The Relationship between Retinal Vascular Reactivity and Arteriolar Diameter in Response to Metabolic Provocation

Abstract

purpose. To compare the magnitude of vascular reactivity in response to metabolic provocation in retinal arterioles of varying diameter in healthy young subjects.

methods. Ten healthy young subjects (26.2 ± 3.5 years [mean ± SD]) attended for three sessions. Session 1 was used to select two discrete hemodynamic measurement sites along the superior temporal arteriole. Retinal arteriolar blood flow was assessed at relatively narrow and wide sites. At sessions 2 and 3, CO₂ and O₂ were sequentially administered (and alternated across sessions) using manual gas flow control via a modified sequential rebreathing circuit to achieve target hypercapnia and hyperoxia. Blood flow was assessed for each gas phase. Total vascular reactivity capacity (TVRC) was taken as the difference in flow between hypercapnia and hyperoxia.

results. The baseline diameter for the narrow and wide measurement sites was 92.4 μm (±13.6) and 116.7 μm (±12.7), respectively (ReANOVA; P < 0.0001). Hyperoxia induced a decrease in blood flow, whereas hypercapnia increased flow (P < 0.0001). TVRC was greater for the wide than for the narrow measurement sites (Δ flow narrow = 3.0 μL/min versus Δ flow wide = 6.6 μL/min; P < 0.0001). In terms of percentage change in flow relative to baseline, TVRC was the same between the wide and narrow sites (Δ narrow = 67% versus Δ wide = 61%; P > 0.05).

conclusions. In response to metabolic provocation, absolute TVRC was greater for retinal arteriolar measurement sites with wider baseline vessel diameters. However, percentage change in retinal blood flow was the same irrespective of initial arteriolar diameter.

The eye is often described as a “window into the body” because it uniquely facilitates the assessment of vascular physiology by permitting direct visualization and measurement of the retinal microvasculature. Noninvasive and quantitative measurements of retinal blood flow in combination with standardized vasoactive stimuli have enabled in vivo hemodynamic studies of human retinal vascular bed regulation in health and disease. Hyperoxia (i.e., elevated PaO₂) reduces retinal blood flow, ^{1 2 3 4 5 6 7 8 9 10 11 12 13 14 15} whereas hypercapnia (i.e., elevated PaCO₂) increases flow. ^{2 14 15 16 17 18 19} Vascular reactivity is defined as the magnitude of change of hemodynamic parameters (vessel diameter, velocity, or flow) in response to provocative stimuli, such as a change in the arterial partial pressure of CO₂ (PaCO₂) or O₂ (PaO₂). 

The degree of dilatation or constriction induced by a vasoactive stimulus depends, among other things, on the extent of vasodilatory or vasoconstrictive reserve of the vessels. It is often stated that smaller diameter retinal vessels demonstrate greater magnitude of vascular reactivity than larger vessels. However, this difference is based on the comparison of relatively large and rigid conducting cerebral vessels (approximately 1.5–3.0-mm diameter) and the downstream resistance arterioles (approximately 100-μm diameter). ²⁰ In the retina, some studies have found that smaller diameter venules and arterioles respond to a greater degree to provocation than larger diameter vessels, ^{21 22} whereas other studies have found no difference. ^{23 24 25} Retinal vessel wall pulsatility in humans is greater in smaller diameter retinal vessels, ²⁶ and diameter changes related to myogenic provocation are especially variable in narrow-diameter dissected hamster arterioles. ^{27 28} None of the human retinal hemodynamic studies, however, have systematically assessed the relationship between initial retinal vessel diameter and vascular reactivity in terms of absolute change in blood flow (in real units). ^{21 22 23 24 25} Rather, they have relied on changes in vessel diameter or velocity to reflect change in blood flow. 

