoking habits in recent year were recorded.²³ We excluded those with a history of ischemic or hemorrhagic neurological accidents (evidenced by neurological imaging such as MRI), those with psychological disorders, and those with a history of cardiovascular diseases such as acute coronary syndrome, documented coronary artery disease or undergoing coronary revascularization. Those with claustrophobia were also excluded from the study. The study protocol was approved by the institutional review board (IRB) and the medical ethics committee of the Shahid Beheshti University of Medical Sciences and all of the patients provided their informed written consent before inclusion in the study.

Study protocol

All of the patients were evaluated for their eligibility to be included in the study by a fourth-year resident of neurology. A complete history and physical examination was performed for each patient and the age, gender, race, education, anthropometric indices, and the cardiovascular risk factors were recorded. The cognitive function of each individual was assessed using the Persian version of the Montreal Cognitive Assessment (MoCA)²⁴ which has been previously validated.²⁵ The MoCA is a measure of global cognitive function and is a 30-point scale that assesses a wide range of cognitive abilities including memory, visuospatial function, executive function, attention, language, and orientation. All of the patients were first instructed how to fill the form. Then all of the included individuals underwent a brain MRI in our center. The Framingham general cardiovascular disease risk prediction model was used to assess the cardiovascular risk^26,27 and the results were correlated with the MRI findings and the cognitive function assessment.

Brain MRI protocol

All of the brain imaging was performed using a 1.5 Tesla MRI machine (Siemens, Avanto, Germany) which was used to acquire images on fluid-attenuated inversion recovery (FLAIR) sequences. MRI scans were prepared with following metrics: axial, 2D FLAIR, repetition time (TR)/echo time (TE) = 2500/100, matrix = 20 × 240, the field of view = 220.

All of the MRI images were reviewed independently by a professor and associate professor of radiology with at least 18 years of experience who were unaware of the cognitive function assessment results of the patients (Figure 1). The reports were collected randomly and the results were checked by both radiologists. To assess absolute inter-observer agreement between the two radiologists, the interclass correlation coefficient (ICC) was calculated. Agreement was interpreted as poor agreement (ICC, 0–0.2), fair agreement (ICC, 0.3–0.4), moderate agreement (0.5–0.6), strong agreement (ICC, 0.7–0.8) or excellent agreement (ICC, >0.8). In this regard, ICC was shown to be 0.79 for qualifying images for lacunar infarcts and 0.88 for WMH. We defined a lacunar infarct as an area of low signal intensity measuring ≤15 mm and ≥ 3 mm on T1-weighted images that was visible as a hyperintense lesion on T2-weighted and FLAIR images in a patient with no history of stroke or TIA. Hyperintense punctuate lesions, evident only on the T2-weighted and FLAIR images, were not considered lacunar infarcts. We defined WMH as hyperintensities in the white matter areas on T2-weighted and FLAIR images. The volume of the WMH was calculated with the Medical Image Processing, Analysis, and Visualization segmentation program (MIPAV) available at https://mipav.cit.nih.gov.

Figure 1.

Multiple focal and confluent abnormal hyper signal foci in periventricular and deep white matter in (a) 67-year-old female and (b) 55-year-old male with cardiovascular risk factors, showing small vessel disease.

The entire lesions were summed to obtain the total volume of white matter lesion (WML). We also defined brain atrophy as the percentage of the intracranial volume occupied by brain tissue. Ventricular enlargement, an indicator for subcortical brain atrophy, was assessed with the ventricular fraction and was calculated as the percentage ventricular volume of the total intracranial volume.

Cognitive function assessment

All of the patients were assessed regarding their cognitive function using the MoCA. The MoCA testing was conducted by a fourth-year neurology resident. Cognition was measured by the total MoCA score and several subdomains. Memory was tested with a free recall of five words, visuospatial function with clock drawing and complex figure copying tasks, executive function with trail-making, phonemic fluency and verbal abstraction, attention with targeted tapping, serial subtraction and digit spans forward and backward, language with confrontation naming, repetition and phonemic fluency, and orientation with place and time.

