Increased Oxygen Saturation in Retinal Venules During Isometric Exercise Is Accompanied With Increased Peripheral Blood Flow in Normal Persons

Abstract

Purpose: A recent study has shown that an increase in the arterial blood pressure of approximately 10 mm Hg in healthy persons can increase the oxygen saturation in venules from the retinal periphery but not from the macular area. The purpose of the present study was to investigate whether a higher increase in blood pressure has further effects on oxygen saturations and whether this is accompanied with changes in retinal blood flow.

Methods: In 30 healthy persons, oxygen saturation, diameter, and blood flow were measured in arterioles to and venules from the retinal periphery and the macular area. The experiments were performed before and during an experimental increase in arterial blood pressure of (mean ± SD) 18.3 ± 6.2 mm Hg.

Results: A higher number of venules than arterioles branching from the temporal vascular arcades to the macular area was balanced by a smaller diameter of the venules. Isometric exercise induced significant contraction of both peripheral and macular arterioles (P < 0.01 for both comparisons) and significant increase in oxygen saturation in both peripheral and macular venules (P < 0.001 for both comparisons). This was accompanied with a significant increase in the blood flow in the peripheral arterioles and venules (P = 0.4 for both comparisons), but not in their macular counterparts (P > 0.06 for both comparisons).

Conclusions: Increased systemic blood pressure leading to arterial contraction and increased venous oxygen saturation in the retina in normal persons can increase peripheral blood flow without significant effects on macular blood flow. This may contribute to explaining regional differences in the response pattern of retinal vascular disease.

In a recent study, it was shown that an increase in the arterial blood pressure of approximately 10 mm Hg in healthy persons can increase the oxygen saturation in venules from the retinal periphery but not from the macular area.¹ However, it was not clarified whether a higher increase in blood pressure could have further effects on regional oxygen saturations and whether these effects could be due to changes in blood flow. This might represent changes in transit times or in shunting of blood to bypass the microcirculation. Therefore, a further elucidation of the finding requires that the experiment is extended with a higher increase in the arterial blood pressure and measurement of the accompanying changes in blood flow in peripheral and macular vessel branches. 

Retinal blood flow, that is, blood velocity multiplied by the cross-sectional area of the vessels, has traditionally been estimated from the systemic blood pressure and the vascular resistance derived from the diameter of retinal resistance vessels.^2,3 This approach does not fully take into account that the intraluminal pressure decreases with increasing distance from the heart and that the vascular resistance depends on the diameter along the total length of the vascular system and not only at single locations on the resistance vessels. However, recent technological advances have introduced new methods for measuring retinal blood flow based on, for example, adaptive optics,⁴ the recording of light scattering,⁵ and the Doppler effect.⁶ One of these techniques is Doppler optical coherence tomography (OCT) that allows the quantification of blood flow in individual retinal vessels on the basis of light scattering from moving blood cells.⁷ 

Therefore, 30 healthy persons where subjected to oximetry and measurement of blood flow by Doppler OCT in arterioles to and venules from the retinal periphery and the macular area. The experiments were performed before and during an increase in the arterial blood pressure of approximately 18 mm Hg induced by isometric exercise. 

Materials and Methods

Participants

Thirty persons (14 men and 16 women) aged 20 to 33 years with no known previous or present systemic or ocular disease were recruited by public announcement among employees and students at the Department of Ophthalmology, Aarhus University Hospital. Table 1 shows the demographic and clinical characteristics of the participants. The study had been approved by the Regional Committee for Scientific Ethics (application number: 1-10-72-351-21) and was conducted in accordance with the Declaration of Helsinki. This included the obtainment of written and oral informed consent before enrollment. 

Table 1.

 

Characteristics of the Participants and Studied Eyes (Mean ± SD)

 

Examination

The participants were instructed to abstain from intake of caffein 12 hours prior to the examination,⁸ which was carried out on the same day for each participant. 

Ophthalmologic Examination

A routine ophthalmologic examination was performed including measurement of best corrected visual acuity (BCVA) in accordance with Early Treatment Diabetic Retinopathy Study (ETDRS) standards.⁹ Mydriasis was induced by phenylephrine 10% (Amgros I/S, Copenhagen, Denmark) and tropicamide 1% (Bausch & Lomb, Surrey, UK), followed by slit lamp examination and fundus photography (Canon CF 60Z; Canon, Amstelveen, The Netherlands) with a 60 degrees fundus photograph centered on the fovea and the optic disk. These photographs were used to document the presence of normal central retinal morphology and to identify the vessels to be studied. The eye that contained the most complete set of four branches from the upper temporal arcade, that is, an arteriolar and a venular branch toward, respectively, the retinal periphery and the macular area with a diameter of at least one third of the main vascular arcade, was selected. If the number of eligible vessel branches was identical in the two eyes, the eye with the shortest distance between the most distal and proximal of these branches was selected. This resulted in the identification of 59 arteriolar branches (30 peripheral and 29 macular) and 60 venular branches (30 peripheral and 30 macular) from 18 right eyes and 12 left eyes. Together with the 30 peripapillary arterioles and 30 peripapillary venules, this resulted in 179 vascular segments to be studied. 

