1. Introduction
Obstructive sleep apnea syndrome (OSAS) is a major public health issue. The prevalence of OSAS in the global population is difficult to assess, as it is underdiagnosed. OSAS is associated with an increased risk of cardiovascular events and diabetes. If left untreated, it can result in difficult-to-control high blood pressure, heart rhythm disorders, and blood hypercoagulability [
1]. OSAS is also associated with heart failure, pulmonary hypertension, and chronic kidney disease [
1]. Risk factors for OSAS include obesity, age, male sex, African ethnicity, smoking, and certain craniofacial morphologies [
2].
OSAS is characterized by repeated episodes of complete (apnea) or partial (hypopnea) obstruction of the upper airway during sleep, resulting in episodes of oxygen desaturation, hypercapnia, and micro-arousals that impair sleep quality [
2,
3]. It is diagnosed when polysomnography reveals an obstructive apnea–hypopnea index (OAHI) greater than 5 (i.e., at least five episodes of obstructive apnea/hypopnea per hour of sleep). Intermittent hypoxia induces hyper-activation of the sympathetic nervous system, inflammatory response, and oxidative stress. These are the primary mechanisms responsible for the cardiovascular and metabolic consequences associated with OSAS [
4]. Acute hypoxia also induces self-regulated retinal vasodilation [
5].
Studies have suggested that OSAS is associated with various ophthalmologic conditions, such as non-arteritic anterior ischemic optic neuropathy, retinal vein occlusion, glaucoma, central serous chorioretinopathy, and severe diabetic macular edema [
6,
7]. Several studies have assessed choroidal thickness in patients with OSAS; however, these studies have presented divergent results. One meta-analysis concluded that there was a significant decrease in choroidal thickness in patients with OSAS [
8].
Optical coherence tomography angiography (OCTA) is an imaging technique that allows visualization of the retinal and choroidal vasculatures without contrast injection. The retina is irrigated by two vascular networks: retinal and choroidal vascularization. OCTA highlights the superficial and deep capillary plexuses (SCP and DCP, respectively) of the retinal vascular network (RVN). Limited studies have investigated the RVN density using OCTA in patients with OSAS, and the available evidence is inconsistent. For example, Yu et al., who were among the first to investigate the macular perfusion density (PD) in patients with OSAS, reported a decrease in the density in the parafoveolar area with increasing OSAS severity [
9]. Meanwhile, Moyal et al. reported no significant differences in the macular PD between their control, mild OSAS, moderate OSAS, and severe OSAS groups [
10]. Additionally, while Cai et al. observed a significant increase in the PD in the parafoveolar and perifoveolar DCP regions in patients with severe OSAS compared with that in controls, no difference was seen in the PD of the SCP [
3].
The Ophthalmology Department of Erasme Hospital in Brussels is equipped with PLEX Elite 9000. Owing to its powerful algorithm, this equipment allows accurate quantitative analysis of the retinal vascular density in terms of both the PD and total length of perfused vessels (VD). This study primarily assesses the retinal vascular density and has the following objectives:
- (1)
To assess the macular VD of both the SCP and DCP in patients with OSAS;
- (2)
To assess the macular PD of both the SCP and DCP in patients with OSAS;
- (3)
To assess the possible correlation between the macular VD and OSAS severity.
4. Discussion
This study compared the VD and PD between patients with OSAS and control patients. There were significant increases in the VD and PD of the DCP in the perifoveolar area in the OSAS group compared with those in the control group. The VD of the DCP in the parafoveolar area also significantly increased in the OSAS group. There were no significant differences in the VD and PD of the SCP and RVN. To our knowledge, this study is the first to analyze the macular vascular density in terms of vessel length in patients with OSAS. Nevertheless, several studies have previously analyzed the PD in these patients.
The current results are consistent with those of Cai et al., who reported an increase in the PD in the parafoveolar and perifoveolar DCP in the severe OSAS group compared with that in the control group [
3]. Furthermore, Moyal et al., Cai et al., and Colak et al. all reported no significant difference in the PD of the SCP between groups [
3,
10,
17]. This is because the SCP consists of capillaries, arterioles, and venules, whereas the DCP consists of only capillaries and venules [
16,
18]. Therefore, the oxygen supply is more stable in the SCP owing to the direct connection to the retinal arterioles from the central retinal artery. This could explain why hypoxia affects the DCP first [
17].
