Obstructive sleep apnea (OSA) is characterized by repetitive complete or partial collapse of the upper airway and reduction of airflow during sleep. It is associated with significantly increased daytime muscle sympathetic nerve activity thought to result from the repetitive intermittent periods of hypoxemia during sleep and brain alterations that are likely to result. Different brain regions are affected by subsequent hypoxia/anoxia. Neurodegenerative processes result in measurable atrophy of cortical gray matter in the temporal lobes and posterior cingulate cortex, as well as in subcortical structures such as the hippocampus, amygdala, and thalamus. This study involved a group of firstly diagnosed, therapy
Obstructive sleep apnea (OSA) is characterized by repetitive complete or partial collapse of the upper airway and reduction of airflow during sleep. It is associated with significantly increased daytime muscle sympathetic nerve activity thought to result from the repetitive intermittent periods of hypoxemia during sleep and brain alterations that are likely to result [
Studies conducted in younger and middle-aged adults indicate that the effects of sleep fragmentation and nocturnal hypoxemia probably support the cognitive deficits associated with OSA [
Nonetheless, a relative paucity of investigations dealing with the interrelationships between OSA, brain integrity, and cognitive decline in older adults should be noted. As adults age, they may experience neurodegenerative processes resulting in measurable atrophy of cortical gray matter in the temporal lobes and posterior cingulate cortex, as well as subcortical structures such as the hippocampus, amygdala, and thalamus. These changes are evident even in the transitional or “at-risk” stages between normal aging and dementia, defined as those with subjective memory concerns and mild cognitive impairment [
A study by Elliott et al. [
They estimated that NAFLD was associated independently with reduced cognitive performance independent of cardiovascular disease and its risk factors. Previous studies on animals have shown a very strong connection between OSA and NAFLD [
However, no study has examined the tissue-specific effects of recurrent airway closure in humans. A few studies have shown that liver enzymes may be acutely elevated in OSA and are lowered with CPAP [
We aimed to assess whether OSA is associated with structural brain changes in various brain regions and whether OSA consecutively leads to cognitive impairment compared to NAFLD patients. Our goal was to compare the differences in cognitive functions in individuals with OSA and NAFLD relative to those with NAFLD but without OSA. We hypothesized that significant differences in cognitive statuses exist between those two groups of interest.
This study involved a group of firstly diagnosed, therapy-naive NAFLD patients, out of which 144 (96 males and 48 females), aged 34–57 (mean 47.88 ± 6.07), satisfied recruiting criteria for the study and control groups. The grouping criterion for the division into the studied and control group was the presence of the OSA, so the studied group included the patients with NAFLD and OSA, and the controls were the individuals with NAFLD but without OSA. The study was approved by the Ethical Committee of the Clinical and Hospital Center “Dr Dragisa Misovic–Dedinje,” Belgrade.
All the participants were acquainted in detail with the study aim and design before entering the program. They all signed a written consent afterward.
A selection flow diagram is shown in Figure
Patient selection flowchart.
Recruiting criteria were as follows: Older than 18. No previous history of viral hepatitis of any kind, haemochromatosis, autoimmune hepatitis, cirrhosis, or other chronic liver diseases. No presence of severe cardiopulmonary disease. The absence of endocrinological disorders: hypothyroidism, hypercorticism, and syndrome of the polycystic ovaries. No history or clinical signs of excessive alcohol abuse (>20 g/day for males and >10 g/day for females). No neuropsychiatric disease involving signs of any kind of dementia, and/or neuropsychiatric medication history, or any other hepatotoxic drugs. No visible traces of illicit drugs abuse. Negative urine multiple drug test on 10 kinds of drugs: cannabinoids, opiates, amphetamines, 3, 4-methylenedioxymethamphetamine, cocaine/crack, benzodiazepines, tricyclic antidepressants, barbiturates, methadone, and buprenorphine. No visible focal or diffuse changes in the gray matter of the brain on MRI. Fazekas score 0 on MRI scan. Fazekas score is the estimated level of the white matter vascular changes and is the aftermath of the brain vessels’ atherosclerotic changes. Absence of any rheumatologic disease. Patients who used antidiabetic drugs, insulin, antilipemic drugs, uricosuric drugs, steroids, and oral contraceptives were excluded from the study.