The axial length of the eye was measured by ultrasound biometry (Lenstar LS 900; Haag-Streit AG, Köniz, Switzerland), and the intraocular pressure (IOP) was measured using a non-contact tonometer (Tonoref II; Nidek, Gamagori, Aichi, Japan). 

Blood Flow

Blood flow was measured in the peripapillary vessels between one half and two disk diameters from the disk margin and in the branches toward, respectively, the retinal periphery and the macular area immediately after the branching point. The measurements were performed using a dual-beam Doppler Fourier-Domain OCT system developed at the Center for Biomedical Engineering and Physics in Vienna, Austria.^7,10 The scans were oriented either vertically or horizontally depending on the direction that would allow the crossing of the scan to be nearest to perpendicular to the direction of the vessel. 

Oxygen Saturation

The oxygen saturation in larger retinal vessels was measured using a dual wavelength oximeter (model T1; Oxymap, Reykavik, Iceland), which is a fundus camera (Topcon TRC-50DX; Topcon Corporation, Tokyo, Japan) that captures images simultaneously at two different wavelengths (570 nm and 600 nm). Photographs were centered on, respectively, the fovea, the optic disk, and on the upper and the lower temporal vascular arcades¹¹ to enable the extraction of oxygen saturation values corresponding to the locations where blood flow was measured by Doppler OCT. 

Examination Protocol

The examination had the following steps: 

• 1) The participant was seated in a chair and rested for 5 minutes, followed by measurement of the arterial blood pressure using an automated oscillometric blood pressure monitor (Omron 705IT; Omron Healthcare, Kyoto, Japan) with the cuff on the upper left arm.
• 2) Subsequently, the participant was asked to lift a 2 kg dumbbell for 2 minutes with the right arm held as laterally extended as possible, followed by a new measurement of the arterial blood pressure.

The two-step procedure was repeated 5 to 7 times, each separated by 5 minutes of rest. At the first repetition, the blood pressure measurement was replaced by the capture of oximetry images, the second repetition with a measurement of the IOP, and the last 3 to 5 repetitions with the capture of Doppler OCT scans until all the preselected vascular segments had been covered. 

Previous studies have shown that the inclusion of 27 persons is necessary to measure changes in retinal oxygen saturation and blood flow with an alpha-risk of 5% and a power of 80% using the described techniques.¹² Therefore, 30 persons were included in the study. 

Data Analysis

Blood Pressure

The mean arterial pressure (MAP) was calculated as MAP = 1/3 * BP_s + 2/3 * BP_d, where BP_s was the systolic and BP_d the diastolic blood pressure. 

Doppler OCT

The scanning of the preselected 179 vessels at rest and during isometric exercise resulted in 358 scans of which 60 were excluded due to insufficient scan quality, 31 because of instability of the recording due to eye movements, and 29 because of blurring due to lack of focus. 

The remaining scans were analyzed using specially developed software (DOCTstudio, version 0.9, 30.11.2016), which translated the phase shift of the light backscattered from moving red blood cells to a linear velocity. The software corrected for the deviation of the scanning plane from perpendicular to the vessel and for the angle of incidence of the two beams at the retinal plane.¹⁰ However, high linear velocities measured by Doppler OCT are sensitive to phase shift values outside the range of –π to π that erroneously appear to represent reverse flow values.⁷ This phenomenon was observed within the central third of the diameter of 4 out of 25 peripapillary arterioles at rest and in 8 out of 22 of these vessels during isometric exercise. 

In order to correct for this artifact, the linear velocity of the blood was calculated in both the total vascular cross section and in the peripheral third of the diameter in the peripapillary arterioles of all 21 vessels where the central phase shift artifact was not observed. The plotting of these two measures in an x-y diagram showed that the linear velocity integrated over the full cross section (LVf) could be predicted from the velocity in the peripheral third of the vessel (LVp) according to the linear equation: LVf = LVp*1.03 + 2.59, R² = 0.95, P < 0.0001. This equation was used to calculate the LVf in the vessels where the velocity could only be measured in the peripheral third of the vessel. In two scans, the phase shift was out of range in more than one third of vessel diameter in the Doppler OCT recording and these scans were discarded. This resulted in Doppler OCT measurements from 158 of 179 vessel segments at rest and 138 of 179 vessel segments during isometric exercise. 