The current results demonstrate some discrepancy with previous reports regarding the alteration of the PD of the DCP. Colak et al. and Ucak et al. observed a significant decrease in the PD of the DCP in the parafoveolar area in their OSAS groups [
17,
19]. Colak et al. also observed this phenomenon in the perifoveolar area [
17]. Similarly, Yu et al. reported a significant decrease in the PD of the RVN in the parafoveolar and perifoveolar regions [
9]. This discrepancy in results could be explained by differences in ethnicity, age, and duration of OSAS symptoms between the study populations. For example, the populations in the current study and in the study by Cai et al. were younger than the populations studied by Ucak et al. and Colak et al. Furthermore, differences in the definition of the groups, particularly the control group, could reduce the relevance of comparisons between these studies. Moreover, most previous studies used SD RTVue-XR Avanti, which is a less powerful device in terms of image acquisition and analysis than SS PLEX Elite 9000.
The OAHI is the classic index used to evaluate OSAS severity. The current study revealed no correlation between the VD and OAHI in the patients with OSAS. Similarly, Cai et al. and Colak et al. did not report correlations between the PD and OAHI [
3,
17]. However, Yu et al. and Ucak et al. reported negative correlations between the OAHI and PD [
9,
19]. Some experts have questioned the definition of OSAS based only on the OAHI, as the OAHI may not be the most reliable indicator of OSAS severity. A new definition of OSAS could include the oxygen desaturation index (ODI) [
20]. Cai et al. supported the inclusion of the oxygen saturation in the definition of OSAS, as their results showed a negative correlation between the perifoveolar PD and the lowest hemoglobin oxygen saturation [
3]. However, Colak et al. did not report a correlation between the PD of the SCP and DCP and time spent under 90% oxygen saturation [
17].
The pathophysiological mechanisms driving alterations in the VD and PD of patients with OSAS are not yet well-understood. It is well-established that OSAS-related intermittent hypoxia results in the activation of the orthosympathetic system, which leads to peripheral vasoconstriction [
4,
21]. However, unlike choroidal vessels, retinal vessels lack autonomic innervation [
5]. Hypoxia and hypercapnia result in retinal vasodilation via an autoregulatory mechanism that maintains appropriate blood flow based on metabolic tissue needs. This autoregulatory mechanism is mediated by local factors, such as vasoactive molecules released by the endothelium. Nitric oxide (NO), certain prostaglandins (i.e., PGI2 and PGE2), and extracellular lactate are involved in retinal arterial vasodilation in response to hypoxia and hypercapnia [
4,
22]. These physiological mechanisms could explain the increase in the PD of the DCP in the perifoveolar area observed in the patients with OSAS in the current study and in the severe OSAS group in the study by Cai et al. [
3].
It is speculated that an increase in the VD in terms of vessel length occurs secondary to hypoxia; however, the exact nature of the vessels (i.e., whether they are neovessels or collateral vessels) is unknown. The non-anarchic organization of the vessels observed in OCTA images supports the hypothesis that they are collateral vessels that develop in response to hypoxia. This phenomenon has been established at the coronary level in patients with chronic ischemic heart disease [
23]. One study has suggested that coronary collateral vessels also develop in patients with OSAS [
24]. Patients with OSAS have higher blood levels of vascular endothelial growth factor (VEGF) than individuals without OSAS. Intermittent hypoxia stimulates VEGF gene transcription via hypoxia-induced factors [
25]. Oxidative stress may also be involved in the development of collateral vessels [
26]. The increased levels of VEGF and oxidative stress present in patients with OSAS are likely to be the two mechanisms involved in the collateral vessel formation observed in the current study.
The decrease in the retinal PD described in previous studies could be explained by endothelial dysfunction and atherosclerosis, which may be associated with long-term OSAS [
4,
27]. According to a recent meta-analysis, severe OSAS is associated with a high risk of endothelial dysfunction [
27]. Long-term endothelial dysfunction and oxidative stress could induce a decrease in NO production, which would decrease vasodilation and, therefore, macular vascular perfusion in patients with OSAS [
28,
29]. The impairment of vascular reactivity, regardless of the relation to endothelial dysfunction, would be more marked in patients experiencing apnea with an oxygen desaturation index of >20 [
30]. In the long term, these mechanisms could be responsible for capillary occlusion or destruction at the origin of a PD-related decrease. Therefore, it would be interesting to continue the current study to investigate the VD and PD in the included patients for several years to observe whether a two-step reaction in the retinal vasculature exists in patients with OSAS.
Limitations
Although this study was a prospective controlled study, it involved a single center, and its power is limited by the number of patients included. In addition, the control group included only subjects with complaints related to sleep but without sleep disordered breathing. Therefore, this group may potentially not be representative of the population without apnea. However, in order to avoid as much as possible any risk of selection bias during recruitment for this study, all subjects eligible according to the inclusion and exclusion criteria were invited to participate. Nevertheless, despite this systematic invitation, only patients who agreed to participate in this study were included, which may potentially limit the generalizability of our results. Furthermore, it was difficult to estimate how long the included patients had experienced OSAS symptoms. Studies with longer follow-up periods are necessary to confirm the pathophysiological hypotheses and clinical implications of the current results.