Altered cerebral blood flow and white matter during wakeful rest in patients with obstructive sleep apnea: a population-based retrospective study

Abstract

Objectives:

To explore changes in cerebral blood flow (CBF) and white matter during wakeful rest in patients with obstructive sleep apnea (OSA).

Methods:

The subjects comprised OSA patients and age- and sex-matched non-sleep apnea (NSA) subjects from December 2020 to December 2021. All subjects underwent structural and arterial spin labeling MRI examinations using a 3.0 T MRI scanner. Intergroup differences in regional and global CBF and white matter hyperintensities (WMHs) were analyzed.

Results:

In this study, 100 (74 males) of 750 (439 males) subjects were diagnosed with OSA, so the prevalence of OSA in the general population was 13.3% (100/750), with 16.9% (74/439) in males and 8.4% (26/311) in females. Excluding four patients with incomplete imaging data, 96 OSA patients and 103 age- and sex-matched NSA subjects were included. At global level, OSA patients showed significantly decreased CBF values in gray matter and whole brain compared to NSA subjects (gray matter: p = 0.010; whole brain: p = 0.021). No significant difference in CBF values was found in WM between the two groups (p = 0.250). At regional level, compared with NSA subjects, patients with OSA exhibited significantly decreased regional CBF values mainly in right parietal lobe and right temporal lobe. Moreover, OSA patients had significantly higher WMHs burden than NSA subjects (p = 0.017).

Conclusions:

OSA patients exhibit decreased global and regional CBF values and increased WMHs burden.

Advances in knowledge:

These findings provide a basis for exploring neuropathological changes of OSA and for early and appropriate treatment.

Introduction

Obstructive sleep apnea (OSA) is a chronic respiratory disorder characterized by partial (hypopnea) or complete (apnea) upper airway obstruction with associated intermittent hypoxemia or awakening from sleep.¹ This disturbance leads to snoring, fragmented sleep, and excessive daytime sleepiness. Decreased oxygen saturation in patients with OSA can increase blood pressure and heart rate.² Intermittent nocturnal hypoxemia, and the resulting periodic episodes of hypoxia and reoxygenation, are causally linked to multiple pathological states, such as hypertension, stroke, cardiovascular disease, and Type two diabetes.^3,4

Recently, the impact of OSA on brain structure and function has been increasingly emphasized. The brain, which is only 2% of the body by weight, accounts for 20% of whole-body energy consumption.⁵ The brain has scarce energy reserves and is highly dependent on cerebral blood flow (CBF) for its energy supply. CBF is tightly regulated to maintain steady cerebral perfusion. Several studies have investigated cerebral perfusion changes in OSA patients based on imaging techniques. Fernandes et al used positron emission tomography (PET) to explore regional CBF changes in OSA patients and confirmed that the CBF in the bilateral precuneus, posterior cingulate cortex and frontal lobe was significantly decreased.⁶ Subsequently, several single-photon emission computed tomography (SPECT) studies further confirmed the reduction of CBF in multiple brain regions in association with OSA,^7,8 and that the respiratory events characterizing OSA during rapid eye movement and non-rapid eye movement sleep are associated with different cerebral perfusion patterns during the day.⁹ Although PET and SPECT are the current gold standards for assessing cerebral perfusion, both of these imaging techniques are invasive and require radioactive tracers, which limits their wide clinical application. Moreover, the above studies all focused on regional CBF changes in OSA patients; the pattern of change in global CBF is still unclear. In addition, the formation of white matter (WM) hyperintensities is closely related to hypoperfusion^10,11; therefore, WM abnormalities in OSA patients also merit further evaluation.

Arterial spin labeling (ASL), which uses magnetically labeled protons in arterial blood as an endogenous contrast agent, is currently the only noninvasive perfusion imaging technique.¹² Compared with other perfusion imaging techniques, such as CT perfusion, dynamic contrast-enhanced, and dynamic susceptibility contrast MRI, ASL has the advantages of non-invasiveness, simplicity and high reproducibility. Moreover, brain perfusion is tightly coupled to metabolism.¹³ Previous studies have confirmed that CBF measured by ASL technique^14–17 has a good correlation with brain metabolism measured by PET^18–21 in determining mild cognitive impairment and Alzheimer’s dementia. Thus, ASL has emerged as a promising alternative technique that is widely applied in various disease states.²² Innes et al has identified CBF alterations in multiple brain regions in OSA patients by applying ASL.²³

In this study, we used 3D pseudo-continuous ASL technique to comprehensively explore regional and global CBF alteration patterns in OSA patients. In addition, we analyzed WM change in OSA patients.

