Authors: Shuxin Guo, Chunguang Liang, Jinrui Fei, Ying Ma, Weiwei Su, Huameng Xu, Jie Kong
Categories: Research, Obstructive sleep apnea, Daytime sleepiness, Mild cognitive impairment, Sleep apnea severity, Sleep quality
Source: BMC Psychiatry
Authors: Shuxin Guo, Chunguang Liang, Jinrui Fei, Ying Ma, Weiwei Su, Huameng Xu, Jie Kong
It is becoming increasingly acknowledged that obstructive sleep apnea (OSA) is a variable factor influencing cognitive health. The aims of this study were to explore whether the severity of OSA is related to the occurrence of mild cognitive impairment (MCI) in people with OSA and whether different degrees of daytime sleepiness and nighttime sleep quality are related to MCI.
The study was cross-sectional. For our subjects, we selected individuals who visited the Sleep Medicine Center of Jinzhou Medical University’s First Affiliated Hospital between May 2023 and October 2024, underwent polysomnography (PSG) or the home sleep apnea test (HSAT), and were diagnosed with OSA. The patients were split into two one for normal cognitive function (NC) and the other for MCI. MCI was defined as Montreal Cognitive Assessment (MOCA) < 26 points. The apnea-hypopnea index (AHI) and the oxygen desaturation index (ODI) were used to assess the severity of sleep apnea. The Epworth Sleepiness Scale (ESS) was used to assess patients’ daytime sleepiness, and the Pittsburgh Sleep Quality Index Scale (PSQI) was used to assess their sleep quality. Multivariate logistic regression analysis was used to evaluate the correlation between the variables.
In this study, 387 patients with OSA (45.3 ± 12.6 years, 82.4% male) were included, of whom 38% had MCI (52.4 ± 11.9 years, 74.1% male). In the unadjusted model, the sleep apnea severity, daytime sleepiness severity, and different sleep quality at night were positively related to MCI. After controlling for confounding factors, this correlation was no longer significant. Only severe sleep apnea (AHI ≥ 30/h, p < 0.001), poor nighttime sleep quality (PSQI ≥ 9, p = 0.020), and sleepiness (ESS ≥ 11, p < 0.05) were associated with increased risk of MCI.
Severe sleep apnea, poor sleep quality, and sleepiness were relevant to increased risk of MCI. It provides a basis for a more comprehensive understanding of the relationship between OSA and MCI.
Complete or partial blockage of the upper respiratory tract during sleep is the hallmark of obstructive sleep apnea (OSA), a respiratory disease that can cause hypercapnia, intermittent hypoxia, and fragmented sleep [1]. Patients with OSA may suffer from snoring, nocturnal decrease in blood saturation, poor quality of sleep at night, and daytime drowsiness, which can affect patients’ work and life, and even threaten their life and health. OSA has become a global health concern, and it is estimated that about 1 billion people between the ages of 30 and 65 have mild to severe OSA, and 425 million people have moderate to severe OSA [2]. Severe OSA leads to dysfunction of several organ systems, including cardiovascular, endocrine, and neurological systems [3], and its neurological effects are mainly manifested in cognitive impairment. Patients with OSA frequently suffer from mild cognitive impairment (MCI). MCI is a common age-related cognitive syndrome that lies between normal aging and dementia [4]. At present, the proportion of OSA combined with MCI is high, and the prevalence rate of MCI in people with OSA is about 37–48% [5, 6]. Therefore, early attention to their cognitive problems can prevent irreversible cognitive impairment, improve patients’ quality of life, and reduce economic and care burdens.
OSA raises the risk of cognitive impairment. According to a major meta-analysis, people with sleep-related breathing disorders had a 26% higher possibility of developing cognitive impairment [7]. The severity of OSA and MCI has been found to be linked in a number of studies [8, 9], particularly the greater incidence of MCI in people having moderate to severe OSA [6]. However, others have also found no link between MCI and the existence or severity of OSA [10, 11].
The relationship between OSA and MCI is controversial, and these differences may be due to the size and type of the study population, the MCI screening tool used, definitions describing the characteristics of OSA, and various confounders. The Montreal Cognitive Assessment (MOCA), one of the most commonly utilized screening tools, has been shown to be more sensitive and comprehensive in evaluating MCI than the Mini-Mental State Examination (MMSE) and is better able to identify the presence of MCI [12]. Therefore, we used the MOCA to examine the link between MCI and sleep apnea severity. Furthermore, the statistical power and adaptability of effect sizes were constrained by the very small sample sizes utilized in many earlier studies in clinical populations [8]. To address this issue, we recruited more patients with OSA to participate in this study.
