Relationship between Retinal Oxygen Saturation and the Severity of Visual Field Damage in Glaucoma

Abstract

Purpose:

To investigate the association between retinal oxygen saturation (StO₂) % and the severity of visual field (VF) loss in glaucoma.

Methods:

A total of 198 eyes from 131 glaucoma patients were included in this cross-sectional study. Participants underwent imaging using ocular oximetry (Zilia; Quebec City, Canada) and 24-2 SITA standard VF (Carl Zeiss-Meditec, San Leandro). StO₂ (%), was measured at 2 locations of the peripapillary optic nerve head (ONH) (superotemporal, and inferotemporal). Measurements were reported as the mean of at least 5 measurements in each location. Associations between the severity of VF loss, reported as mean deviation (MD), and StO₂ (%) was calculated.

Results:

198 eyes of 131 patients (mean (95% CI) age, 71.1 (68.9,73.3) years, 68 females [51.9%], 63 males (48.1%)) were analyzed. In univariable analysis, higher StO₂ −0.06 (−0.12, 0.00) was associated with severity in all hemifields (P=0.047). Multivariate regression analysis showed that each 1% increase in StO₂ was associated with −0.06 (−0.12,−0.00) dB loss in MD in all hemifields (P=0.043). In multivariate regression analysis in the superior hemifields, higher StO₂ −0.07 (−0.16, 0.01) tended to be associated with superior hemifield severity (P=0.09).

Conclusions:

Retinal oximetry enabled the continuous quantitative measurement of retinal StO_2. Increased StO₂ was significantly associated with the severity of VF damage in glaucoma patients.

Précis

Increased oxygen saturation was significantly associated with the severity of VF damage in glaucoma patients.

Introduction

Glaucoma is a progressive optic neuropathy characterized by the loss of retinal ganglion cells (RGC) and accompanying visual field (VF) damage.¹ Although intraocular pressure (IOP) is a well-known risk factor for glaucoma, some patients continue to progress despite IOP reduction.² Apart from tissue degeneration as a result of increased IOP, insufficient or poorly regulated blood supply has been proposed to occur in glaucoma.^3–5

The physiological regulation of ocular perfusion is complex. The autoregulation of retinal vessels balances fluctuations in perfusion pressure which depends on variations of IOP and blood pressure.^6,7 Systemic and localized vascular abnormalities have been shown to be associated with glaucoma.^2,8,9 A low perfusion pressure is a possible risk factor not only for the development but also the progression of open-angle glaucoma.⁶ Moreover, a failure of adaptation of blood flow to the tissue’s demand (neurovascular coupling) has been proposed as relevant to the pathogenesis of glaucoma.^7,10,11 However, there is still limited knowledge regarding impaired blood flow and metabolic alterations in the retina of glaucomatous subjects.

Capillary oxygenation, regardless of the underlying mechanism of glaucoma damage, may provide insights into the metabolic function of the optic nerve and retina in glaucoma.^11,12 The accurate measurement of retinal oxygen levels has been challenging.¹³ Since its adoption as a tool for clinical research, retinal oximetry has led to the discovery of new biomarkers in several retinal diseases including diabetic retinopathy, central retinal vein occlusion, retinitis pigmentosa, and glaucoma, as well as diseases of the brain, including Alzheimer’s disease.¹² Studies using various retinal oximetry devices to measure the oxygen saturation of the retinal vasculature in glaucoma have been reported, with the main focus on measuring O₂ saturation (StO₂) in major vessels of optic nerve head (ONH).^14–18 In each of these earlier studies, retinal blood vessel oxygen saturation was measured rather than retinal tissue oxygenation.