Methods and materials

Study population

This study was approved by the Ethics Committee of our institution. Written informed consent was obtained from each subject. The data for this study were derived from the multimodality medical imaging study based on KaiLuan Study (META-KLS), a cohort study based on the Kailuan community. From December 2020 to December 2021, a total of 750 subjects (439 males, 311 females) were enrolled in this study and underwent detailed clinical and imaging examinations. Patients with OSA and age-, sex-, and body mass index (BMI)-matched non-sleep apnea (NSA) subjects were included for further analysis. All patients underwent detailed clinical interviews and physical examinations and were definitively diagnosed with OSA. Subjects meeting any of the following criteria were excluded: (1) central nervous system abnormalities (stroke, psychiatric disorder, head injury, neurodegenerative disorder); (2) uncontrolled hypertension and diabetes; and (3) alcohol, cigarette or drug abuse in the past 3 months. The OSA duration and Montreal Cognitive Assessment (MoCA) were recorded for all subjects.

Imaging protocol

All subjects underwent MRI examination (General Electric 750W, Milwaukee, WI, USA). ASL was performed using a 3D pseudo-continuous ASL sequence with the following imaging parameters: repetition time (TR)/echo time (TE), 5313/10.7 ms; field of view (FOV), 256× 256 mm; inversion time (TI) = 2525 ms; slice thickness = 4 mm; flip angle (FA), 111°; postlabel delay (PLD), 2525 ms; and in-plane resolution, 3.37 × 3.37 mm. T₂-weighted imaging was performed with the following imaging parameters: TR/TE, 5113.3/110.3 ms; slice thickness, 5 mm; FA, 142°; Sagittal T2 fluid-attenuated inversion recovery (FLAIR) imaging was performed with the following imaging parameters: TR/TE, 5000/117.5 ms; TI, 1635 ms; slice thickness, 1 mm; FA, 90°.

CBF quantification

The processing of CBF images was described in previous studies.^24,25 First, CBF maps were generated by pairwise subtraction onboard the GE MRI scanner (AW4.6, GE Healthcare). Then, the CBF maps were normalized to the Montreal Neurological Institute (MNI) space using Statistical Parametric Mapping (SPM8) software in MATLAB 2013b. Finally, the 8 mm full width at half maximum Gaussian kernel was used to smooth the CBF maps. All images were visually inspected.

To explore perfusion changes at the global level in patients with OSA, we extracted the CBF values of the whole brain, gray matter (GM) and WM using corresponding masks.²⁵ Furthermore, we further explored regional CBF alterations at the voxel level.

Classification of WM hyperintensities

Periventricular and deep WM hyperintensities were assessed by a neuroradiologist with 18 years of experience in neuroradiology on T2WI and FLAIR sequences using the Fazekas scale.²⁶ This neuroradiologist was blinded to all clinical information. Ratings of periventricular and deep WM hyperintensities were summarized into total scores (range: 0–6).

Statistical analysis

Statistical analyses were performed using SPSS 22.0 software. Shapiro-Wilk test was used to explore the normality of the data. The clinical and demographic characteristics were compared between patients with OSA and NSA subjects by two-samples t-test for continuous variables with a normal distribution, Mann-Whitney U-test for continuous variables with skewed distribution, and Chi-Square test for categorical variables. Mann-Whitney-U test was used to compare differences in global CBF between groups. To explore the group difference in regional CBF, we performed two-samples t-test with gender and age as covariates (cluster-level family-wise error (FWE) correction). Chi-square test was used to compare the differences in severity of WM hyperintensities between the two groups. A p-value < 0.05 was considered to indicate a statistically significant difference.