Definition of the relationship between change in blood flow and initial vessel diameter is important for understanding the physiology of retinal vascular regulation and functional differences with other vascular beds in the body, for understanding changes in vascular reactivity associated with disease, and ultimately for achieving standardization of this promising functional vascular marker. Unlike previous studies, total vascular reactivity capacity (TVRC) was quantified by calculating the change in retinal blood flow that occurred between sequentially administered hypercapnic and hyperoxic gas challenges. The aim of the study was, therefore, to compare the magnitude of retinal vascular reactivity in response to metabolic provocation at measurement sites of different diameters along the superior temporal arteriole (STA) in healthy young subjects. We hypothesized that the wider diameter measurement site would exhibit a greater magnitude of absolute TVRC than the narrower site considering that it is endowed with a greater number of smooth muscle cells. A secondary aim of the study was to calculate the coefficient of variation (CV) to determine the site, if any, that offers optimal signal detection. 

Materials and Methods

Sample

This study received approval by the University of Waterloo Office of Research Ethics. Informed consent was obtained from each subject after explanation of the nature and possible consequences of the study according to the tenets of the Declaration of Helsinki. The sample comprised one randomly chosen eye from each of five healthy women and five healthy men (mean age, 26.2 years [range, 21–33; SD, 3.5]). All subjects had a logMAR visual acuity of 0.0 or better. Exclusion criteria included any refractive error > ±6.00 diopters (D) sphere or ±1.50 D cylinder, habitual smoking, treatable respiratory disorders (e.g., asthma), cardiovascular disease, systemic hypertension, family history of glaucoma or diabetes, or history of any ocular disease. All participants were asked to abstain from caffeine and red meat for 24 hours before their study visits. An impact of caffeine on cerebral blood flow is well established, ²⁹ and saturated fats are known to influence vascular reactivity. ³⁰  

Visits

Each subject attended for three visits. The first visit was used to determine eligibility, to define the retinal measurement sites for quantitative blood flow assessment, to acquire baseline data, and to familiarize the subjects with the blood flow measurement technique. The second and third visits were used to assess TVRC. The inspired gas stimuli were administered sequentially at each visit, but the order of hyperoxia and hypercapnia was alternated across the two visits (Fig. 1) . The order of stimuli was also alternated between consecutive volunteers. 

Quantitative Assessment of Retinal Arteriolar Blood Flow

The noninvasive, quantitative assessment of retinal arteriolar blood flow was achieved with a flowmeter (Canon Laser Blood Flowmeter, model 100 [CLBF]; Canon, Irvine, CA). Details of the CLBF have been published elsewhere. ^{31 32 33} Briefly, two sequential, bidirectional laser Doppler velocimetry measurements of 2 seconds duration each were collected from the STA using a 675-nm laser (frequency, 50 Hz). Simultaneous vessel diameter measurements were acquired using a rectangular 543-nm laser that was manually orientated at 90° to a relatively straight segment of vessel. Densitometry analysis of the transmittance profile allowed the determination of the vessel diameter. An eye tracker mechanism used the position of the projected 543-nm laser on the retina to stabilize the selected measurement site with respect to involuntary eye movement. Flow values were derived from the velocity and diameter measurements assuming Poiseuille flow and a circular vessel profile. 

Gas Provocation Protocol

The subject breathed through a face mask connected to a sequential gas delivery circuit (Hi-Ox80; Viasys Healthcare, Yorba Linda, CA) modified by the addition of an exhaled gas reservoir. ^{34 35} Medical tape (Tegaderm; 3M Health Care, St. Paul, MN) was applied as necessary to maintain an air-tight seal at the face. Details of the function of the sequential rebreathing circuit have been published elsewhere. ^{27 31 32} The flow of air and oxygen to the sequential gas delivery circuit was controlled manually to achieve the gas challenges. Tidal gas concentrations were continuously sampled using a rapid response critical care gas analyzer (Cardiocap 5; Datex-Ohmeda, Madison, WI). Hemoglobin oxygen saturation via pulse oximetry, respiratory, and pulse rate was also continually recorded. All data outputs were downloaded to an electronic data acquisition system (S5 Collect; Datex-Ohmeda). 