Excessive daytime sleepiness (EDS) and poor nighttime sleep quality are the common clinical symptoms in people with OSA. EDS usually manifests as unconscious sleep or subjective perception of sleepiness during daytime daily activities, which is a tendency to sleep that cannot be controlled autonomously, and it has been attested that people with OSA who have a combination of EDS can be accompanied by persistent cognitive impairment [13], which affects driving and work performance and severely impacts quality of life, but the relationship has not been further validated. Previous scholars believe that OSA-related cognitive impairment is age-dependent [14], while poorer sleep quality also affects cognitive function [15, 16], and there is a correlation between insomnia severity index and cognitive impairment [17]. No study has simultaneously explored the correlation between sleep apnea severity, daytime sleepiness, nighttime sleep quality, and MCI. Our study may provide the principle for early clinical diagnosis of MCI so that appropriate interventions can be given at an early stage in the clinic to improve patients’ prognosis and quality of life, and provide a reference for the overall clinical management of people with OSA.
Thus, the study’s objectives were to (1) evaluate the prevalence rate of MCI in people with OSA using MOCA, (2) determine whether sleep apnea severity was significantly related to MCI, and (3) determine whether daytime sleepiness and nighttime sleep quality were significantly related to MCI. We hypothesized that the prevalence of MCI would be highest in people with severe OSA and that there would be a positive correlation between MCI and sleep apnea severity. MCI was positively relevant to the degree of daytime sleepiness and poor sleep quality in people with OSA.
The study was cross-sectional. The data used in the study were obtained from people who visited the Sleep Medicine Center of Jinzhou Medical University’s First Affiliated Hospital between May 2023 and October 2024 and underwent polysomnography (PSG) or home sleep apnea test examinations. (1) aged ≥ 18 years with informed consent to participate in this study voluntarily; (2) patients with OSA (AHI ≥ 5/h) diagnosed by polysomnography (PSG) or the home sleep apnea test (HSAT); (3) consciousness, stable condition, and able to cooperate with the completion of the investigation. Exclusion criteria (1) people who had been or were undergoing treatment such as a ventilator for OSA; (2) those with a history of epilepsy, serious mental illness, serious heart, liver, or kidney disease, malignant tumor, etc.; (3) those with functional deficiencies such as serious visual and hearing impairment, comprehension impairment, etc., that affect the conduct of the survey study; (4) those with incomplete data; (5) those who were diagnosed as having dementia or Alzheimer’s disease (AD) according to the diagnostic criteria of dementia and Alzheimer’s disease in the 2018 Chinese Guidelines for the Diagnosis and Treatment of Dementia and Cognitive Disorders. Patients who satisfied the requirements were assigned to two those with MCI (MOCA < 26) and those with NC (MOCA ≥ 26), respectively, based on their MOCA scores. The study was performed in accordance with the Declaration of Helsinki, and all subjects signed the informed consent form. Jinzhou Medical University’s Ethics Committee has authorized our study with ethical number JZMULL2024005.
This study used the United States Embla Polysomnography Test system (Model: N7000) or the home sleep apnea test (Model: Embletta MPR) to monitor sleep in OSA patients, collecting data for more than 7 h per night. Patients receiving PSG are required to stay overnight in the sleep monitoring room, and those receiving HSAT do so at home. The HSAT device records raw signals (e.g., respiration, pulse rate, blood oxygen, airflow, chest and abdominal movements, body position, and snore) through sensors, allowing for simplified polysomnographic monitoring capabilities. After the data is saved through the built-in memory card, the patient’s sleep is automatically analyzed by the Remlogic system software, then manually evaluated by a professional sleep technician, and finally reviewed by a medical professional. In this study, patients with an AHI ≥ 5/h were diagnosed as having OSA, and the AHI was the sum of the average number of hypopneas and apneas per hour. Respiratory events were scored according to the 2012 American Academy of Sleep Medicine criteria [18]: hypopnea is defined as a decrease in airflow amplitude of ≥ 30% compared to baseline for at least 10 s, along with a 3% drop in oxygen saturation or arousal. Apnea is defined as a decrease in airflow amplitude of ≥ 90% relative to baseline for at least 10 s. The severity of OSA was assessed by either the apnea-hypopnea index (AHI) or the oxygen desaturation index (ODI). According to the AHI, traditional thresholds were classified as mild OSA (5 ≤ AHI < 15), moderate OSA (15 ≤ AHI < 30), and severe OSA (AHI ≥ 30). According to the ODI, critical values were classified as mild OSA (5 ≤ ODI < 15), moderate OSA (15 ≤ ODI < 30), and severe OSA (ODI ≥ 30). In addition, the OSA parameters we obtained from the patients’ sleep reports were total sleep time (TST), percentage of time oxygen saturation < 90% (T90), mean oxygen saturation (MSaO2), and lowest oxygen saturation (LSaO2).
People with OSA had their cognitive functioning evaluated using the MOCA [19]. The MoCA consists of 12 subtests covering eight cognitive domains, including visuospatial and executive, orientation, naming, memory, abstraction, attention, and language. The total score is 30, and scores < 26 are considered MCI, and the lower the score, the more severe the cognitive impairment. The total score is corrected by adding 1 point if the length of education is less than 12 years.