The Zilia Ocular is a new technology that enables the continuous quantitative measurement of retinal StO₂. It utilizes diffuse reflectance spectroscopy to determine oxygen saturation at a specific target point in the tissue including the microvasculature of the of retina or in the ONH. Low variability between the measurements of Zilia ocular oximetry has previously been reported.¹⁹ In a recent study, the Zilia oximeter showed good macular test-retest repeatability.^20,21 This study reported 0.78 as an average intraclass correlation and 8.4 for the average intrasubject repeatability coefficient for the three acquisition times.²⁰ It was also reported that retinal oxygenation saturation undergoes diurnal variations over 24 h with an amplitude of 5.84 ± 3.86% and acrophase of 2.35 h similar to IOP, mean ocular perfusion pressure (MOPP), axial length, choroidal thickness, superficial vessel density, heart rate, systolic blood pressure, and mean arterial pressure (MAP).^20,21 Therefore, this technology may help elucidate the role of microcirculation and optic nerve blood flow in the pathogenesis of glaucoma.

The purpose of this study was to assess the relationship of peripapillary retinal oxygen saturation (StO₂), rather than blood vessel oxygen saturation, in glaucoma eyes. We hypothesized that there would be a change in StO₂ in glaucomatous eyes with worsening VFs.

Methods

This was an observational cross-sectional study that included glaucoma suspects and glaucoma participants. Participants were enrolled from patients at the Shiley Eye Institute of the University of California, San Diego. The research protocol adhered to the tenets of the Declaration of Helsinki and was approved by the University of California, San Diego Institutional Review Board. Written informed consent was obtained from all study participants.

All participants underwent a comprehensive ophthalmologic examination, including best-corrected visual acuity, slit-lamp biomicroscopy, IOP measurement with Goldmann applanation tonometry, gonioscopy, dilated fundus examination, stereoscopic optic disc photography, ultrasound pachymetry, and standard automated perimetry (Humphrey Field Analyzer; Carl Zeiss Meditec, Dublin, CA), in both eyes. Self-reported history of hypertension and diabetes were also collected.

Glaucoma suspect eyes were defined as those with elevated IOP (≥22mmHg) or glaucomatous-appearing optic discs without repeatable glaucomatous VF damage. Glaucomatous-appearing optic discs were defined as those with observable neuroretinal rim narrowing or notching, excavation, or a localized or diffuse retinal nerve fiber layer defect suggestive of glaucoma based on the standard review of stereophotographs. Primary open angle glaucoma (POAG) was defined as eyes showing at least two reliable (≤33% fixation losses, and false-positives) and repeatable abnormal (GHT outside normal limits or PSD outside 95% normal limits) VF results using the 24-2 SITA Standard with glaucomatous-looking discs on dilated eye exam. Glaucoma disease severity was classified as early (VF mean deviation (MD)>-6 dB) or moderate to advanced (VF MD≤-6 dB) by HFA SITA Standard 24-2.²² All participants were confirmed to have open angles on gonioscopy.

Only participants older than 18 years of age with open angles on gonioscopy and a best-corrected visual acuity of 20/40 or better were included. Exclusion criteria included (1) history of trauma or intraocular surgery (except for uncomplicated cataract surgery or glaucoma surgery), (2) coexisting retinal disease, (3) uveitis, (4) or non-glaucomatous optic neuropathy. Participants with the diagnosis of systemic diseases such as Parkinson’s disease, Alzheimer’s disease, dementia, or a history of stroke were also excluded.

Standard Automated Perimetry

All participants underwent VF testing using the 24-2 pattern Swedish interactive threshold algorithm on the Humphrey Field Analyzer (Carl Zeiss Meditec, Dublin, CA) within 6 months of imaging. Only reliable tests (≤33% fixation losses and false-negatives, and ≤15% false-positives) were included. The quality of VF tests was also reviewed to identify and exclude VFs with evidence of inattention or inappropriate fixation, artifacts such as eyelid and lens rim artifacts, fatigue effects, and abnormal results caused by diseases other than glaucoma.

Ocular Imaging and oxygen saturation

The participants underwent measurements of retinal oxygen levels in both eyes using ocular oximetry. Continuous quantitative measurement of retinal oxygen saturation (StO₂) was assessed with retinal oximetry (Zilia; Quebec City, Canada). The Zilia Ocular device (version D, software version 01.2109.279f76) utilizes diffuse reflectance spectroscopy to determine oxygen concentration at selected points (Superotemporal and inferotemporal) around the ONH, Figure 1. These points provide sensitive localized assessments of O2 saturation in areas of 1.5 degree for each region.