Subjects underwent a gas provocation protocol consisting of seven phases at each of two visits (Fig. 1) . Either of two protocols, repeated hypercapnia or repeated hyperoxia, was undertaken at each visit. Hypercapnia was defined as an increase of the end-tidal Pco ₂ (PetCO₂) by approximately 25% relative to baseline. This was accomplished by reducing the total flow of air to the sequential gas delivery breathing circuit. Reducing the air flow resulted in a proportional increase in the rate of rebreathing of previously exhaled gas, raising the PetCO₂. The reduction in air flow was titrated to the required PetCO₂. Once target PetCO₂ values were attained, maintaining stable gas flows resulted in stable PetCO₂ and PetO₂ for the duration required to acquire the hemodynamic measurements. ³⁶ For the hyperoxic gas phase, switching from air to pure oxygen raised the fractional (percentage) inspired O₂ (FiO₂) to approximately 95%. Isocapnia was maintained and stabilized by restricting the O₂ flow to each subject’s minute ventilation (determined while breathing air at baseline). ³⁶ Hemodynamic measurements were acquired when PetCO₂ and FiO₂ were stable (Fig. 1) . 

At the completion of either gas phase, subjects were returned to breathing of air without restriction to establish recovery values, and retinal blood flow measurements were reacquired when F_ETCO₂ had returned to baseline (Fig. 1) . After the initial gas phase (hypercapnia for vasodilatation or hyperoxia for vasoconstriction) and the subsequent recovery phase, the gas challenge that resulted in the opposite vascular response was initiated, and a second recovery phase was provided. Finally, the initial gas phase was repeated and a third recovery phase was provided. Each of the seven phases was maintained for 10 minutes: the first 3 minutes were used to stabilize tidal gas values, and data were acquired during the remaining 7 minutes of each phase. Volunteers underwent vascular reactivity assessment over two separate visits in which hyperoxia or hypercapnia were initiated after baseline readings (Fig. 1) . The initiation of hyperoxia or hypercapnia after baseline was alternated between consecutive subjects (thereby instituting the repeated hypercapnia, or repeated hyperoxia, protocols) and was alternated between visits 2 and 3 for each subject. 

Comparison of the arteriolar responses with hypercapnia and hyperoxia, or vice versa, permitted assessment of the TVRC. The two gas provocation protocols were designed to permit the assessment of TVRC, irrespective of any persistent effects associated with one or the other of the gas challenges. In a previous related study (Tayyari F, et al., manuscript submitted, 2009), the hyperoxic gas challenge resulted in persistent vasoconstriction, whereas termination of hypercapnia promptly returned the hemodynamics to baseline. As a result, TVRC was taken as the difference in flow between hypercapnia and the subsequent hyperoxia (Fig. 1) . 

Procedures

The pupil of the study eye was dilated using tropicamide 1% (Alcon; Mississauga, Canada). Subjects rested for 10 minutes before the start of each study visit to stabilize baseline cardiovascular and respiratory parameters. At visit 1, two separate measurement locations, representing different diameter values along the STA, were sought for each subject. Relatively straight STA segments, distant from vessel branches, were selected. A wide measurement site was selected on the STA as close to the optic nerve head as possible, and a distal narrower site was selected on the main branch of the STA to provide maximum separation between sites, within the constraint that the diameter was large enough to allow repeated blood flow measurements. There was at least one major bifurcation between the two measurement sites. Arterioles were used in this study because they are responsible, at least in part, for determining vascular resistance in the retina. At all three visits, five retinal blood flow measurements were acquired at the narrow and wide measurement sites for each phase (Fig. 1) . The order of measurement site was randomized between subjects and alternated across visits with respect to proximity from the optic nerve head. Blood pressure was noninvasively recorded every 2.5 minutes over the course of each gas provocation (Cardiocap 5; Datex-Ohmeda). 

