Associations between daytime and nighttime plasma orexin A levels and cognitive function in patients with obstructive sleep apnea

Abstract

The relationship between plasma orexin A (OXA) levels and cognitive function in patients with obstructive sleep apnea (OSA) remains unclear. This study aimed to evaluate associations between daytime and nighttime plasma OXA levels and cognitive function in patients with OSA. Subjects with suspected OSA underwent overnight polysomnography (PSG), Montreal Cognitive Assessment (MoCA), and Epworth Sleepiness Scale (ESS) assessment. Subjects were considered controls or having OSA according to the apnea–hypopnea index (AHI). Daytime and nighttime plasma OXA levels were determined by ELISA. Receiver-operating characteristics curves were used to evaluate the diagnostic value of plasma OXA levels for assessing cognitive impairment in OSA patients. One hundred and six subjects met the inclusion criteria. MoCA scores and plasma OXA concentrations were significantly lower in OSA patients than controls (p < 0.01). Patients with moderate and severe OSA had significantly lower MoCA scores than controls and mild OSA patients (p < 0.01). Daytime and nighttime OXA levels were significantly lower in OSA patients with cognitive impairment than those without cognitive impairment (p < 0.01). Both daytime and nighttime plasma OXA levels in patients with OSA were positively correlated with MoCA scores and nadir SaO₂, negatively correlated with AHI, oxygen desaturation index, and percentage of time spent with an SaO₂ below 90% (all p < 0.05), and not correlated with ESS scores. The optimal threshold of daytime plasma OXA to diagnose OSA with cognitive impairment was 49.34 pg/ml, with a sensitivity of 80.0% and a specificity of 74.3%. We concluded that plasma OXA concentrations might be related to cognitive function and daytime plasma OXA levels have diagnostic value for assessing cognitive impairment in OSA patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41105-022-00387-4.

Introduction

Obstructive sleep apnea (OSA) is a breathing disorder characterized by episodes of upper airway obstruction or collapse resulting in repeated hypoxia and reoxygenation during sleep. OSA is frequently accompanied by excessive daytime sleepiness and disordered sleep architecture that results in cognitive dysfunction and chronic diseases such as hypertension and diabetes [1, 2]. The negative effects of OSA on cognitive function mainly manifest as impairments in attention and executive function [3]. Moreover, the mild cognitive impairment resulting from OSA can be difficult for patients to detect subjectively. One meta-analysis reported a 26% cognitive decline in adult patients with OSA [4]. The risk of cognitive impairment in OSA patients increases with age and, moreover, OSA is also a risk factor for dementia [5]. Therefore, the early detection and treatment of cognitive impairment in OSA remains a clinical problem.

Orexins, including orexin A (OXA) and OXB, are synthesized solely by neurons located in the lateral hypothalamus (LH) [6]. Orexin neuron projections target the entire brain except for the cerebellum and play important physiological roles [6]. Orexin neuron activity is controlled by the circadian clock in the suprachiasmatic nucleus [7]. OXA passes through the blood–brain barrier to distribute to human peripheral tissues and plasma [8]. In animal models, cerebrospinal fluid (CSF) orexin levels are low just after waking but gradually increase with daytime activity before decreasing again at night, suggesting that orexin levels might be related to the day length [9]. Many studies have now shown that the orexin system plays an important role in maintaining wakefulness, and in its most severe form, the loss of orexin neurons can lead to narcolepsy with cataplexy [10, 11]. Indeed, low CSF orexin levels have been used clinically as a biomarker of narcolepsy type 1 [12].

The data on interactions between OXA and neurocognitive function are conflicting. Liguori et al. found that increased OXA levels in the CSF were associated with deteriorated sleep quality and cognitive decline in patients with Alzheimer’s disease [13]. Other studies have suggested that OXA might protect against cognitive decline [14]. The cognitive impairment in OSA may be related to sleep fragmentation and nocturnal hypoxia during sleep, whereas OXA is involved in the regulation of sleep and wakefulness [1, 11]. Plasma OXA levels in OSA patients have been reported to be significantly higher than in control patients and positively associated with arousal index and the apnea–hypopnea index (AHI) [15], although some studies have found the opposite [16, 17]. Plasma OXA levels can be affected by environmental factors such as diet, smoking, and the time of blood collection [18, 19]. Therefore, the relationship between plasma OXA levels and cognitive function in OSA patients remains unclear, and a full understanding of the temporal dynamics of plasma OXA concentrations would be useful for developing non-invasive but accurate biomarkers of cognitive decline in OSA.