Figure 1.

The Zilia Ocular device utilizes diffuse reflectance spectroscopy to determine oxygen concentration at a selected point (Superotemporal (A), and inferotemporal (B)) around the optic nerve head (ONH).

The mean StO₂ at each target point was considered as the average percentage of at least 5 measurements obtained in each area, with a standard deviation (SD) of less than 5. The mean StO2 at each target point was considered as the average percentage of at least 5 measurements obtained in each area, with a SD of less than 5%. We considered SD above 95% of our dataset as a cut-off to exclude eyes that are extreme outliers and have considerable fluctuations during measurement that could be related to saccadic eye movement during imaging or low-quality spectra. Therefore, we excluded eyes with less than 5 data points and SD of more than 5%. Trained graders reviewed fundus photo of Zilia images and excluded poor-quality images that exhibited artifacts such as darkness, no StO₂ measurement, defocus, or measurement on major vessels; the trained graders were GM and EM. The graders evaluated the image quality and any disagreements between them were further evaluated by SM. There are no quantitative measures like OCT for the device currently and the quality control was performed according to the artifacts or issues that have been listed and found to affect the image quality, spectral quality, and measurement during imaging and was consulted with the manufacturer. During the StO2 acquisition process, some events affecting measurements could occur (saccades, presence of a large blood vessel, defocus, etc.). In addition, the device evaluates spectral quality. When the spectral quality criteria were not reached (which can be due to saccades, low signal to noise ratio, blinking or other events affecting spectral quality) those images and StO2 measurements were removed from the analysis.

Statistical Analysis

Patient and eye characteristics data were presented as means (95% confidence interval (CI)) for continuous variables and counts (%) for categorical variables. The average total deviation (TD) was calculated for each hemifield based on individual test points, excluding the blind spot. The sensitivity in decibels for each hemifield was converted to a linear scale of 1/Lambert (1/L) =10^dB/10 and then averaged to obtain mean sensitivity on a linear scale.²³ Subsequently, the mean sensitivity on a linear scale was converted to the dB scale. Linear mixed models were used to evaluate the effect of StO₂ and baseline characteristics on VF hemifield loss for the inferior and superior hemifield while adjusting for potential correlations between both eyes from the same individual.^24,25 Multivariable linear mixed models were adjusted for any variable with a p-value < 0.10. All statistical analyses were performed using the commercially available software Stata version 15 (StataCorp LP, College Station, TX), and the alpha level (type I error) was set at 0.05.

Results

A total of 198 eyes (31 glaucoma suspect and 167 POAG) from 131 participants were included in the analysis. The mean age (95% CI) was 71.1 (68.9, 73.3) years. Among the participants, 127 (64.1%) and 71 (35.9%) patients had mild glaucoma, and moderate to severe glaucoma, respectively. (Table 1). The mean VF MD (95% CI) was −6.0 dB (−6.9, −5.0) in all eyes. In mild and moderate to severe glaucoma, the mean VF MD losses were −2.0 (−2.3, −1.6) dB and −13.1(−14.5, −11.7) dB, respectively. The mean baseline StO₂ (95% CI) was 69.8 (68.3, 71.3) and 69.3 (67.8, 70.7) in inferior and superior regions, respectively. The mean (95% CI) retinal nerve fiber and ganglion cell inner plexiform layer thickness for all eyes were 69.3 (67.2, 71.3) μm and 65.2 (63.5, 66.9) μm respectively. Demographics and baseline clinical characteristics of the participants are presented in Table 1.

Table 1.