Daytime drowsiness was evaluated by the Epworth Sleepiness Scale (ESS) [20]. The ESS has eight questions. The overall score ranges from 0 to 24 points, with each item being scored on a four-point scale (range 0 to 3). The higher score is associated with a more severe level of sleepiness. The threshold for issues with daytime sleepiness was set at a total ESS score of more than 10 [21]. People were assigned to four groups according to the degree of sleepiness; ESS scores ≤ 10 were classified as the non-sleepiness group, 11 ≤ ESS scores ≤ 12 were classified as the mild sleepiness group, 13 ≤ ESS scores ≤ 15 were classified as the moderate sleepiness group, and ESS scores ≥ 16 were classified as the severe sleepiness group [22].
Sleep quality was assessed by the Pittsburgh Sleep Quality Index Scale (PSQI) [23]. There are seven themes and eighteen questions in the PSQI. The overall score ranges from 0 to 21 points, with each item being scored on a four-point scale (range 0 to 3), similar to the ESS. The threshold for issues with sleep quality was set at a total PSQI score of more than 5. The patients were grouped according to their sleep quality; PSQI scores ≤ 5 were classified into the good sleep group, 6 ≤ PSQI scores ≤ 8 were classified into the average sleep group, and PSQI scores ≥ 9 were classified into the poor sleep group [24].
Since age, gender, cognitive reserve, economic status, diabetes, cardiovascular disease, physical activity, and social activity may impact the relationship between OSA and MCI [25], we assessed these variables. Sex, age, education level, habitation, occupation, income, and family history of dementia, which were completed based on self-reporting by the study participants. Hypertension, diabetes, coronary disease, and hyperlipidemia were collected according to whether or not a previous diagnosis had been made. Subjects’ height, weight, and neck circumference were measured in the field, and body mass index (BMI) was calculated based on weight (kg) divided by the square of height (m). We defined smoking as an average of ≥ 1 cigarette per day for more than three months and alcohol consumption as an average of ≥ 1 drink per week with more than 50 g per drink. Participating in group activities or shopping with others, traveling with others, etc., ≥ 3 days per week was defined as engaging in social activities. In addition to those necessary for subjects’ daily life activities, a series of purposeful, intense, and repetitive physical activities [26], such as brisk walking, jogging, swimming, tai chi, etc., ≥ 20 min per session, ≥ 3 days per week, was defined as physical exercise.
For continuous variables, if they were distributed normally, they were expressed as the mean values and standard deviations; comparisons were made applying Student’s t-tests; if they were not distributed normally, they were expressed as a median and interquartile range, and comparisons were made applying the Mann-Whitney U-test. Categorical variables were expressed as frequencies and proportions; comparisons were made using Pearson’s chi-square test or Fisher’s exact test. Independent associations between sleep apnea severity, daytime sleepiness, sleep quality, and MCI in people having OSA were assessed by multivariate logistic regression. MCI was analyzed as the dependent variable (MOCA < 26) in different models. Independent variables included the severity of OSA (assessed by AHI or ODI), daytime sleepiness (assessed by ESS), and nighttime sleep quality (assessed by PSQI). In order to prevent multicollinearity between independent variables, covariance diagnostics were performed. Include variables with a variance inflation factor (VIF) < 5 and some variables that were significantly related to the dependent variable in univariate analysis (p < 0.05) into the multivariate logistic regression model for control. These confounders included sex (1 = male, 2 = female), age, education level (1 = illiteracy, 2 = primary school, 3 = middle school, 4 = high school, 5 = university), habitation (1 = rural, 2 = urban), occupation (1 = physical labor, 2 = mental labor), income (1 = < 3000, 2 = 3000 , 3 = 5000 , and LSaO, 4 = ≥ 8000), BMI, hypertension (0 = no, 1 = yes), diabetes (0 = no, 1 = yes), coronary disease (0 = no, 1 = yes), hyperlipidemia (0 = no, 1 = yes), family history of dementia (0 = no, 1 = yes), physical exercise (0 = no, 1 = yes), social activity (0 = no, 1 = yes), TST (1 = < 6, 2 = 6 ~ 7, 3 = > 7), MSaO22. Among them, age, BMI, MSaO2, and LSaO2 were continuous variables, while the other variables were categorical variables.
The Hosmer-Lemeshow test was used to assess the model’s goodness of fit; a p-value of less than 0.05 indicated a poor model fit. Every statistical test was two-tailed, and a significance level of p < 0.05 was used. Version 27.0 of IBM SPSS Statistics was applied to analyze these data.