Here, we investigated potential interactions between OXA levels, cognitive function, and OSA severity by evaluating plasma OXA levels at different time points and their association with Montreal Cognitive Assessment (MoCA) scores and sleep-related parameters in patients with OSA.

Methods

Subjects

Subjects with suspected OSA (snoring symptoms) were recruited from the outpatient department of Renmin Hospital of Wuhan University between October 2019 and November 2020. This study procedure was reviewed and approved by the Ethics Committee of Renmin Hospital of Wuhan University (WDRY2021-K019). Written informed consent was obtained from each participant. The inclusion criteria of participants were as follows: (i) all participants were at least 18 years of age and were not pregnant; (ii) participants had normal lung function. Participants were excluded if they had a history of bronchial asthma, diabetes, mental illness, non-OSA disorders affecting sleep or cognitive impairment, and renal dysfunction. Participants who refused to undergo polysomnography (PSG), MoCA testing, or have two intravenous blood samples taken were excluded. Overall, 106 subjects met the inclusion criteria for the study and were classified into two groups: a control group (AHI < 5 events/hour) and the OSA group (AHI ≥ 5 events/hour) (Fig. 1).

Fig. 1.

Study flow diagram. OSA obstructive sleep apnea, PSG polysomnography, MoCA Montreal Cognitive Function Assessment, AHI apnea-hypopnea index

Epworth sleepiness scale (ESS)

ESS is a validated instrument used to assess daytime sleepiness [20]. The scale includes eight scenarios for which subjects must rate their likelihood of falling asleep. The higher the ESS score, the more likely the subject is to experience daytime sleepiness. ESS questionnaires were completed by subjects prior to PSG.

PSG

PSG is indispensable for the diagnosis of OSA. PSG was performed by a trained professional. OSA was diagnosed as an AHI ≥ 5 events/hour during sleep by the overnight PSG (SOMNOscreen Plus Tele PSG, SOMNOmedics GmbH, Randersacker, Germany) according to American Academy of Sleep Medicine guidelines [21]. The AHI, average oxyhemoglobin saturation (SaO₂), nadir SaO₂, percentage of time spent with a SaO₂ below 90% (TS90%), and oxygen desaturation index (ODI) were recorded from PSG.

Assessment of cognitive function

Cognitive function was assessed in a quiet and comfortable environment 1 h before receiving PSG by the same professional using the MoCA scale, which is widely used to screen for cognitive dysfunction and has a high sensitivity and specificity [22]. The scale includes visuospatial and executive ability, naming, memory, attention, language, abstraction, delayed recall, and orientation. The total MoCA score is 30, but if the participant has less than 12 years of education, an extra point is added to the score. An MoCA score < 26 defined cognitive impairment [23].

Collection of intravenous blood and isolation of plasma

Three milliliter samples of peripheral blood were obtained from each subject at nighttime (21:00–22:00) and during the day (06:00–07:00). Plasma samples were collected by centrifugation and preserved at − 80 °C until analysis.

Detection of OXA by enzyme-linked immunosorbent assay (ELISA)

OXA protein concentrations in plasma were measured with an ELISA kit following the manufacturer’s instructions (Jiangsu Meimian Industry Co., Ltd, Nanjing, China). The detection limit was 1 pg/ml. Intra-and inter-assay coefficients of variation for all ELISAs were < 10% and < 12%, respectively. All samples were measured in duplicate.