Results

Study population

In this study, a total of 100 (74 males, 26 females) of 750 (439 males, 311 females) subjects were diagnosed with OSA. Thus, the prevalence of OSA in general population is 13.3% (100/750), with 16.9% (74/439) in males and 8.4% (26/311) in females. Four patients with OSA were excluded due to lack of ASL images. Finally, 96 patients with OSA and 103 NSA subjects were analyzed in this study. The median age of OSA patients was 59.5 years, and the median duration of OSA was 4 years. There was no significant difference in age, gender, BMI and MoCA scores between patients with OSA and NSA subjects. All subjects were right-handed. The clinical and demographic data are shown in Table 1.

Table 1.

Clinical and demographic data of patients with OSA and NSA subjects

	OSA patients (n = 96)	NSA subjects (n = 103)	P
Age (years)	59.5	57.0	0.526a
Sex (male/female)	72/24	66/37	0.095b
BMI (kg/m2)	25.89 ± 4.13	24.92 ± 3.43	0.072c
MoCA score	25	25	0.162a
OSA duration (years)	4	-	-

BMI: body mass index; MoCA: Montreal Cognitive Assessment; NSA: non-sleep apnea;OSA: obstructive sleep apnea.

Normally distributed data are expressed as the mean ± standard deviation. Nonnormally distributed data are represented as the median

^aMann-Whitney U test.

^bchi-square test.

^cTwo-example t-test.

Global and regional CBF changes

The global CBF in all subjects is summarized in Figure 1. In whole brain and GM, patients with OSA showed significantly decreased CBF values compared to NSA subjects (whole brain: 46.86 ± 6.15 ml/100 g/min vs 48.63 ± 5.85, p = 0.021; GM: 48.61 ± 6.54 ml/100 g/min vs 50.75 ± 6.32, p = 0.010). No significant difference was found in WM CBF values between the two groups (46.41 ± 6.14 ml/100 g/min vs 47.19 ± 5.22, p = 0.250).

Figure 1.

Global CBF difference between patients with OSA and NSA subjects. CBF: cerebral blood flow; OSA: obstructive sleep apnea; NSA: non-sleep apnea

The regional CBF changes of patients with OSA is exhibited in Figure 2 and Table 2. Compared with NSA subjects, patients with OSA exhibited significantly decreased regional CBF values mainly in the right parietal lobe and right temporal lobe (FWE corrected, p < 0.05). No significant increase in regional CBF was found in OSA patients compared with NSA subjects.

Figure 2.

Altered regional CBF between patients with OSA and NSA subjects. Compared with NSA subjects, patients with OSA exhibited significantly decreased regional CBF values in right superior temporal gyrus and right precuneus (cluster-level FWE corrected, p < 0.05). CBF: cerebral blood flow; OSA: obstructive sleep apnea; NSA: non-sleep apnea; FWE: family-wise error.

Table 2.

Brain regions with significant group differences in CBF

Brain region	Peak MNI (mm)			Peak T value	Cluster size (mm3)
	x	y	z
NSA subjects > OSA patients
Cluster 1					3509
Right precuneus	18	–68	47	−3.792	978
Left precuneus	-7	–76	52	−3.985	380
Right superior parietal gyrus	14	–63	52	−3.776	222
Cluster 2					1297
Right superior temporal gyrus	63	–23	15	−3.676	555
Right middle temporal gyrus	64	–33	6	−3.547	491

OSA: obstructive sleep apnea; NSA: non-sleep apnea; MNI: Montreal Neurological Institute

WM hyperintensities changes

The proportions of patients with different Fazekas scores in OSA patients were 17.7% (score of 0), 30.2% (score of 1), 21.9% (score of 2), 8.3% (score of 3), 12.5% (score of 4), 4.2% (score of 5), and 5.2% (score of 6). In NSA subjects, the proportions of different Fazekas scores were as follows: 22.3% (score of 0), 42.7% (score of 1), 15.5% (score of 2), 13.6% (score of 3), 3.9% (score of 4), 0% (score of 5), and 1.9% (score of 6). Patients with OSA had significantly higher Fazekas scores than NSA subjects (p = 0.017).

Discussion

In this study, the prevalence of OSA in the KaiLuan community population was first reported. We further reported that patients with OSA had significantly decreased CBF at the global level as well as in regional brain regions, such as the right parietal and temporal lobes. In addition, this study confirmed that the severity of WM hyperintensities was significantly higher in OSA patients than in NSA subjects. These findings provide a basis for exploring changes in the central nervous system of OSA and for early and appropriate treatment.