In this study, information was collected from 420 subjects; of all subjects, 11 were on ventilator therapy for OSA, another 3 had severe audiovisual or comprehension impairments, 2 met criteria for dementia as well as Alzheimer’s disease, and 17 subjects were excluded because of incomplete information. Thus, 387 subjects participated in this study. Of the 387 subjects, 147 had MCI and 240 had normal cognition. Tables 1 and 2 summarize the social demography and sleep characteristics of people with OSA, and the study samples are grouped by cognitive status.Table 1Sociodemographic and health characteristics of the final sampleCharacteristicFeaturesOSA subjectsMCI groupNC groupP valueTotal (n = 387)(n = 147)(n = 240)Age, years, mean (SD)45.3 (12.6)52.4 (11.9)41.0 (11.1)< 0.001Sex, n (%)male319 (82.4%)109 (74.1%)210 (87.5%)< 0.001female68 (17.6%)38 (25.9%)30 (12.5%)Education, n (%)illiteracy7 (1.8%)7 (4.8%)0 (0%)< 0.001primary school39 (10.1%)29 (19.7%)10 (4.2%)middle school102 (26.4%)66 (44.9%)36 (15%)high school85 (22%)18 (12.2%)67 (27.9%)university154 (39.8)27 (18.4%)127 (52.9%)Habitation, n (%)rural96 (24.8%)65 (44.2%)31 (12.9%)< 0.001urban291 (75.2%)82 (55.8%)209 (87.1%)Occupation, n (%)physical labor215 (55.6%)113 (76.9%)102 (42.5%)< 0.001mental labor172 (44.4%)34 (23.1%)138 (57.5%)Income, yuan, n (%)< 300082 (21.2%)59 (40.1%)23 (9.6%)< 0.0013000150 (38.8%)59 (40.1%)91 (37.9%)5000107 (27.6%)17 (11.6%)90 (37.5%)≥ 800048 (12.4%)12 (8.2%)36 (15%)Hypertension, n (%)No246 (63.6%)59 (40.1%)187 (77.9%)< 0.001Yes141 (36.4%)88 (59.9%)53 (22.1%)Diabetes, n (%)No343 (88.6%)117 (79.6%)226 (94.2%)< 0.001Yes44 (11.4%)30 (20.4%)14 (5.8%)Coronary disease, n (%)No345 (89.1%)122 (83.0%)223 (92.9%)< 0.001Yes42 (10.9%)25 (17.0%)17 (7.1%)Hyperlipidemia, n (%)No255 (65.9%)74 (50.3%)181 (75.4%)< 0.001Yes132 (34.1%)73 (49.7%)59 (24.6%)Family history of dementia, n (%)No346 (89.4%)118 (80.3%)228 (95%)< 0.001Yes41 (10.6%)29 (19.7%)12 (5%)Neck circumference, cm, mean (SD)41.2 (3.9)41.3 (3.8)41.2 (4.0)0.476BMI, kg/m^2^, mean (SD)29.0 (4.3)29.3 (4.4)28.9 (4.3)0.327Alcohol consumption, n (%)No166 (42.9%)65 (44.2%)101 (42.1%)0.681Yes221 (57.1%)82 (55.8%)139 (57.9%)Smoking, n (%)No238 (61.5%)90 (61.2%)148 (61.7%)0.931Yes149 (38.5%)57 (38.8%)92 (38.3%)Physical exercise, n (%)No207 (53.5%)110 (74.8%)97 (40.4%)< 0.001Yes180 (46.5%)37 (25.2%)143 (59.6%)Social activity, n (%)No87 (22.5%)71 (48.3%)16 (6.7%)< 0.001Yes300 (77.5%)76 (51.7%)224 (93.3%)*BMI *Body mass index, *IQR *Interquartile range, *MCI *Mild cognitive impairment, *n *number, *NC *Normal cognitive, *OSA Obstructive sleep apnea, SD Standard deviationTable 2Sleep-related characteristics of the final sampleCharacteristicFeaturesOSA subjectsMCI groupNC groupP valueTotal(n = 387)(n = 147)(n = 240)Total sleep time, hours, n (%)< 6101 (26.1%)60 (40.8%)41 (17.1%)< 0.0016 ~ 7165 (42.6%)58 (39.5%)107 (44.6%)> 7121 (31.3%)29 (19.7%)92 (38.3%)Severity of sleep apnea, n (%)5 ≤ AHI < 1589 (23%)10 (6.8%)79 (32.9%)< 0.00115 ≤ AHI < 3066 (17.1%)20 (13.6%)46 (19.2%)AHI ≥ 30232 (59.9%)117 (79.6%)115 (47.9%)AHI (events/hour), median (IQR)38.4 (17.5,66.9)53.9 (31.5,82.6)27.4 (12.1,55.3)< 0.001ODI (events/hour), median (IQR)35.3 (14.8,58.9)46.7 (27,74.5)28 (10.5,51.5)< 0.001MSaO2, %, median (IQR)93.3 (91,94.4)92 (90,93.6)93.7 (92,94.8)< 0.001T90, %, median (IQR)7.3 (1.5,27.7)16.8 (3.4,39.2)4.4 (0.7,17.8)< 0.001LSaO2, %, median (IQR)78 (69,84)73 (63,81)81 (72,85)< 0.001ESS, mean (SD)11.5 (4.8)14.1 (4)9.9 (4.5)< 0.001PSQI, mean (SD)8.4 (3.8)10.8 (3.4)7 (3.3)< 0.001AHI *Apnea-hypopnea index, *ESS *Epworth Sleepiness Scale, *IQR *Interquartile range, LSaO2 Lowest oxygen saturation, *MCI *Mild cognitive impairment, MSaO2Mean oxygen saturation, *NC *Mormal cognitive, *n *number, *ODI *Oxygen desaturation index, *OSA *Obstructive sleep apnea, *PSQI *Pittsburgh Sleep Quality Index, *SD *Standard deviation, *T90 *Percentage of time oxygen saturation <90%