Statistical analysis

Quantitative data with normal distributions were expressed as means ± standard deviations (SD), and one-way analysis of variance (ANOVA) was used to compare differences between groups. Non-parametric data were expressed as medians (interquartile range) and were analyzed with the Kruskal–Wallis test. The paired t test was used to compare daytime and nighttime data. Spearman’s correlation coefficients were used to assess correlations between the two quantitative variables. Qualitative data were compared between groups using Fisher’s exact probability method. Diagnostic tests were evaluated by receiver-operating characteristic (ROC) analysis. The maximum Youden index value was used as a cut-off, where the Youden index was calculated as (sensitivity + specificity-1). A p value < 0.05 was considered statistically significant. All analyses were tested using SPSS 19.0 (IBM Statistics, Chicago, IL, USA).

Results

Baseline data and clinical characteristics

The baseline characteristics of the study subjects are shown in Table 1. One hundred and six subjects were eligible to participate in the study, 28 participants in the control group and 78 in the OSA group. There were no significant differences in age, gender, body mass index (BMI), smoking status, alcohol consumption, or FEV₁/FVC between groups (p > 0.05). Neck circumference, waist circumference, AHI, ODI, ESS scores, and hypertension status were significantly higher in the OSA group than in the control group, while the nadir SaO₂, average SaO₂, MoCA score, and plasma OXA concentration were significantly lower in the OSA group than in the control group (all p < 0.01).

Table 1.

Baseline characteristics of the subjects (n = 106)

Characteristic	Control group (n = 28)	OSA group (n = 78)	P value
Sex, male, n (%)	17 (60.7)	54 (69.2)	0.28
Age, yr	55.54 ± 8.94	56.55 ± 10.9	0.66
BMI, kg/m2	26.2 ± 3.9	26.9 ± 3.2	0.35
Neck circumference, cm	36.0 ± 3.3	38.1 ± 3.1	0.01
Waist circumference, cm	91.0 ± 11.6	96.2 ± 7.9	0.02
Smoking status, n (%)			0.90
Never	15 (53.6)	38 (48.7)
Former	4 (14.3)	13 (16.7)
Current	9 (32.1)	27 (34.6)
Alcohol consumption, n (%)			0.74
Never	17 (60.7)	41 (52.5)
Former	2 (7.2)	8 (10.3)
Current	9 (32.1)	29 (37.2)
FEV1/FVC% Predicted	82.3 ± 7.41	82.77 ± 6.83	0.81
AHI, events/h, median (Q1,Q3)	3.4 (1.6, 4.4)	25.8 (13.8, 43.3)	< 0.01
ODI, median (Q1,Q3)	3.4 (1.7, 4.4)	20.5 (10.8, 32.4)	< 0.01
TS90%, median (Q1,Q3)	0.3 (0.0, 2.0)	6.15 (1.0, 13.0)	< 0.01
Nadir SaO2,%	86.21 ± 6.02	75.94 ± 11.35	< 0.01
Average SaO2,%	95.54 ± 0.92	93.29 ± 2.32	< 0.01
ESS	7.14 ± 4.32	12.33 ± 5.19	< 0.01
Hypertension, n (%)	12 (42.9)	52 (66.7)	< 0.01
MoCA (score)	28.64 ± 1.10	25.62 ± 2.10	< 0.01
Orexin A(pg/ml)
Daytime#	71.00 ± 8.19	50.66 ± 10.94	< 0.01
Nighttime&	65.51 ± 7.37	49.25 ± 9.51	< 0.01

Continuous variables which are obviously skewed are presented as median (Q1,Q3), other continuous variables are described as means ± SD. Categorical variables are expressed as percentages. P value for control group vs. OSA group

OSA Obstructive sleep apnea, FEV₁ forced expiratory volume in the first second, FVC forced vital capacity, AHI apnea-hypopnea index, ODI Oxygen decay index, SaO₂ Oxygen saturation, TS90% percentage of time spent with a SaO₂ below 90%, ESS Epworth sleep scale, MoCA Montreal Cognitive Function Assessment

^#Presents number of daytime subjects including control group (n = 27) and OSA group (n = 75); ^&presents number of nighttime subjects including control group (n = 22) and OSA group (n = 74)

Differences in daytime and nighttime plasma OXA levels

OXA levels in the CSF are regulated by circadian rhythms. Here, plasma OXA concentrations during the day were significantly higher than at nighttime in control group (71.11 ± 7.50 pg/ml vs. 64.43 ± 5.48 pg/ml, p < 0.01), while no difference was observed in the OSA group (50.71 ± 11.13 pg/ml vs. 49.42 ± 9.26 pg/ml, p > 0.05) (Fig. 2).