Retinal Oximetry

The Oxymap Analyzer software was used to delimit the peripapillary vessel segments in accordance with the method described previously.¹ In short, a circle was placed by the eye to best fit the margin of the optic disk, and this circle was used to define an inner circle with a 30 pixels larger diameter and an outer circle with a 3 times larger diameter. In each image, the inner and the outer circles delimited the proximal and distal ends of the upper temporal peripapillary vessel segments that were studied. Peripheral and macular branch segments were selected to start from the branching point and consist of the longest unbranched segment not exceeding a length of 50 pixels. Because 1 pixel corresponded to approximately 9.3 micrometers at the retinal plane,¹³ the maximum length of the vessel segment was approximately 465 micrometers. 

The oxygen saturation was calculated from the light reflected from the vessel (I) and the perivascular retina (I₀) at 570 nm and 600 nm. For each of the two wavelengths, the optical density (OD) was calculated as OD = log (I/I₀), which was subsequently used to calculate the optical density ratio (ODR), ODR = OD₆₀₀/OD₅₇₀.¹⁴ Finally, the oxygen saturation (SatO₂) could be calculated using the formula: SatO₂ = (a * ODR + b) + (c * d + k), with the calibration set to: a = –1.28, b = 1.24, c = 0.0097, and k = –0.14 (Oxymap Analyzer version: 2.5.2, v2), and d = the vessel diameter extracted by the oximetry software.¹⁵ 

Distance From Optic Disk to Branching Points

In the oximetry images, there were no significant differences between the distances (mean ± SD) along the vessels from the optic disc to the branching points toward the retinal periphery and macular area for arterioles (peripheral 513.1 ± 44.1 pixels and macular 499.0 ± 44.1 pixels, P = 0.82) or venules (peripheral 540.7 ± 40.9 pixels and macular 609.6 ± 57.9 pixels, P = 0.33). 

Diameter Measurements

The fundus photograph obtained by the oximeter at 570 nm was saved in jpg-format and imported into the Automated Retinal Image Analyzer (ARIA, V1-09-12-11), an open-source software for the MATLAB platform (MathWorks Inc., Natick, MA, USA), in order to define vessel borders and calculate the vessel diameter.¹⁶ The software calculated the mean of individual diameter measurements for every pixel along the vessel. The mean vessel diameters were also extracted from the Doppler OCT scans, but because the definition of the vessel border using this technique was more variable,¹⁷ the vessel diameters obtained from fundus photographs were used for the further analysis. 

Vascular Branches Toward the Retinal Periphery and the Macular Area

On each of the fovea centered 60 degrees fundus photographs, the fovea was marked manually to constitute the center of two circles (Fig. 1). An outer circle was defined with a radius extending from the fovea to the nasal margin of the optic disk and an inner circle with half this diameter. Arterioles and venules with a diameter larger than the detection limit of 8 pixels, corresponding to approximately 42 micrometers in a standard Gullstrand eye,¹⁸ were marked if they branched from the upper or lower temporal vascular arcades between a vertical line through the center of the optic disk and the crossing of the arcade with the outer circle. The marked vessels were counted and their average diameter from the branching at the vascular arcade to the crossing of the outer circle (peripheral vessels) or the inner circle (macular vessels) were measured using the ARIA software. An example is shown in Figure 1. 

Figure 1.

 

Fundus photograph with the two (black) circles centered on the fovea. Branches from the upper and lower temporal arcades have been identified between the vertical white line through the center of the optic disk and the temporal crossings of these branches with the large circle (small white lines). The identified branches are marked in red (arterioles) and blue (venules).

Blood Flow

On the basis of the linear velocity of the blood (v) obtained by Doppler OCT and the vessel diameter (d), the blood flow (Q) could be calculated from the equation: Q = (v * d² * π)/4.¹⁹ 

Statistical Analysis

The statistical analyses were performed using STATA (version 17.0; STATA, College Station, TX, USA). Probability plots showed that all data were normal distributed. The unpaired t-test was used to test if the distance from the optic disc to the branching point, the number of vessels, vessel diameter, saturation, and blood flow between the peripheral and macular branches differed significantly at rest. The paired t-test was used to test if vessel diameter, oxygen saturation, and blood flow changed significantly from rest to isometric exercise. 