There were 387 subjects with OSA (319 male and 68 female; mean age 45.3 ± 12.6 years); the mean age of the MCI group was 52.4 (± 11.9) years, and that of the NC group was 41.0 (± 11.1) years; the MCI group was older compared with the NC group. The group with MCI had lower education (69.4% below high school) and lower monthly income compared to the group with NC (p < 0.05). The prevalence rate of hypertension, diabetes, coronary disease, and hyperlipidemia in people with OSA was 36.4%, 11.4%, 10.9%, and 34.1%, respectively. The group with MCI had a higher prevalence rate of clinical comorbidities compared with the NC group (Table 1).
Total sleep duration < 6 h was discovered in 26.1% of people with OSA, 40.8% in the MCI group, and 17.1% in the NC group. People in the MCI group slept less and had poorer self-reported sleep quality compared to the NC group (7 ± 3.3 vs. 10.8 ± 3.4, p < 0.05). The percentages of mild, moderate, and severe OSA among all subjects with OSA were 23%, 17.1%, and 59.9%, respectively. The median AHI was 38.4/h (IQR, 17.5/h to 66.9/h). The MCI and NC groups had the highest proportions of patients with severe OSA, 79.6% and 47.9%, respectively. The MCI group had more severe hypoxia (28.0 vs. 46.7, 93.7% vs. 92.0%, 4.4% vs. 16.8%, 81% vs. 73%, p < 0.05) and was more likely to have daytime sleepiness compared to the NC group (9.9 ± 4.5 vs. 14.1 ± 4.0, p < 0.05) (Table 2).
When the AHI was used to categorize the severity of sleep apnea, the prevalence rate of MCI was highest in people with severe OSA (50.4%) and lowest in people with mild OSA (11.2%) (Fig. 1). When sleep apnea severity was grouped according to the ODI, the results were similar to those classified according to the AHI; the more severe the degree of sleep apnea, the higher the prevalence of MCI (Fig. 1).Fig. 1The prevalence rates of mild cognitive impairment in people with obstructive sleep apnea with different severities (n = 387). A Severities of OSA were categorized by AHI. B Severities of OSA were categorized by ODI. AHI, apnea-hypopnea index; MCI, mild cognitive impairment; ODI, oxygen desaturation index; OSA, obstructive sleep apnea
In the unadjusted model, the severity of OSA was positively related to MOCA scores < 26 (Table 3). We adjusted for age, sex, education, occupational attributes, monthly income, BMI, clinical comorbidities, physical exercise, social activity, total sleep time (TST), and oxygen saturation. Specifically, in the adjusted model, severe OSA was compared with mild OSA; severe OSA was more highly related to MOCA scores < 26 than mild OSA (odds ratio [OR], 10.562; 95% confidence interval [CI], 2.929-35.095; p < 0.001). We removed TS < 90% from the logistic regression model due to the existence of multicollinearity (VIF > 5). Moderate OSA didn’t differ from mild OSA in the association with MOCA scores < 26. In the unadjusted model, OSA severity assessed using the ODI was positively related to both MOCA scores < 26 (Table 3). In particular, in the adjusted model, severe and moderate OSA were more highly related to MOCA scores < 26 than mild OSA (ORs 5.675 and 4.432, respectively; both p < 0.05).Table 3Logistic regression analysis with mild cognitive impairment as a dependent variable and sleep apnea severity as the independent variable (n = 387)Dependent variableIndependent variableUnadjusted modelAdjusted model ^a^BSEOR95%CIP valueBSEOR95%CIP valueMOCA < 26Defined by AHI ^b^ (reference: mild)Moderate1.2340.4293.4351.480, 7.9690.0041.1860.6683.2730.884, 12.1210.076Severe2.0840.3608.0373.966, 16.289< 0.0012.3570.65410.5622.929, 35.095< 0.001Defined by ODI ^c^ (reference: mild)Moderate1.5600.3884.7612.224, 10.193< 0.0011.4890.5934.4321.386, 14.1730.012Severe1.7590.3275.8043.060, 11.012< 0.0011.7360.6075.6751.725, 18.6670.004*AHI *Apnea-hypopnea index, *MOCA *Montreal Cognitive Assessment, *n *number, *ODI *Oxygen desaturation index^a^Adjusted by age, sex, education, habitation, occupation, monthly income, body mass index, hypertension, diabetes, coronary disease, hyperlipidemia, family history of dementia, physical exercise, social activity, total sleep time, mean oxygen saturation, and lowest oxygen saturation^b^Mild, 5< AHI ≤15; moderate, 15≤ AHI <30; severe, AHI ≥30^c^Mild, 5< ODI ≤15; moderate, 15≤ ODI <30; severe, ODI ≥30