OSA is an extremely common sleep-related breathing disorder characterized by partial or complete obstruction of the upper airway. OSA is more common in adults 30–70 years old and affects 13% of males and 6% of females in the United States.²⁷ A recent European study based on poly(somno)graphy in the general population showed a prevalence of OSA of 30% in males and 13% in females.²⁸ In this study, the prevalence of OSA in the general population was 13.3% (100/750), with 16.9% (74/439) in males and 8.4% (26/311) in females, which may provide a basis for exploring the prevalence of OSA in Kailuan community. Besides, the above studies have confirmed that OSA is more common in males, which is consistent with our study. How males are more likely to develop obstructive sleep apnea is unclear. Whittle et al believes that males are more likely to store more fat in their upper airways and abdomen than women.²⁹ Other studies have suggested that OSA may be associated with anatomically longer airways and higher androgen levels in men.^30,31 Further research on the role of gender in the pathogenesis of OSA is required.

Most of the previous studies only focus on regional CBF changes,^7,23 and the changes of CBF at the global level in OSA patients are still unclear. In this study, we employed non-invasive technique of ASL to comprehensively investigate cerebral perfusion patterns in patients with OSA. We confirmed that the CBF values of the whole brain and GM in these patients were significantly lower than those of NSA subjects. Subsequently, regional cerebral perfusion changes in OSA patients were further explored, and it was found that the CBF of the temporal and parietal lobes was significantly decreased. These regions with decreased CBF were all located in the right cerebral hemisphere, which is consistent with Yaouhi’s study.³² Moreover, decreased CBF mainly in these regions of OSA patients in this study was also confirmed by previous studies.^7,23 Hypopnea in OSA is characterized by varying degrees of hypoxemia, resulting in decreased cerebral perfusion.⁷ Moreover, parietal and temporal hypoperfusion is a biomarker of early Alzheimer’s disease.^33,34 Considering that OSA is a risk factor for mild cognitive impairment and dementia,^35,36 patients with severe OSA may have underlying neurodegenerative changes. To understand whether parietal and temporal hypoperfusion are early markers of dementia secondary to OSA, further cohort studies will be used to assess cerebral perfusion status among OSA patients in different cognitive states.

Several studies have confirmed microscopic WM changes in OSA patients. By using diffusion tensor imaging, Lin et al found significantly decreased fractional anisotropy values in multiple brain regions of male OSA patients, which confirmed that OSA could lead to the destruction of WM integrity.³⁷ Subsequently, Kacar et al found that the apparent diffusion coefficient value of the periventricular WM was significantly reduced.³⁸ A MR spectroscopy study further confirmed these WM changes.³⁹ In line with these findings, we assessed periventricular and deep WM changes using the Fazekas scale, which is widely used in clinical practice; we found that OSA patients had significantly higher Fazekas scores than NSA subjects. WM hyperintensities has been confirmed to be closely associated with hypoperfusion.⁴⁰ In the present study, we found reduced global and regional CBF in patients with OSA, and the evidence suggested that prolonged ischemia and hypoxia might lead to the formation and exacerbation of WM hyperintensities. Longitudinal studies of patients with OSA are needed to determine how OSA contributes to increased WM hyperintensities burden and whether hypoperfusion is an early marker of WM hyperintensities in OSA.

This study has several limitations. The severity of OSA was not assessed in this study, and we will further explore the changes in cerebral perfusion and WM in patients with OSA of different severity in the future. Second, OSA had a higher prevalence in males than in females in this study. OSA has been confirmed to occur predominantly in males,⁴¹ and we will further explore the pattern of CBF changes in OSA subjects of different genders. Third, in this study, all subjects were right-handed. We will enroll more subjects to investigate the effect of handedness on CBF change in OSA patients. Finally, this is a cross-sectional study, and the pattern of changes in CBF and WM hyperintensities before and after OSA treatment remains to be explored.

Conclusions

In this study, we report brain perfusion and WM hyperintensities alterations in patients with OSA. Both global and regional CBF were significantly decreased in OSA patients. Moreover, the severity of WM hyperintensities was significantly higher in OSA patients than in NSA subjects. Future work should assess changes in CBF and WM hyperintensities in OSA patients after treatment to demonstrate that CBF can serve as an early biomarker of neuropathology in this population.