Statistical analysis

We conducted a pilot study on 24 patients and the correlation between the MoCA score and total volume of the lesion was calculated to be 0.27. Thus, the sample size population was calculated to be 140 patients as follows:

$$\text{The}\hspace{1em}\text{standard}\hspace{1em}\text{normal}\hspace{1em}\text{deviate}\hspace{1em}\text{ for }\alpha ={\text{Z}}_{\alpha }=1.960$$

$$\text{The}\hspace{1em}\text{standard}\hspace{1em}\text{normal}\hspace{1em}\text{deviate}\hspace{1em}\text{ for }\hspace{1em}\beta ={\text{Z}}_{\beta }=1.282$$

$$\text{C}=0.5\times \text{ln}[(1+r)/(1-r)]=0.277$$

$$\text{Total}\hspace{1em}\text{sample}\hspace{1em}\text{size}=\text{N}=[({\text{Z}}_{\alpha }+{\text{Z}}_{\beta })/\text{C}{]}^2+3=140$$

We included a total number of 156 patients to compensate for non-evaluable patients. All data were presented as mean ± standard deviation (SD) (95% confidence interval (CI)) for quantitative variables and frequency, percentage (95% CI) for categorical variables. The Kolmogorov-Smirnov (K-S) test was used to compare distribution of the main variables with normal distribution. In this regard, except for educational level that skewed to the left, other baseline variables including age, BMI, and cardiovascular risk factors had a normal distribution. The association between the quantitative variables was examined using the Spearman's correlation tests. A two-sided p-value of less than 0.05 was considered to be statistically significant.

Results

Overall, we included a total number of 156 patients with cardiovascular risk factors who underwent brain MRI and the total volume of the WML and lacunar infarcts were calculated. There were 102 (65.4%) men and 54 (34.6%) women among the patients with mean age of 62.01 ± 11.2 years. The demographic and baseline characteristics of the patients are summarized in Table 1. We found that a total number of 23 (14.74%) patients had lacunar infarcts in brain MRI. The total volume of WMH was found to be 3.62 ± 5.09 ml being distributed in periventricular and subcortical regions (Table 1). The results of the cognitive function assessment using MoCA and the Framingham general cardiovascular risk prediction are summarized in Table 2. Regarding the association between MoCA score with cardiovascular risk profiles (Table 3), an adverse association was revealed between MoCA cognition score and different cardiovascular risk profiles including Framingham BMI score (r = −0.252, p < 0.001), Framingham lipid score (r = −0.344, p < 0.001), vascular age lipid score (r = −0.486, p < 0.001), and vascular age BMI score (r = −0.360, p = 0.001). With respect to association between the volume of WMH with different components of cognitive functional status (Table 3), we found that the total volume of WMH was negatively associated with total MoCA cognition score (r = −0.477, p < 0.001) as well as with some components of cognitive function including memory (r = −0.288, p = 0.007), attention (r = −0.265, p = 0.011), language (r = −0.230, p = 0.027) and orientation (r = −0.306, p = 0.003). Our study also showed an adverse association between the total number of lacunar infarcts and total MoCA cognition score (r = −0.218, p = 0.038) and consequently with some cognition components including memory (r = −0.304, p = 0.013), attention (r = −0.248, p = 0.037), abstraction (r = −0.189, p = 0.046) and orientation (r = −0.253, p = 0.002).

Table 1.

The baseline characteristics of 156 patients with cardiovascular risk factors included in the current study.

Age, years	62.01 ± 11.2
Gender, %
Male	54 (34.6, 95% CI: 27.6–42.4)
Female	102 (65.4, 95% CI: 57.6–72.4)
Body mass index (kg/m2)
Education level, %
Illiterate	36 (23.1, 95% CI: 17.2–30.3)
Diploma	80 (51.3, 95% CI: 43.5–59.0)
Associate degree	10 (6.4, 95% CI: 3.5–11.4)
Bachelor of science	26 (16.7, 95% CI: 11.6–23.3)
Master of science	4 (2.6, 95% CI: 1.0–6.4)
Cardiovascular risk factors, %
Hypercholesterolemia	111 (72.4, 95% CI: 63.6–77.7)
Hyperlipidemia	100 (64.1, 95% CI: 56.3–71.2)
High triglyceride	55 (35.2, 95% CI: 28.2–43.0)
Hypertension	33 (21.1, 95% CI: 15.5–28.2)
Diabetes mellitus	31 (19.8, 95% CI: 14.4–26.8)
High hemoglobin A1C	25 (16.1, 95% CI: 11.1–22.6)
White matter lesions volume
Paraventricular (ml)	1.54 ± 2.31 (95% CI: 1.18–1.90)
Subcortical (ml)	2.08 ± 1.77 (95% CI: 1.80–2.36)
Total (ml)	3.62 ± 5.09 (95% CI: 2.82–4.42)
Lacunar infarcts, %
One lacunar	17 (10.9, 95% CI: 6.9–16.7)
More than one lacunar	6 (3.8, 95% CI: 1.8–8.1)