Volume measurements of the gray and white matter and lateral ventricles of the brain were performed on 3D T1-weighted MR images (Phillips Inc. Holland). Acquisition parameters were as follows: TR = 9.8 ms; TE = 4.6 ms; flip angle = 8; section thickness = 1.2 mm; number of sections = 120; no section gap; whole-brain coverage; FOV = 224 mm; matrix = 192; reconstruction matrix = 256. Routine T2-weighted and FLAIR images were performed to rule out a mass lesion as a contributory factor to memory loss or cognitive decline. The structures were manually outlined and compared with automatic extraction of the regions of interest in commercially available software. The software finally computed the volumes required.
After the diagnostic procedures, all the subjects underwent psychological testing of cognitive impairments using the Montreal Cognitive Assessment (MoCA) test, Serbian version. The test has several levels of testing: alternating connection (connect Figure
For body weight and height, the patients were measured in bare feet and light clothing in the morning with the same equipment. Body mass index (BMI) was calculated by dividing body weight by height square (kg/m2).
Fasting blood was taken in the morning for the measurement of serum glucose, and lipid profile comprising total cholesterol (TC), high-density lipoprotein-cholesterol (HDL-C), low-density lipoprotein-cholesterol (LDL-C), and triglycerides (TG). Adipokines, adiponectin, and leptin were analyzed and compared. All the tests were run by AO-BK-200 mini Auto Biochemistry Analyzer, Alpha Omega Electronics, Madrid, Spain.
All participants, examined and controls, underwent polysomnography (PSG) in the sleep department. It was performed within four weeks of the MRI scan and neuropsychological testing. Nocturnal PSGs were collected on an ambulatory recording system with the Alice PDx portable diagnostic recording device (Philips Respironics), together with nasal airflow which was recorded with the nasal pressure transducer. Respiratory effort was assessed using thoracic and abdominal bands; blood oxygen saturation was revealed by pulse oximetry. Patients were advised not to disturb their usual bedtime weekly rhythm and were required to abstain from caffeinated beverages (coffee, caffeinated soda) at least eight hours before and, especially, during PSG data collection. The study was reported by an accredited sleep physician.
Sleep staging was scored according to the criteria of the American Academy of Sleep Medicine [
The atherogenic index of plasma (AIP) represents the risk for atherosclerosis. It is calculated as the logarithmic value of the triglyceride/HDL score. The risk was interpreted as follows: AIP < 0.11 low risk; AIP = 0.11–0.21 intermediate risk; and API > 0.21 increased risk. The following index values were analyzed: cholesterol/HDL, triglycerides/HDL, and HDL/LDL.
The liver was assessed as normal when the consistency was homogeneous, displayed fine level echoes, minimally hyperechoic or even isoechoic in contrast to the regular renal cortex. Mild steatosis was evaluated as a minor increase in liver echogenicity. In moderate steatosis, there were visual images associated with intrahepatic vessels, the slightly damaged diaphragm, and the existence of increased liver organ echogenicity. Severe steatosis was evaluated as a marked increase in hepatic echogenicity, poor penetration of posterior segment from the right lobe of the liver, and poor or no visual images from the hepatic vessels and the diaphragm [
FibroScan® (Echosens group) was used to determine the fibrosis grade in the liver parenchyma. The normal range for a FibroScan is between 2 and 7 kPa. The average normal result is 5.3 kPa. The results vary based on the liver disease in question. For NASH/NAFLD there are 4 stages of scarring: •
Statistical testing was performed by the commercially available software (SPSS 17.0, Inc., Chicago Il, US). Besides measures of the central tendency (mean and standard deviation (SD), minimum and maximum), potential differences of mean values were assessed with one-way analysis of variance (ANOVA) with the Bonferroni post hoc correction, Student’s
Demographic data are shown in Table
Demographic parameters of examined population.