Demographics and Baseline Clinical Characteristics of the Study

Characteristics	n=258 eyes of 165 patients
Age (years)	69.9(68.0,71.8)
Sex (Female/ Male), n(%)	82(49.7) / 83(50.3)
Race, n(%)
African American	7(4.2)
Asian	16(9.7)
White	111(67.3)
Others	31 (18.8)
IOP (mmHg)	15.1(14.3, 15.8)
Self-reported diabetes, n(%)	9(5.5)
Self-reported hypertension, n(%)	38(23.0)
CCT (μm)	5.4 (5.4,5.5)
Diagnosis
Glaucoma/ Glaucoma suspect
Eyes, n(%)	141(85.5)/24(14.5)
Disease Severity by baseline 24-2 VF MD
Early glaucoma, n(%)	−1.66(−2.0,−1.3)
Moderate and advanced glaucoma, n(%)	−13.1(−14.6,−11.6)
Superior hemifield severity by baseline 24-2 VF MD	−6.6(−8.1,−5.2)
Inferior hemifield severity by baseline 24-2 VF MD	−5.1(−6.5,−3.7)
VF 24-2 MD (dB)	−5.7(−6.5,−3.7)
VF 24-2 PSD (dB)	5.3(4.7,6.0)
StO2 (%), inferior region	69.8 (68.3,70.7)
StO2 (%), superior region	69.1 (67.7,70.4)
cpRNFL thickness, μm
All eyes	69.3 (67.2, 71.3)
Glaucoma	67.5 (65.5, 69.6)
Glaucoma suspect	79.4 (73.1, 85.7)
GCC thickness, μm
All eyes	65.2 (63.5, 66.9)
Glaucoma	63.9 (62.1, 65.6)
Glaucoma suspect	73.1 (68.5, 77.7)

StO₂=O2 saturation; IOP = intraocular pressure; MD = mean deviation; PSD = pattern standard deviation; VF = visual field; n=number; SD=standard deviation; cp-RNFL= circumpapillary retinal nerve fiber layer thickness; GCC= ganglion cell complex thickness.

Values are shown in mean (95% confidence interval), unless otherwise indicated.

The results of univariate linear regressions for factors affecting both inferior and superior hemifield MD are summarized in Table 2. Higher mean (CI) StO₂ −0.06 (−0.12,-0.00), P=0.047, higher IOP 0.24 (0.14,0.34), P<0.001, and lower central corneal thickness (CCT) 2.87 (1.27,4.47), P<0.001 were associated with severity in all hemifields. Scatter plot shows the linear relationships between StO₂ with MD in all (superior and inferior) hemifields in Figure 2.

Table 2.

Univariable and multivariable linear mixed analysis of factors correlated with each hemifield of VF severity

Variables	Univariable Model	P value	Multivariable Model	P value
StO2 (per 1% lower)	−0.06(−0.12,−0.00)	0.047	−0.06(−0.12, −0.00)	0.043
Age (per 1 yr older)	−0.03 (−0.09,0.03)	0.32
Gender (female)	−0.22 (−1.52,1.08)	0.74
Race (African descent)	−0.29 (−0.89,0.31)	0.35
IOP (per 1 mmHg higher)	0.24 (0.14,0.34)	<0.001	0.21 (0.11, 0.31)	0.000
CCT (per 100 μm thinner)	2.87 (1.27,4.47)	<0.001	2.12(0.58,3.67)	0.007
Hypertension (yes)	1.03 (−0.73,2.79)	0.25
Diabetes (yes)	−0.85(−3.90,2.19)	0.58

Boldface values indicate statistical significance.

Abbreviations: CCT = central corneal thickness; StO₂ = O2 Saturation; IOP = intraocular pressure.

Figure 2.

Scatter plot illustrates the linear (black dashed line) correlation between all (inferior and superior) hemifields and O2 Saturation in corresponding hemifields. Red and green dashed lines show the correlation between inferior and superior hemifield severity and O2 Saturation in their corresponding hemifields, respectively. dB = decibels. *R²: Adjusted, R² from the linear regression model.

Results from the multivariable regression analysis for both superior and inferior hemifield MD as the dependent variable are summarized in Table 2. In multivariate linear regression, each 1% increase in StO₂ was associated with a −0.06 dB loss in MD (P= 0.043), while controlling for the potentially confounding effect of IOP, and CCT. Results from the multivariable regression analysis for the inferior or superior hemifield MD as the dependent variable are summarized in Table 3 and Table 4. In superior hemifield VF severity, multivariate linear regression analysis showed that each 1% increase in StO₂ tended to be associated with a −0.07 dB loss in MD (P= 0.09), while controlling for the potentially confounding effect of IOP, and CCT.