CI: confidence interval.

Table 2.

The results of cognitive function assessment using Montreal Cognitive Assessment (MoCA) and Framingham general cardiovascular disease risk prediction in a 156 sample of patients with cardiovascular risk factors included in the current study.

Variable	Value
MoCA score (points)	20.16 ± 5.15 (95% CI: 19.4–21.0)
Visuoconstructional	2.55 ± 1.59 (95% CI: 2.30–2.80)
Naming	2.50 ± 0.59 (95% CI: 2.41–2.59)
Memory	1.83 ± 1.36 (95% CI: 1.62–2.04)
Attention	4.55 ± 1.56 (95% CI: 4.30–4.79)
Language	1.52 ± 1.08 (95% CI: 1.35–1.69)
Abstraction	0.89 ± 0.78 (95% CI: 0.76–1.01)
Orientation	5.46 ± 0.93 (95% CI: 5.31–5.61)
Framingham risk assessment
Framingham lipid score	19.18 ± 14.48 (95% CI: 16.9–21.4)
Framingham BMI score	22.84 ± 16.22 (95% CI: 20.3–25.4)
Vascular age lipid score	71.94 ± 14.97 (95% CI: 69.6–74.3)
Vascular age BMI score	73.09 ± 14.57 (95% CI: 70.8–45.4)

BMI: body mass index; CI: confidence interval.

Table 3.

Results of correlation analysis utilizing Spearman’s correlation between the cognitive function, general cardiovascular risk prediction score and the volume of white matter lesion and lacunar infarcts.

Item	Correlation coefficient	p-Value
MoCA score
Framingham lipid score	−0.344	<0.001
Framingham BMI score	−0.252	0.001
Vascular age lipid score	−0.486	<0.001
Vascular age BMI score	−0.360	<0.001
Total volume of WMH
Total	−0.477	<0.001
Visuoconstructional	−0.139	0.187
Naming	0.108	0.305
Memory	−0.280	0.007
Attention	−0.265	0.011
Language	−0.230	0.027
Abstraction	0.161	0.124
Orientation	−0.306	0.003
Total number of lacunar infarcts
Total	−0.218	0.038
Visuoconstructional	0.098	0.215
Naming	−0.103	0.067
Memory	−0.304	0.013
Attention	−0.248	0.037
Language	0.085	0.317
Abstraction	−0.189	0.046
Orientation	−0.253	0.002

BMI: body mass index; MoCA: Montreal Cognitive Assessment; WMH: white matter hyperintensity.

With a view to a closer assessing the relation between cognitive function subtypes and the volume of white matter lesion, two different regions of white matter including paraventricular and subcortical regions were also considered (Table 4). We found adverse association between volume of paraventricular region and the score for cognition components of memory (r = −0.233, p = 0.018), and language (r = −0.207, p = 0.033), while volume of subcortical region was inversely associated with the scores for naming (r = −0.210, p = 0.030), attention (r = −0.174, p = 0.048), and language (r = −0.257, p = 0.007), and abstraction (r = −0.201, p = 0.029).

Table 4.

Results of correlation analysis utilizing Spearman’s correlation between the cognitive function subtypes and the volume of white matter lesion in different regions.