Fig. 2.

Differences in daytime and nighttime plasma OXA levels. The paired t test was used for analysis. A Changes in daytime and nighttime plasma OXA levels in controls (n = 21). B Change in daytime and nighttime plasma OXA levels in OSA patients (n = 71). *p < 0.01 for daytime vs. nighttime

Changes in OXA protein levels and MoCA scores in the OSA subgroup

We next assessed the differences in plasma OXA protein levels and MoCA scores according to OSA severity. Daytime plasma OXA concentrations decreased as the severity of OSA increased (p < 0.05). Nighttime plasma OXA concentrations in patients with mild OSA were lower than in normal patients but higher than in patients with moderate and severe OSA (p < 0.05). There was no significant difference in nighttime plasma OXA levels in patients with moderate and severe OSA (p > 0.05) (Fig. 3).

Fig. 3.

Differences in daytime and nighttime plasma OXA levels in patients with OSA. Results were analyzed by one-way ANOVA. A Differences in daytime plasma OXA levels between groups. B Differences in nighttime plasma OXA levels between groups. *p < 0.05 vs. other subgroups. ^#p < 0.05 vs. control and mild OSA group

The Kruskal–Wallis test was used to compare differences in cognitive scores in OSA subgroups, because the data were not normally distributed. As shown in Fig. 4, MoCA scores in patients with moderate and severe OSA were significantly lower than those in controls and patients with mild OSA (p < 0.01), but there was no statistically significant difference in MoCA scores between patients with moderate and severe OSA (p > 0.05).

Fig. 4.

Comparison of cognitive scores between daytime and nighttime subjects in the four groups. Results were analyzed using the Kruskal–Wallis test. A Comparison of cognitive scores between daytime subjects in the four groups (n = 75). B Comparison of cognitive scores between nighttime subjects in the four groups (n = 74). *p < 0.01 vs. the control group and mild OSA group

Relationship between OXA levels and cognitive function in OSA patients

Patients with MoCA score less than 26 were considered to have. OSA patients with cognitive impairment had significantly lower daytime and nighttime plasma OXA levels than in OSA patients without cognitive impairment (p < 0.01) (Fig. 5A, B). In OSA patients with cognitive impairment, there was a significant positive correlation between daytime plasma OXA concentrations and MoCA scores (r = 0.70, p < 0.01). However, there was no significant correlation between nighttime plasma OXA concentrations and MoCA scores (r = 0.25, p = 0.11) (Fig. 5C, D).

Fig. 5.

Differences in OXA levels in OSA patients with and without cognitive impairment and their association with cognitive scores. Student’s t test was used to compare the two groups. Spearman’s test was used for correlation analysis. A, B Differences in daytime and nighttime plasma OXA levels in OSA patients with or without cognitive impairment; C, D Correlation between daytime and nighttime OXA concentrations and cognitive scores in OSA patients with cognitive impairment

Correlations between OXA concentrations, MoCA scores, and sleep parameters in OSA patients

The relationship between plasma OXA concentrations, MoCA scores, and sleep parameters in OSA patients was assessed by Spearman’s test. Daytime plasma OXA concentrations were positively correlated with MoCA scores and nadir SaO₂ (r = 0.68, r = 0.63, p < 0.01) but negatively correlated with AHI, ODI, and TS90% (r = − 0.72, r = − 0.48, r = − 0.50, p < 0.01) (Supplementary Fig. 1). Interestingly, however, there was no association between plasma OXA and ESS scores in OSA patients (r = − 0.16, p = 0.16). Similarly, nighttime plasma OXA levels were also associated with MoCA scores and sleep parameters (p < 0.05), but with lower correlation coefficients than daytime values (Supplementary Fig. 2). Finally, there were no correlations between plasma OXA and BMI and average SaO₂ in OSA patients. These results suggested that OXA was associated with cognitive function and severity of OSA. Moreover, it might not be affected by BMI and ESS score.