Group according to sleep apnea severity | Examined | Controls | Total | Significance |
---|---|---|---|---|
( | ( | ( | ||
Parameter | ||||
Age (years ± SD) | 47.88 ± 6.07 | 47.62 ± 6.97 | 46.94 ± 9.00 | NS |
Gender (male/female) | 58/10 | 38/38 | 96/48 | |
Education level (grammar/high school/university) | 3/7/12 | 25/33/18 | 43/70/31 | NS |
Body mass index (BMI, kg/m2) | 35.34 ± 7.31 | 34.95 ± 8.27 | 31.90 ± 6.61 | |
Leptin (ng/mL) m/f | 11.36 ± 1.97 | 4.39 ± 2.17 | 10.39 ± 2.53 | |
Adiponectin (ng/mL) m/f | 8.06 ± 0.97 | 8.8 ± 1.44 | 8.19 ± 2.44 | NS |
Glucose (mmol/l) | 5.31 ± 0.65 | 5.75 ± 1.55 | 5.52 ± 1.28 | NS |
C reactive protein (mg/l, mean ± SE) | 3.86 ± 1.27 | 3.74 ± 0.39 | 3.39 ± 0.25 | NS |
Cholesterol (mmol/l, mean ± SD) | 5.19 ± 0.63 | 5.74 ± 1.18 | 5.57 ± 1.11 | NS |
HDL (mmol/l, mean ± SE) | 0.95 ± 0.04 | 1.37 ± 0.07 | 1.18 ± 0.04 | |
LDL (mmol/l, mean ± SE) | 3.23 ± 0.17 | 3.27 ± 0.12 | 3.44 ± 0.08 | |
Triglycerides (mmol/l, mean ± SE) | 2.68 ± 0.35 | 1.86 ± 0.13 | 2.24 ± 0.12 | |
Cholesterol/HDL ratio (mean ± SE) | 6.02 ± 0.33 | 4.64 ± 0.18 | 5.27 ± 0.18 | |
Triglycerides/HDL ratio | 3.32 ± 0.48 | 1.62 ± 0.15 | 2.41 ± 0.19 | |
AIP | 0.52 | 0.21 | 0.38 | |
HDL/LDL ratio (mean ± SE) | 0.32 ± 0.03 | 0.43 ± 0.04 | 0.38 ± 0.02 | NS |
AHI per hour (mean ± SE) | 12.23 ± 5.49 | 2.95 ± 0.12 | 12.16 ± 1.31 | |
ODI per hour (mean ± SE) | 14.35 ± 4.51 | 2.78 ± 0.13 | 11.08 ± 1.24 | |
Total brain volume (cm3, mean ± SE) | 1441.19 ± 32.31 | 1403.37 ± 18.85 | 1398.96 ± 13.16 | NS |
MoCA score (mean ± SD) | 24.23 ± 3.14 | 25.53 ± 3.20 | 25.28 ± 3.04 |
NS: not significant.
The concentration of serum triglyceride, HDL, and LDL differed significantly: patients with severe OSA had the lowest concentration of HDL and the highest level of LDL. All the examinees with OSA were at very high risk for atherosclerosis (all above 0.51 risk index).
Although total brain volumes among the groups observed did not differ significantly, volumes of the structures of interest were significantly lower in the group of examinees with OSA. Higher volumes were obtained for the lateral ventricles on both hemispheres in OSA suffering patients, while volumes of the amygdaloid complexes did not differ significantly (Table
Brain volumes according to the sleep apnea of the observed groups.
Group according to sleep apnea severity | Mild ( | Moderate ( | Severe ( | Controls ( | Total ( | Significance | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
L | R | L | R | L | R | L | R | L | R | L | R | |
Caudate nucleus volume (cm3, mean ± SD, L/R) | 3.76 ± 0.11 | 3.80 ± 0.15 | 3.77 ± 0.04 | 3.81 ± 0.16 | 3.79 ± 0.15 | 3.85 ± 0.14 | 4.5 ± 0.65 | 4.56 ± 0.64 | 4.22 ± 0.6 | 4.17 ± 0.6 | ||
Putamen volume (cm3, mean ± SD, L/R) | 5.94 ± 0.10 | 5.95 ± 0.10 | 5.97 ± 0.11 | 5.98 ± 0.11 | 5.97 ± 0.12 | 5.99 ± 0.12 | 6.31 ± 0.62 | 6.30 ± 0.70 | 6.15 ± 0.49 | 6.16 ± 0.54 | ||
Globus pallidus volume (cm3, mean ± SD, L/R) | 2.13 ± 0.10 | 2.12 ± 007 | 2.13 ± 0.08 | 2.61 ± 0.46 | 2.18 ± 0.79 | 2.12 ± 0.07 | 2.61 ± 0.46 | 2.58 ± 0.40 | 2.38 ± 0.41 | 2.36 ± 0.37 | ||
Thalamus volume (cm3, mean ± SD, L/R) | 6.29 ± 0.10 | 6.28 ± 0.09 | 6.27 ± 0.09 | 6.26 ± 0.09 | 6.28 ± 0.07 | 6.28 ± 0.07 | 7.46 ± 0.77 | 7.48 ± 0.90 | 6.90 ± 0.81 | 6.91 ± 0.79 | ||
Hippocampal formation volume (cm3, mean ± SD, L/R) | 2.82 ± 0.12 | 2.81 ± 0.12 | 2.77 ± 0.13 | 2.77 ± 0.13 | 2.80 ± 0.13 | 2.80 ± 0.13 | 3.99 ± 0.63 | 3.96 ± 0.68 | 3.43 ± 0.75 | 3.41 ± 0.77 | ||