	Correlation coefficient	p-Value
Periventricular volume of WMH
Visuoconstructional	−0.199	0.041
Naming	0.068	0.418
Memory	−0.233	0.018
Attention	−0.088	0.097
Language	−0.207	0.033
Abstraction	0.028	0.604
Orientation	−0.132	0.056
Subcortical volume of WMH
Visuoconstructional	−0.127	0.074
Naming	−0.210	0.030
Memory	0.029	0.742
Attention	−0.174	0.048
Language	−0.257	0.007
Abstraction	−0.201	0.029
Orientation	−0.145	0.062

WMH: white matter hyperintensity.

Discussion

As the first finding in our cross-sectional study, we showed a significant relationship between cognitive impairment and different cardiovascular risk factors. Several studies have revealed such associations in different populations, but this was the first survey based on Iranian adults. In this survey, significant associations were shown between cognitive impairment and two major cardiovascular risk determinants including high BMI and hyperlipidemia. In other words, any metabolic disturbances leading to raised BMI (in obese patients) and lipid metabolic disturbances can effectively affect cognition status. As shown by Grodstein,²⁸ there is a central role for DM, hyperlipidemia, and hypertension, all considered as risk factors for cognitive decline, especially in later life. As revealed by Takeda et al.,²⁹ higher levels of triglycerides, total LDL and very-low-density lipoprotein (VLDL) cholesterol were associated with poorer performance on the digit span and category fluency tests. They also showed higher BMI as an indicator for poorer delayed recall performance. In another study by Kim et al. in 2016,³⁰ individuals with obesity (BMI ≥ 25 kg/m²) compared to normal weight were marginally more likely to experience the development of severe cognitive impairment especially in women and thus obesity was associated with risk of cognitive decline among the mid- and old-age population. Ma et al.³¹ revealed that higher blood concentrations of total cholesterol and low-density lipoprotein cholesterol (LDL-C) in were associated with faster global cognitive decline in the very old condition. However, some authors could not demonstrate such causality associations. As shown by Arntzen et al.,³² although diabetes, hypertension, smoking, and low physical activity were all independently associated with lower cognitive ability in both genders, no consistent association was found between total cholesterol, high-density lipoprotein (HDL)-cholesterol, or BMI, and cognitive impairment. In some studies, the association between cardiovascular risk profiles and cognitive decline was shown to be specific to the presence of hypertension, diabetes mellitus, and even hypoglycemia.^33–36 Higher BMI have been well shown to be significantly correlated with reduced gray matter volumes in some brain regions. In this regard, a smaller volume of the left orbitofrontal region, an area underlying complex cognitive tasks, was associated with poorer executive function. This can explain how obesity impacts brain structure and subsequent cognition.³⁷ Some possible biological mechanisms explain the association between high cholesterol levels and cognitive decline. Compared to the cholesterol of the periphery with de novo synthesis and provided by diet in equal parts, brain cholesterol is stable, with the major input solely through de novo synthesis locally rather than transfer from the periphery. Thus, the central nervous system contains its own supply of lipoproteins.³⁸ Brain cholesterol has been implicated in playing a role in altering the degradation of the amyloid precursor protein, which could have an impact on the accumulation of amyloid beta peptides, contributing to the pathogenesis of dementia.³⁸ In addition, hypercholesterolemia is involved in atherogenesis, which could result in macro- and micro-vascular diseases, deteriorating cognitive function both at subclinical^39,40 and clinical levels.⁴¹

We note some limitations to our study. First, the number of included patients was limited and thus the study was calculated to have 80% power for correlation analysis. Thus, further studies with larger study populations should be conducted. The other limitation was that we only assessed cognitive function using the MoCA. Although the MoCA is a validated questionnaire for evaluation of cognitive function, it has shortcomings in evaluating the memory (verbal and short-term), psychiatric, and personality subgroups. Thus, further studies including these variables in cognitive assessment are recommended. This study is among the few available studies addressing the correlation between the WMH and lacunar infarcts in the normal population with cardiovascular infarcts.

In conclusion, the total volume of WMH and number of lacunar infarcts correlated with a lowering in cognitive function of normal population with cardiovascular risk factors. Periventricular lesions are associated with impaired memory, language, and visuoconstructional function while subcortical lesions are associated with impairments in naming, attention, language, and abstraction functions. Further studies are required for extending the results of the current study.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.