In the unadjusted model, the degree of sleepiness assessed by the ESS was positively related to both MOCA scores < 26 (Table 4). We adjusted for age, sex, education, occupational attributes, monthly income, BMI, clinical comorbidities, physical exercise, social activity, TST, and oxygen saturation. In particular, in the adjusted model, the association with MOCA scores < 26 was higher in the severe sleepiness group than in the non-sleepiness group (odds ratio [OR], 5.722; 95% confidence interval [CI], 2.221-14.746; p < 0.05). Both the moderate sleepiness group and the mild sleepiness group were also associated with MOCA scores < 26 (ORs 2.872 and 3.206, respectively; both p < 0.05).Table 4Logistic regression analysis with mild cognitive impairment as a dependent variable and nighttime sleep quality and daytime sleepiness as the independent variables. (n = 387)Dependent variableIndependent variableUnadjusted modelAdjusted model ^a^BSEOR95%CIP valueBSEOR95%CIP valueMOCA < 26Defined by PSQI ^b^(reference: good sleep group)Average sleep group1.3760.4473.9611.649, 9.5120.0020.5130.6091.3750.220, 2.4020.106Poor sleep group3.1020.42422.2519.699, 51.047< 0.0011.3970.6004.0411.247, 13.0930.020Defined by ESS ^c^(reference: non-sleepiness group)Mild sleepiness group1.5470.4014.6982.142, 10.304< 0.0011.1650.5563.2061.079, 9.5300.036Moderate sleepiness group1.7360.2995.6723.155, 10.199< 0.0011.0550.4792.8721.123, 7.3480.028severe sleepiness group2.2460.3009.4505.254, 16.998< 0.0011.7440.4835.7222.221, 14.746< 0.001*ESS *Epworth Sleepiness Scale, *MOCA *Montreal Cognitive Assessment,, *n *number, *PSQI *Pittsburgh Sleep Quality Index Scale^a^Adjusted by age, sex, education, habitation, occupation, monthly income, body mass index, hypertension, diabetes, coronary disease, hyperlipidemia, family history of dementia, physical exercise, social activity, total sleep time, mean oxygen saturation, and lowest oxygen saturation^b^good sleep group, PSQI scores ≤5; Average sleep group, 6≤ PSQI scores ≤8; Poor sleep group, PSQI scores ≥9^c^non-sleepiness group, ESS scores ≤10; Mild sleepiness group, 11≤ ESS scores ≤12; Moderate sleepiness group, 13≤ ESS scores ≤15; severe sleepiness group, ESS scores ≥16
In the unadjusted model, the different sleep quality groups assessed by the PSQI were all positively related to MOCA scores < 26 (Table 4). We adjusted for age, sex, education, occupational attributes, monthly income, BMI, clinical comorbidities, physical exercise, social activity, TST, and oxygen saturation. Specifically, in the adjusted model, only the poor sleep group was positively related to MOCA scores < 26. The poor sleep group had a higher association with MOCA scores < 26 than the good sleep group (odds ratio [OR], 4.041; 95% confidence interval [CI], 1.247-13.093; p < 0.05).
The study aimed to determine whether sleep apnea severity was related to MCI and whether daytime sleepiness and nighttime sleep quality were related to MCI. The results found that after adjusting for a range of confounders, severe OSA, poor sleep quality, and sleepiness were independent risk factors for developing MCI in people with OSA. The outcomes of the current study may provide ideas for the diagnosis and therapy of MCI in people with OSA, i.e., whether the severity of OSA, sleepiness, and sleep quality can be used as potential risk predictors for MCI and whether cognitive function can be improved by improving the severity of OSA, sleepiness, and nighttime sleep quality.