Results

Figure 2 shows an example of an oximetry image with corresponding Doppler OCT scans of the peripapillary, peripheral, and macular retinal vessels. Table 2 shows that at rest there was a significantly higher oxygen saturation in the macular than in the peripheral venules (P < 0.001), but no significant difference between this saturation in their arterial counterparts. The table also shows that the macular venules had a smaller diameter (P < 0.001) and a lower blood flow (P = 0.01) than their peripheral counterparts whereas there were no significant differences in these parameters between the macular and peripheral arterioles. However, the counting of vascular branches showed significantly more venules than arterioles originating from the temporal vascular arcades toward the macular area (mean ± SD) 5.67 ± 1.12 versus 2.87 ± 0.78, P < 0.001 whereas there was no significant difference between the number of arteriolar and venular branches toward the retinal periphery (mean ± SD) 3.80 ± 1.52 versus 3.70 ± 1.32, P = 0.79. Therefore, the ratio between the sum of the cross-sectional areas of the studied venules and arterioles was 1.21, both for branches toward the macular area and toward the retinal periphery. 

Figure 2.

 

Retinal oximetry image with color coding of the oxygen saturation of retinal vessels so that the arterioles appear red and the venules greenish. Four Doppler OCT scans are shown from which shaded rhombi extend into the oximetry image to end with the small side corresponding to the scanning plane. In the oximetry image, arteriolar cross sections are marked in red and venular cross sections in light blue. The same colors highlight the areas in which the corresponding vessels are located on the Doppler OCT scans. The color coding of the vessels in the scans refer to the direction of the blood flow in relation to the scanning plane.

Table 2.

 

Oxygen Saturation, Diameter, and Blood Flow of the Four Studied Types of Vascular Branches at Rest (Mean ± SD)

 

Isometric exercise increased the MAP by (mean ± SD) 18.3 ± 6.2 mm Hg (from 91.3 ± 7.3 mm Hg to 109.3 ± 10.8 mm Hg, P < 0.001) and the IOP by 1.1 ± 1.5 mm Hg (from 14.6 ± 2.3 mm Hg to 15.7 ± 2.9 mm Hg, P < 0.001). Table 3 shows that isometric exercise induced a significant contraction of all arteriolar branching types (P < 0.01 for all comparisons) and a significant increase in oxygen saturation of all venular branching types (P < 0.001 for all comparisons). This was accompanied with a significant increase in the blood flow in peripheral arterioles and venules (P = 0.04 for both comparisons), but not in their macular counterparts (P > 0.06 for both comparisons). 

Table 3.

 

Changes in Oxygen Saturation, Diameter, and Blood Flow From Rest to Exercise (Mean ± SD)

 

Discussion

The present study provides new evidence to our understanding of regional variations in retinal oxygen extraction and flow regulation during isometric exercise. The study design required several repetitions of exercise followed by rest with the risk of not reaching a resting condition between the measurements. However, previous studies have shown that 2 minutes of rest is sufficient for the blood pressure to normalize after isometric exercise.^20,21 Therefore, the choice of resting periods lasting 5 minutes was considered to ensure that carry over effects of the blood pressure were avoided. Retinal blood flow was measured by the experimental technique Doppler OCT. The measurements depended on the detection of light scattered from moving blood cells in which the phase shift is limited to values between -π and π. This limitation was experienced in the central part of some arterioles, especially when the blood flow increased secondary to an increase in the arterial blood pressure. However, in the vessels where this artifactual phenomenon was not observed, the blood flow in the peripheral thirds of these vessels could predict the total blood flow that also included the central portion of the vessels. This enabled an estimation of the blood flow in the majority of the vessels in which the phase shift artefact could be observed. The experiments were performed on young healthy subjects because the Doppler OCT method requires a sufficiently high visual acuity to ensure stable fixation, large pupils, and clear refractive media. In spite of this, approximately one out of six scans had to be discarded due to insufficient scan quality. This may point to limitations in the potential for transferring the technique to the study of retinal vascular diseases that affect older persons with media opacities and smaller pupils. 

The validity of the Doppler OCT technique has previously been confirmed by showing that the blood flow measured to enter and leave the eye are comparable,^10,22 and the blood flow observed in the peripapillary vessels in the present study was consistent with these findings in healthy persons.²³ The validity of the findings is also supported by the lack of significant differences between the distances from the optic disc to the branching points of peripheral and macular arterial and venular branches. This argues against differences in intraluminal pressure and diffusion loss of oxygen from the vessels as a source of bias for the findings. 