Table 3.

Univariable and multivariable linear mixed analysis of factors correlated with inferior hemifield VF severity

Variables	Univariable Model	P value	Multivariable Model	P value
StO2 (per 1% lower)	−0.04(−0.12,0.05)	0.41	−0.04(−0.12, 0.04)	0.33
Age (per 1 yr older)	−0.03 (−0.10,0.04)	0.38
Gender (female)	−0.03(−1.49,1.42)	0.96
Race (African descent)	−0.51 (−1.21,0.20)	0.16
IOP (per 1 mmHg higher)	0.25 (0.13,0.38)	<0.000	0.19 (0.06, 0.32)	0.004
CCT (per 100 μm thinner)	3.80 (1.95,5.65)	<0.000	2.78(0.91,4.64)	0.004
Hypertension (yes)	1.58 (−0.37,3.53)	0.11
Diabetes (yes)	−0.43 (−3.70,2.84)	0.80

Boldface values indicate statistical significance.

Abbreviations: CCT = central corneal thickness; StO₂ = O2 Saturation; IOP = intraocular pressure.

Table 4.

Univariable and multivariable linear mixed analysis of factors correlated with superior hemifield VF severity

Variables	Univariable Model	P value	Multivariable Model	P value
StO2 (per 1% lower)	−0.10 (−0.19, −0.00)	0.05	−0.07 (−0.16,0.01)	0.09
Age (per 1 yr older)	−0.05 (−0.13,0.02)	0.16
Gender (female)	0.27 (−1.58, 2.12)	0.77
Race (African descent)	0.25 (−0.50,0.99)	0.52
IOP (per 1 mmHg higher)	0.23(0.10,0.37)	0.001	0.22 (0.08,0.35)	0.001
CCT (per 100 μm thinner)	2.03 (−0.04, 4.09)	0.06	1.53(−0.43,3.49)	0.13
Hypertension (yes)	0.11 (−2.23,2.46)	0.93
Diabetes (yes)	−1.23 (−5.22, 2.76)	0.55

Boldface values indicate statistical significance.

Abbreviations: CCT = central corneal thickness; StO₂ = O2 Saturation; IOP = intraocular pressure.

Discussion

The present study reveals a significant correlation between retinal StO₂ and the severity of VF loss. Our results indicate that higher retinal StO₂ levels are linked to worse VF damage in the corresponding visual hemifield. The study findings further suggest a topographical relationship between the area of increased StO₂ and the area of VF loss, suggesting that reduced oxygen consumption in tissue may be associated with glaucomatous VF damage.

While some studies have found an association between StO₂ levels and VF severity, several others have not found any such association.^18,26–31 In the present study, higher StO₂ levels in retinal tissue were significantly associated with worse VF MD in all eyes. Specifically, higher StO₂ levels in the inferotemporal peripapillary region tended to be associated with worse superior VF MD. Our findings align with previous reports that found an association between retinal StO₂ in glaucomatous patients and VF severity.^18,26–28 However, these previous studies measured StO₂ in blood vessels rather than ocular tissue, which uses infrared oxygen utilization in the retinal tissues.^18,26,27 For instance, Shimakaji et al. used a non-invasive spectrophotometric retinal oximeter to measure StO₂ in the retinal vessels of glaucomatous eyes.²⁷ They found that retinal venous oxygen saturation was increased in the hemifield with the most advanced glaucomatous VF defects relative to the less affected hemifield in patients with NTG as well as patients with POAG only when the difference between the worse and better hemifields was more than 10 dB.²⁷ Additionally, higher StO₂ in retinal venules and a lower arteriovenous difference were found to be associated with worse VF damage in glaucoma eyes, suggesting reduced oxygen consumption in eyes with glaucoma.^18,28