MCI (defined as a MOCA score < 26) was present in 38% of OSA patients in this study. Because non-OSA patients were not included, it is not possible to conclude that the OSA population is more likely to develop MCI, and the higher rate of MCI in this group may partly reflect demographic factors such as higher age and lower education level, as well as more comorbidities. Three groups of patients were created for the study based on AHI: mild, moderate, and severe, and the prevalence rate of MCI was 11.2%, 30.3%, and 50.4%, respectively. The sleep apnea severity in people with OSA affects the prevalence of MCI. The prevalence rate of MCI increased significantly with the increasing severity of OSA. In this study, people having severe OSA were far more likely to develop MCI compared to people having mild OSA. This positive association persisted after adjusting for a series of confounding factors, such as age, sex, education, occupational attributes, monthly income, BMI, clinical comorbidities, physical exercise, social activity, total sleep time, and oxygen saturation. Notably, a population-based study also reported an association between severe OSA and poorer cognitive outcomes [27]. The findings of Román et al. confirmed the relationship between PSG-related markers and cerebrospinal fluid markers in people with MCI [28], determining that higher AHI was related to higher cerebrospinal fluid levels of p-tau and t-tau, and this association persisted after adjusting for confounding factors, similar to our findings. In addition, Arslan assessed plasma NfL levels in adults with OSA for the first time and found a direct correlation between ODI and NfL levels in the moderate-to-severe OSA group [29], whereas NfL is a marker of neural axonal injury, which can be released into the bloodstream and cerebrospinal fluid after the injury and can be applied to evaluate cognitive impairment [30]. The above studies illustrate the correlation of OSA with markers of MCI, and this association is related to the severity of OSA. This is because people with severe OSA experience more apnea and hypoventilation events at night, triggering hypoxia and more frequent awakenings. For every 0.01 unit increase in sleep fragmentation, the annual decline rate of cognitive function increased by 22% [31]. Studies have shown that sleep fragmentation can cause Aβ protein deposition, astrocyte activation, etc., resulting in neurodegenerative diseases [32, 33]. However, the cross-sectional study conducted by Dlugaj et al. didn’t find the contact between severe sleep apnea (AHI ≥ 30/h) and MCI and its subtypes [34], which may be related to the generally young age of the study cohort and the cognitive testing methodology used. Therefore, it is particularly important to use a more comprehensive cognitive testing approach to observe the connection between the severity of OSA and MCI in patients of different ages. In addition, we considered classifying the severity of sleep apnea based on blood oxygenation status, and the ODI, as an indicator based on changes in oxygen saturation, can provide a direct and objective assessment of nocturnal oxygenation status [35], compensating for the shortcomings posed by the traditional indicator for classifying the severity of OSA. The model was adjusted to find that moderate and severe OSA were positively related to MCI. Many studies have also shown that the associated hypoxemic symptoms due to OSA are closely related to cognitive decline [37, 38]. Pathophysiologic changes triggered by hypoxia are an important mechanism for OSA-associated cognitive dysfunction. Furthermore, OSA also promotes cognitive impairment through vascular brain injury (blood-brain barrier disruption, white matter lesions, abnormal cerebrovascular regulation) in conjunction with hypoxia and sleep fragmentation [39, 40]. Further studies are recommended in the future to assess the relationship between different severities of OSA determined by other hypoxia burden indicators and MCI.
We assessed sleep quality in people with OSA by PSQI, and patients in the MCI group showed worse sleep quality. The adjusted model indicated that poor sleep quality could become a potential risk marker for early screening of MCI in people with OSA. Clinically, people with OSA often suffer from frequent nighttime awakenings and sleep fragmentation, which can cause decreased sleep quality in people with OSA, causing daytime sleepiness and decreased attention, which can impair cognitive function over time. Several studies have discovered that poorer sleep is related to reduced executive function and slower disposal speed [41, 42]. Dlugaj found that in the general population, poor sleep quality was related to overall MCI, independent of traditional cardiovascular risk factors [34]. Especially, difficulty falling asleep and difficulty awakening in the morning were related to the amnestic MCI subtype, whereas difficulty maintaining sleep was related to the non-amnestic MCI subtype [34]. The mechanisms by which OSA patients with poor sleep easily trigger MCI are not fully clear. Hypoxia seems, at least to some extent, likely to generate these cognitive changes [43, 44]. Repeated airway obstruction and collapse during sleep in people with OSA cause chronic intermittent hypoxia, which induces a series of pathophysiological changes that produce inflammation, blood-brain-barrier damage, and oxidative stress, leading to structural and microvascular brain pathological changes and thus cognitive dysfunction [45]. On the one hand, chronic intermittent hypoxia can activate astrocytes and microglia directly to cause a central nervous system inflammatory response. On the other hand, it can cause the release of peripheral inflammatory factors, impair the blood-brain barrier, and stimulate the vagus to enter the center, causing synaptic damage and loss, and necrosis and apoptosis of neurons [46, 47]. In addition, hypoxia promotes Aβ protein deposition and tau protein hyperphosphorylation, accelerating the transition to dementia [45]. We also observed that the MCI group’s T90 was significantly higher than the NC group’s and that the two groups’ mean and lowest oxygen saturation levels differed significantly, with the MCI group exhibiting more severe hypoxia. To overcome the adverse impacts of poor sleep quality on cognitive functioning, it is recommended that other factors affecting sleep quality be investigated and that sleep quality be assessed at follow-up after appropriate treatment of OSA.