Diagnostic value of daytime plasma OXA for assessing cognitive impairment

An ROC curve was used to evaluate the diagnostic value of daytime plasma OXA concentrations for assessing cognitive impairment. As shown in Fig. 6, our results showed that the area under the ROC curve (AUC) was 0.783 (95% CI 0.68–0.89; p < 0.01). According to the highest Youden index, the best diagnostic threshold of daytime plasma OXA for cognitive impairment in OSA was 49.34 pg/ml, with a sensitivity of 80% and a specificity of 74.3%.

Fig. 6.

ROC curve of OXA for diagnosing cognitive impairment in OSA patients

Discussion

The pathophysiological basis of OSA is chronic intermittent hypoxia resulting in multi-system dysfunction including hypertension and cognitive impairment [1, 2]. In this study, we confirmed that the probability of having hypertension was significantly higher in OSA patients than in controls. Furthermore, OSA patients had significantly lower cognitive scores than control patients. Mechanistically, cognitive impairment might be related to impairments in oxidative stress caused by chronic intermittent hypoxia [24]. Obesity and old age are not only risk factors for OSA but could also affect OXA production from orexin neurons [25, 26]. However, we found that plasma OXA levels were not correlated with BMI and age in OSA patients (data not shown), perhaps due to differences in our study population and those in the previous studies [25, 26].

Reported plasma OXA levels in OSA patients have been inconsistent in previous studies, perhaps due to the detection methods used and the time point of blood collection [18]. Here, blood samples were collected at two time points (daytime and nighttime), and plasma OXA concentrations were measured using the currently recommended ELISA method rather than the previously used radioimmunoassay [27]. We found that daytime OXA levels were higher than nighttime OXA levels in controls, which seemed to correspond to the circadian rhythm. However, this cyclical expression was significantly weakened in the OSA group. This might be because OXA expression changes in OSA patients prior to sleep to prevent apnea during sleep. However, in a study of nine young healthy males, plasma OXA concentrations did not change according to the circadian rhythm [28]. In our study, the inclusion criteria did not control for gender and age, which may have affected OXA secretion. Similar to our results, Salomon et al. reported that the circadian amplitude of CSF OXA levels in depressed patients was significantly lower than in controls [29]. Collecting blood samples from OSA patients over more time points would be useful to accurately assess cyclical changes in OXA levels.

OXA levels in the OSA group were significantly lower than those in controls, consistent with the results of previous studies [16, 30]. Reduced CSF OXA levels are diagnostic criteria for narcolepsy type 1 [12]. Higuchi et al. found that plasma orexin levels were significantly lower in narcolepsy patients than in healthy subjects [31]. Furthermore, a strong association between CSF and plasma OXA levels has been reported in traumatic stress disorder and healthy subjects [32]. These results seem to support the hypothesis that OXA levels in the peripheral blood reflect CSF OXA levels. Because it has been reported that some of the peripheral tissues can also secret OXA, such as from male reproductive organs, pancreas, and adrenal gland [33, 34], it is possible that plasma OXA may not reliably reflect central nervous system (CNS) OXA levels. However, OXA is detectable in low concentrations in most human samples [28], and OXA is produced throughout the CNS and can cross the blood brain barrier [8], low amounts of OXA could diffuse from brain tissues or CSF to the plasma. In our subgroup analysis, we found that daytime plasma OXA concentrations decreased with increasing OSA severity. Nighttime plasma OXA concentrations in patients with mild OSA were obviously higher than those of patients with moderate and severe OSA, but there were no differences in nighttime plasma OXA levels in moderate and severe OSA patients. Hypoxemia occurring during sleep and enhancement of sympathetic activity in patients with moderate and severe OSA chronically overstimulates sympathetic nerves and exhausts orexin neurons, further reducing OXA secretion [35, 36]. In addition, a negative feedback mechanism counteracting these stresses might inhibit OXA secretion in patients with moderate and severe OSA. However, the situation might be more disordered at night, with sleep fragmentation partly increasing OXA secretion in patients with severe OSA. Our correlation analysis also supported the notion that plasma OXA levels are related to AHI and the degree of nocturnal hypoxia, suggesting that plasma OXA levels may be a biomarker of OSA severity, especially daytime plasma OXA, consistent with the findings of Busquet et al. [37]. While ESS reflects excessive daytime sleepiness in OSA patients, we found that plasma OXA levels were not related to ESS scores, perhaps because ESS is a subjective self-reporting test. Moreover, it has also been reported that ESS is not an accurate screening tool for OSA [38, 39].