The underlying mechanisms of this association between higher StO₂ levels and worse VF MD remain unknown, but we speculate that the retinal damage caused by glaucoma progression may be associated with reduced consumption of oxygen, leading to increased StO₂ in the retinal tissues. A similar phenomenon has been described in eyes treated with panretinal photocoagulation (PRP).³² PRP treat conditions that cause extensive peripheral ischemia such as proliferative diabetic retinopathy and retinal vein occlusions. PRP is believed to halt the progression of ischemic disease by improving retinal oxygenation and reducing the drive for vascular endothelial growth factor (VEGF) production by the retina. Improved oxygenation with PRP occurs through different mechanisms, including laser burns causing retinal thinning, which brings the choriocapillaris and its microvasculature physically closer to the inner retinal layers. Additionally, highly metabolically active photoreceptors are obliterated with PRP, reducing overall retinal oxygen demand.^33,34

Areas of choriocapillaris loss have been detected in the parapapillary area and are associated with the severity of the disease.^35,36 Decreased vessel density has also been found in the superficial retinal layers, linked to RNFL thinning and ganglion cell death.³⁶ Therefore, the loss of choriocapillaris, retinal vessels or tissue atrophy associated with glaucomatous retinal thinning may lead to increased StO₂ in the tissues.¹³ Additionally, hypotensive medications and surgery may be another contributing factor to the increase in StO₂, possibly due to lower IOP.^18,37

It is important to note that although the data showed increased StO₂, this does not exclude the possibility of ischemia and hypoxia contributing to glaucomatous VF loss. While elevated IOP in glaucoma is known as the leading risk factor for glaucomatous tissue damage, growing evidence suggests that vascular dysfunction and ultimately decreased retinal and choroidal blood flow are closely associated with increased risk of glaucoma incidence and progression especially in low-tension glaucoma.^4,5,38 Ito et al. reported lower StO₂ levels in normal tension glaucoma (NTG) patients at superior, nasal, and inferotemporal juxta-papillary retinal points compared to high tension glaucoma and normal patients.³⁹ This supports the theory of ischemic insult to the optic disc and retina, which may have a role in the structural and functional damage in glaucoma.^4,5 Additionally, these conditions with low StO₂ may be intermittent and more likely to be present when glaucoma treatment is poor, when IOP spikes, or when blood pressure drops, especially at night.^40,41 Considering that our patients were under glaucoma treatment and had good IOP control at the time of the study (mean IOP: 14.9), retinal hypoxia might have been prevented.

This study has several other limitations. Previous reports suggested that standard structural measures such as RNFL thickness reach a floor effect, and VF tests are highly variable in eyes with advanced glaucoma.⁴² Therefore, it is important to evaluate whether oximetry has a sufficient dynamic range to provide clinically relevant information across the full spectrum of glaucoma severity. Additionally, we did not evaluate the possible confounding impact of various systemic conditions such as blood pressure and perfusion pressure, systemic medications, and glaucoma eye drops on StO₂ and its relationship to functional measures. As patients were taking different glaucoma medications and some of them had undergone glaucoma surgery, an effect on the retinal ocular saturation by influencing ocular blood flow may be seen.^43–45 Furthermore, this current study does not provide information on the behavior of oxygen saturation when ocular blood flow is compromised in these patients. Longitudinal studies are necessary to evaluate the topographic and temporal relationship between changes in StO₂ and glaucomatous changes in standard structural and functional measures in healthy participants, glaucoma suspects, and those with glaucoma. Moreover, pigmentation differences may alter oximetry measurements based on light absorption.⁴⁶ Future studies should evaluate if racial differences can affect StO2 measurements in patients. In addition, precise structure-function is important when evaluating the association between StO2 and hemifield severity. Due to the size of locations and our methods to avoid major vessels during oximetry, there were small variations in the locations of oximetry measurements (superior/suprotemporal as well as inferotemporal/ inferior regions). Therefore, a more global / hemifield approach was chosen to better reflect the association of oxygenation and visual function.

In conclusion, StO₂ measurements are significantly associated with the severity of VF damage, particularly in the superior hemifield. Oximetry may allow monitoring of metabolic changes in glaucoma to enhance our understanding of glaucoma pathophysiology. Future studies will assess whether these measurements can be used in clinical practice to assess glaucoma progression.