EDS is a major clinical symptom of OSA, which is considered to be relevant to cognitive dysfunction [48, 49]. In clinical practice, for OSA individuals with EDS, sleep doctors will focus on the impact of the destruction of neurophysiological activities on cognitive functioning during sleep, which is also a research hotspot. Most of the current research mainly focuses on whether or not EDS exists, which has not differentiated between the severity of EDS. In the present study, we discovered that the severity of sleepiness was positively related to MCI, and after adjusting for various confounders, sleepiness remained a significant risk factor for MCI. Neuroimaging research has observed variation in the grey and white matter of the cerebrum in individuals with OSA and EDS [50], but the exact mechanisms remain unknown. Sleep fragmentation and chronic intermittent hypoxia were thought to induce oxidative damage and alterations in the connections between neurons and brain circuits, involving neurotransmission of noradrenergic and dopaminergic. Dopaminergic controls arousal regions in the brain. Our results show that sleepiness affects cognitive function independently when assessing and predicting the potential value of cognitive function. Sforza et al. also discovered that subjective EDS was related to bilateral hippocampal volume reduction regardless of the severity of AHI or nocturnal hypoxia-related parameters [51], which also provides significant evidence that EDS affects patients’ cognitive function independently of OSA, and provides important evidence. Therefore, it is important to increase the importance of sleepiness and cognitive function in the clinic and then to avoid the risks and improve the quality of life, which is something that needs to be told to patients.
In this study, severe OSA, poor sleep quality, and sleepiness were risk factors for MCI, so reducing AHI, improving sleep quality, and diminishing daytime sleepiness probably ameliorate cognitive function in OSA patients. Earlier studies showed that CPAP therapy can help people with OSA effectively reduce AHI, decrease daytime sleepiness, improve alertness and cognition, and improve quality of life [52, 53]. A study exploring the cognitive impact of CPAP was carried out by Wang et al. They discovered that CPAP treatment can significantly improve attention and information processing speed, partially improving cognitive deficits in individuals with severe OSA [54]. Furthermore, the meta-analysis demonstrated that OSA patients’ subjective and objective somnolence was improved by CPAP therapy. Patients who had moderate-to-severe OSA and baseline ESS scores ≥ 11 experienced a significant decline in ESS scores after CPAP therapy, suggesting that high ESS scores are a significant predictor of benefit from CPAP therapy in these patients [55]. However, there have also been multicenter trials showing no effect [56, 57] or only slight improvements in cognitive function with CPAP treatment. [58, 59]. However, it should be mentioned that good adherence is expected to increase the benefits of CPAP [57, 59]. The outcomes of this study may have potential clinical value, suggesting that patients with OSA should be actively engaged and adhere to CPAP therapy over the long term to reduce the degree of OSA, improve the quality of nighttime sleep, alleviate the symptoms of daytime sleepiness, and reduce risk factors for the development of MCI.
There are various restrictions on this study. First, the AHI has limitations in evaluating OSA because it only reflects the number of occurrences of apnea and hypoventilation in slumber, overlooking the duration and severity of hypoxia. As a result, the AHI is unable to accurately catch the real pathological characteristics of OSA. Second, we didn’t get detailed information about various psychiatric illnesses and medications that may occur at the same time as OSA, and the use of certain drugs may aggravate sleep apnea. In addition, the confounding factors we adjusted were limited, and the possibility of residual confusion still exists. Third, moderated effects analyses were not conducted, and effect heterogeneity may be present. Fourth, the study lacked a non-OSA control group; therefore, it was not possible to determine whether the non-OSA population also had a higher prevalence of MCI. Finally, the assessment of MCI is based on a screening tool rather than a clinical diagnosis, and using only a single MOCA threshold (< 26) to assess overall cognitive functioning increases the risk of false positives, especially when the effects of age and education are not taken into account, which may incorrectly categorize normal aging as MCI. However, to date, there are no uniform diagnostic criteria for MCI, and the MOCA, with its good sensitivity and specificity, is a measurement tool developed specifically for screening MCI and has been widely used in population-based studies of MCI. It is recommended that future studies use more comprehensive cognitive assessments.
In brief, 38% of individuals with OSA suffered from MCI in this study. The sleep apnea severity assessed by ODI was positively related to mild cognitive impairment. In addition, there was a positive relation between severe OSA assessed by AHI, poor sleep quality, sleepiness, and increased risk of MCI. The results of this study deepen the understanding of patients with OSA comorbid with MCI and may be of great help in implementing preventive measures and improving clinical management and treatment decisions.