The study showed that retinal blood flow at rest was lower and the diameter smaller in the macular than in peripheral venules (P > 0.01 for both comparisons) but were comparable in the macular and peripheral arterioles. However, the smaller diameter and lower blood flow of macular venules were counterbalanced by a higher number of these vessels so that the total cross-sectional area of the arterioles as compared to the venules were similar for vessels to the macular area and the retinal periphery. The counts included vessels that were slightly smaller than those in which oxygen saturation and blood flow were measured, which ensured a sufficiently large number for a meaningful statistical analysis to be performed. Ideally, the blood flow and oxygen saturation should have been measured in all these branches, but this could not be achieved because of the limitations in the number of vascular cross sections that could be contained within the Doppler OCT scans and the limitations in the resolution of both oximetry and Doppler OCT.^19,24 On the fundus photographs, it could be observed that the asymmetric branching pattern with a higher number of primary venular than arteriolar branches to the macular area were compensated by more extensive branching of the arterioles than the venules closer to the fovea. This ensured the normal alternating pattern of pre-capillary arterioles supplying and post-capillary venules draining the capillary bed.²⁵ The increase in the arterial blood pressure induced by isometric exercise was within the physiological range and it can therefore be expected that the observed changes in diameter, blood flow, and oxygen saturation had reflected responses that might be observed during normal life. The effect of the intervention on retinal blood flow might potentially be modified by changes in the IOP, but the observed increase in this parameter during exercise was within the known variation of repeated measurements²⁶ and can therefore be considered to have been too low to have affected the results. Although the increase in the arterial blood pressure contracted all the studied arterioles, the contraction was insufficient to maintain the blood flow constant in the peripapillary and peripheral arterioles where a significant increase in the blood flow could be observed. The increased peripapillary and peripheral blood flow during exercise may have resulted in an underestimation of the measured oxygen saturation,¹² which argues that the observation of an increase in venous oxygen saturation in these vessels had been true. In a previous study where isometric exercise for 30 seconds increased the arterial blood pressure by approximately 10 mm Hg, an increased oxygen saturation was only observed in the peripapillary and peripheral venules.¹ However, in the present study, where 2 minutes of exercise increased the arterial blood pressure by approximately 18 mm Hg, the oxygen saturation also increased significantly in the macular venules. Therefore, it is possible that the blood pressure necessary to induce a significant autoregulatory response with derived effects on blood flow and oxygen saturation may differ among peripheral and macular arterioles.²⁷ Such regional differences is supported by previous findings that spontaneous diameter oscillations and ischemic conditioning in the retina in healthy persons may differ among the peripheral and macular vessels.^28–30 The background for these differences, as well as the higher oxygen saturation in the macular than in the peripheral venules at rest¹¹ is unknown, but may be related to the shorter arterio-venous distance through the microcirculation in macular than in the peripheral circulation.^25,31 The observations are also in accordance with previous findings that diameter regulation is affected differently in arterioles supplying the retinal periphery and the macular area in patients with diabetic retinopathy,³² but these studies were not supplemented with measurements of retinal blood flow. This should be the subject of a future study. 

The finding that an increase in the arterial blood pressure can increase the blood flow in the primary vascular branches to the retinal periphery documents a reduced passage time of the blood in these vessels and argues against shunting in these vascular branches as a cause of the increase in venous oxygen saturation. However, it cannot be excluded that shunting may occur at more distal branching levels and the present findings therefore cannot rule out whether the observation is due to bypassing of the capillary bed or an increased passage time that may limit the amount of oxygen that can diffuse out of the vessels. This may also explain that the diameter and blood flow in macular venules were unchanged, whereas the oxygen saturation in these vessels increased. An elucidation of this issue requires methods for studying capillary blood flow in the retinal periphery. 

In conclusion, the study has shown that the branching pattern of arterioles and venules is different in the retinal periphery and the macular area. Increased systemic blood pressure leading to arterial contraction and increased venous oxygen saturation in the retina in normal persons can be accompanied with a significant increase in peripheral retinal blood flow without significant effects on macular blood flow. Differences in vascular structure and blood flow regulation between the retinal periphery and the macular area may help explaining regional differences in the response pattern of retinal vascular disease.^33,34 

Acknowledgments

The skillful assistance of technicians Lars Bisballe and engineer Rene Werkmeister is gratefully acknowledged. 

Supported by the VELUX-Foundation and Fondation Juchum. 

Disclosure: J. Drachmann, None; S.K. Jeppesen, None; T. Bek, None