Lateral ventricle volume (cm3, mean ± SD, L/R) | 7.48 ± 0.14 | 7.49 ± 0.14 | 7.49 ± 0.13 | 7.51 ± 0.13 | 7.50 ± 0.14 | 7.51 ± 0.14 | 5.86 ± 0.55 | 5.81 ± 0.54 | 6.63 ± 0.92 | 6.61 ± 0.94 | ||
Amygdaloid complex volume (cm3, mean ± SD, L/R) | 1.65 ± 0.11 | 1.64 ± 0.12 | 1.61 ± 0.08 | 1.61 ± 0.08 | 1.67 ± 0.08 | 1.68 ± 0.08 | 1.68 ± 0.19 | 1.69 ± 0.21 | 1.67 ± 0.15 | 1.68 ± 0.17 | NS | NS |
Prefrontal cortex volume (cm3, mean ± SD, L/R) | 140.45 ± 6.20 | 141.54 ± 6.07 | 143.43 ± 8.16 | 144.50 ± 8.05 | 145.93 ± 8.43 | 146.98 ± 8.36 | 161.16 ± 18.85 | 158.84 ± 19.32 | 153.30 ± 16.95 | 152.58 ± 16.46 | ||
Gray matter volume (cm3, mean ± SD) | 351.91 ± 4.91 | 348.79 ± 6.02 | 348.93 ± 7.57 | 341.43 ± 40.03 | 335.79 ± 29.93 |
The severity of OSA differed among the observed groups with liver steatosis (Table
Distribution of steatosis severity among groups with different levels of OSA.
OSA | Grade of liver steatosis | |||
---|---|---|---|---|
Mild | Moderate | Severe | Total | |
Mild | 11 | 0 | 0 | 11 |
Moderate | 2 | 8 | 4 | 14 |
Severe | 12 | 13 | 18 | 43 |
Without | 36 | 25 | 15 | 76 |
Total | 61 | 46 | 37 | 144 |
rho = 0.214;
According to the MoCA score, the groups divided by the grade of liver steatosis differed (
Regarding the severity of OSA, the level of the cognitive deficit did not differ among the obtained groups.
Discriminative function analysis outlined cognitive level as the only parameter of importance for the classification of a newly obtained patient into one of the groups of interest: equation = −9.19 + 0.37 × MoCA; centroids for groups: mild, −0.17; moderate, −0.63; severe, 0.25 and cutoff points: mild to moderate, −0.23; moderate to severe, −0.195; goodness of classification, 71.50%.
Polysomnographic parameters had an inverse influence on the volumes of the structures of interest. The tested subjects had lower volumes when both AHI and ODI were higher (Table
Correlation matrix between polysomnographic parameters and volumes obtained.
Parameter | AHI per hour | ODI per hour | Significance AHI | Significance ODI | ||||
---|---|---|---|---|---|---|---|---|
L | R | L | R | L | R | L | R | |
Caudate nucleus volume (cm3) | ||||||||
Putamen volume (cm3) | ||||||||
Globus pallidus volume (cm3) | ||||||||
Thalamus volume (cm3) | ||||||||
Hippocampal formation volume (cm3) | ||||||||
Lateral ventricle volume (cm3) | ||||||||
Prefrontal cortex volume (cm3) | ||||||||
Gray matter volume (cm3) |
L: left hemisphere; R: right hemisphere.
In the multivariate regression analyses the patients with higher levels OSA showed a significant reduction in all volumes of brain structures except for amygdaloid complex and white matter volume (Table
Results of the multivariate regression analyses.
Brain structures | Grade of OSA | Grade of liver steatosis | ||
---|---|---|---|---|
L | R | L | R | |
Caudate nucleus volume (cm3) | ||||
Putamen volumes (cm3) | ||||
Globus pallidus volume (cm3) | ||||
Thalamus volume (cm3) | ||||
Hippocampal formation volume (cm3) | ||||
Lateral ventricle volume (cm3) | ||||
Amygdaloid complex volume (cm3) | ||||
Prefrontal cortex volume (cm3) | ||||
White matter volume (cm3) | ||||
Gray matter volume (cm3) |
The study aimed to reveal whether OSA is associated with structural brain changes in diverse brain regions and whether the grade of liver steatosis influences both OSA appearance and subsequent cognitive alterations, as the result of IH.