We found that cognitive impairment was mainly present in patients with moderate-to-severe OSA. Our results are consistent with recent reports that moderate-to-severe OSA patients had more than 55.3% mild cognitive impairment [23], and indeed, moderate-to-severe OSA has been shown to be an independent risk factor for mild cognitive impairment [23]. Although the effect of OXA on neurocognition is controversial, we found that plasma OXA levels were positively correlated with cognitive function in OSA patients, suggesting that OXA may protect cognitive function in OSA. This might be because: (i) OXA has important anti-inflammatory, anti-oxidant, and immune regulatory functions that protect neurons associated with cognitive function (such as hippocampal neurons) [40]. (ii) OXA might activate other neurons such as cholinergic neurons involved in cognitive protection. (iii) The increased levels of OXA prevent apnea and severe hypoxia during sleep. Our previous study found that exogenous administration of OXA ameliorates cognitive impairment caused by chronic intermittent hypoxia in mice [41]. In addition, Tsuchimine et al. found that plasma OXA levels were significantly reduced in patients with bipolar disorder [42], and Steward et al. found that reduced plasma OXA levels were associated with cognitive deficits in anorexia nervosa [43]. Furthermore, we found that daytime OXA levels in OSA patients with cognitive impairment were positively correlated with cognitive scores, while nighttime OXA levels were not, perhaps because OXA secretion at night was more easily influenced by factors such as activity before falling asleep, mood swings, and insomnia.

Some clinical biomarkers of cognitive impairment in OSA have been reported. Serum insulin-like growth factor-1 was associated with cognitive impairment in OSA patients, and continuous positive airway pressure (CPAP) improved cognitive function in OSA patients by increasing serum levels of insulin-like growth factor-1 [44]. Our results suggested that plasma OXA levels may have diagnostic value for assessing cognitive impairment in OSA patients. CPAP not only improved the symptoms of daytime sleepiness in OSA patients but also cognitive impairment, which further improved after 2–3 months of treatment [45]. Furthermore, Sakurai et al. reported that CPAP for 3–6 months improved decreases in plasma OXA levels in patients with severe OSA [30]. These studies also support the hypothesis that OXA levels might be a biomarker of cognitive impairment in OSA, but larger samples that allow subgroup analysis of individual differences and disease progression would be helpful for further biomarker development.

This study has some limitations. This was a cross-sectional study, so causality cannot be inferred. Although we applied strict inclusion and exclusion criteria, our results must be interpreted with caution, because peripheral plasma OXA levels may be affected by many factors. Finally, the overall sample size in this study was relative small, and larger, prospective, multi-center cohorts should be analyzed to clarify the relationship between plasma OXA levels and cognitive function by multiple regression analysis.

Conclusion

In conclusion, plasma OXA concentrations might be related to cognitive function and daytime plasma OXA levels had diagnostic value for assessing cognitive impairment in OSA patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

JZ and ZZ contributed equally to this work. The idea for the article was raised by JZ and KH. The scales and PSG were completed by JZ, MX, and HM. Subjects’ blood samples were collected and processed by ZZ. The data were collected by ZZ and MJ. The draft of the manuscript was written by JZ and ZZ. KH revised the manuscript and supervised the work.

Funding

This work was supported by research grants from the National Natural Science Foundation of China (81770089, 81970082).

Availability of data and materials

All data generated or analyzed during this study are included in the submitted article and its supplementary files.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

The research protocol was approved by the Ethics Committee of Renmin Hospital of Wuhan University and have been performed in accordance with the 1964 Helsinki declaration and its later amendments.

Informed consent

Informed consent was obtained from all individual participants included in the study.