Our results indicate that the volume changes of the overall cortex and basal nuclei are related to AHI and ODI as the main parameters and both strongly influenced volume decrease. These results are in correlation with previous studies which showed changes in volume values of the brain structures of interest. Kim et al. [
OSA causes nocturnal intermittent hypoxemia and sleep fragmentation in response to oxygen desaturation. Some investigators indicated that OSA, vascular depression, and cognitive impairment are linked to several pathologic processes in the cerebral microvascular and neurovascular systems [
The major risk factors for OSA include obesity, male sex, alcohol and smoking habits, a family history of OSA, and upper airway structural abnormalities such as a large neck girth and craniofacial abnormalities. NAFLD is most commonly associated with metabolic risk factors, such as obesity, diabetes mellitus type 2, and elevated triglyceride levels, but some recent studies have reported that chronic intermittent hypoxia (CIH) can be an independent risk factor to induce liver damage [
Several studies examined OSA influence on structural changes using magnetic resonance spectroscopy, the decreased N-acetyl aspartate, and choline concentration in prefrontal subcortical white matter. Alterations revealed the early structural changes such as neuronal loss and axonal damage [
In some magnetic resonance spectroscopy-guided studies, the left hippocampus was stressed as a region especially sensitive to intermittent hypoxia in OSA suffering patients [
The atherogenic index of plasma was significantly lower in the controls but correlated with AHI, ODI, and BMI indexes. On the contrary, Sparks et al. [
An et al. [
The role of two adipokines, leptin and adiponectin, in cognitive regulation is at least equivocal and insufficiently elucidated, considering literature data. Our patients with OSA had lower concentrations of adiponectin, but this difference was not statistically verified. It has already been published that adiponectin is lower in the examinees with OSA [
Leptin values in our study are higher in patients with OSA and NAFLD, compared to the controls with NAFLD only. The highest values were noted in persons with severe OSA. Leptin is a potent ventilation stimulant acting on central respiratory control nuclei. The central satiety effects of leptin are abrogated in obesity. Leptin resistance is defined as a failure of high-circulating levels of leptin to decrease hunger and promote energy expenditure.
OSA and IH, powerful triggers of oxidative stress, increase peripheral leptin levels and also induce leptin resistance (for a detailed review, see [
The statistical trend for lower leptin levels to be associated with higher cognitive scores was revealed in a large sample of 2731 subjects measured by MoCA. Excessive leptin per unit of fat was associated with lower total MoCA score and memory in black men and with higher MoCA scores in white men [
In our study, two indexes, characteristic for OSA, AHI, and ODI, were found to have a strong influence on the reduction of the volumes of almost all regions of interest. Only amygdaloid complexes were spared from the volume reduction. Total brain volume was found insignificantly different in our sample, which could be the result of larger lateral ventricle volumes and, subsequently, possible higher quantities of cerebrospinal liquor. OSA, that is, impaired AHI and ODI indexes, likely causes cognitive impairment through IH, hormonal imbalance, and/or systemic inflammation, either independently or via the resultant endothelial dysfunction that occurs. Still, the cognitive defect is only partially reparable after CPAP treatment (for a detailed review, see [
CPAP is the first-line treatment for NAFLD patients with OSA, but the effect of CPAP treatment on liver disease is still controversial and unclear. CPAP treatment may be beneficial to NAFLD patients with OSA independent of metabolic risk factors, but a sufficiently long therapeutic duration longer than three months may be needed to achieve positive effects on the liver enzymes and liver steatosis especially in patients with moderate-to-severe OSA [
The limitations of this study are as follows: This is a referral, not a cohort study, restricted only to the patients referred to our department and outpatient clinics A relatively small number of patients with mild and moderate OSA were investigated Only noninvasive tests were performed for NAFLD It is limited to newly diagnosed, therapy-naive patients The diagnoses were dependable on ultrasonographer and sleep doctor’s skills and experience There was an inability to perform functional magnetic resonance imaging because our institution does not possess one
Syndrome of OSA worsens the cognitive status in patients with NAFLD. The possible underlying mechanism is the influence on the reduction of cortical and subcortical structures driven by constant apnea/hypopnea episodes, and consecutive hypoxia that initiates the domino process of deteriorating biochemical reactions in the brain.
The data used to support the findings of this study are deposited in the DOI repository.
The authors declare that there are no conflicts of interest regarding the publication of this study.
The study was partially financed by a grant from the Ministry of Education, Science and Technological Development, Government of the Republic of Serbia (no. III 41020).