CLBF Data Analysis

The CLBF software was used to ensure accurate identification of the original measurement sites selected at visit 1 and was confirmed using a retinal photograph. CLBF analysis software was used to analyze each acquired velocity waveform. All measurements were made by a single experienced operator (FT). The subjects were blinded to the breathing gas mixture composition. Retinal blood flow measurements were made for each condition only when end-tidal gas concentrations were stable. The quality of velocity waveforms was assessed against laboratory-generated agreed-on standards. Using the CLBF eye tracker data feature superimposed on the velocity waveforms, readings with loss of fixation or aberrant velocity waveforms were excluded from the analysis. 

Statistical Analysis

Mean blood flow parameters were calculated for each phase (i.e., baseline, O₂, CO₂, recovery) of each individual (Fig. 1) . Having established the normal distribution of the data, repeated-measures analysis of variance (ReANOVA) was undertaken on the group arteriolar blood flow to determine the significance of any change over the course of each protocol. Arteriolar measurement site (i.e., narrow or wide) was the within-subject factor. In those situations that revealed a significant ANOVA result, post hoc testing was undertaken using Tukey honest significant difference (HSD) test. To test the influence of vessel diameter on the magnitude of retinal vascular reactivity, TVRC was calculated as the difference in flow between hypercapnia and subsequently hyperoxia. The CV for blood flow was calculated for each measurement site using the formula:  

\[\mathrm{CV}{=}(\ \frac{\mathrm{SD}}{\mathrm{mean}}){\times}100\]

 

Results

Baseline Diameter

The group mean baseline diameter for the narrow arteriolar measurement site was 92.4 μm (SD, 13.6 μm), and for the wide measurement site it was 116.7 μm (SD, 12.7 μm; ReANOVA, P < 0.0001). 

Vessel Diameter and Retinal Vascular Reactivity

Hyperoxia induced a decrease in blood flow, whereas hypercapnia increased flow (P < 0.0001), and the combination of measurement site (narrow versus wide) and gas phase (O₂ versus CO₂) was significant (P < 0.0001). TVRC was greater for the wide (absolute Δ flow = 6.6 μL/min ± 3.5) than for the narrow (absolute Δ flow = 3.0 μL/min ± 0.9) measurement sites (ReANOVA, P < 0.0001; Fig. 2 , top). However, TVRC was the same at both sites when normalized for initial vessel diameter (Δ narrow = 67% versus Δ wide = 61%; P > 0.05; Fig. 2 , bottom). There was no effect of visit (P = 0.6504) or visit in combination with measurement site on the vascular reactivity measurements (P = 0.6209). 

Coefficient of Variation

The CVs for arteriolar blood flow for the narrow and wide measurement sites were 31% and 57%, respectively. 

Hyperoxia Repeat Gas Provocation

The first hyperoxic provocation resulted in a decrease in flow relative to the preceding baseline (Tukey HSD, P < 0.0001) and relative to recovery 1 (P < 0.0001) for both measurement sites (Fig. 3) . Hypercapnia resulted in an increase in flow at both the narrow and the wide sites relative to recovery 1 (P < 0.0001) and recovery 2 (narrow, P = 0.0328; wide, P < 0.0001). The repeat hyperoxic provocation resulted in a decrease in flow relative to recovery 2 (P < 0.0001) and relative to recovery 3 (narrow, P = 0.0002; wide, P < 0.0001) for both sites. 

For the hyperoxia repeat protocol, the first hyperoxic provocation resulted in a significant decrease (P < 0.0001) in diameter relative to baseline and recovery 1 for both sites (Fig. 4 , top). Hypercapnia resulted in a significant increase (P < 0.0040) in diameter values for the narrow measurement site relative to recovery 1 and 2. The wide measurement site exhibited significantly larger diameter values in response to hypercapnia relative to recovery 1 (P = 0.0050) but was not different relative to recovery 2 (P ≤ 0.0901). The repeat hyperoxic provocation resulted in a significant decrease (P < 0.0001) in diameter relative to recovery 2 and 3 for both sites. 

For the hypercapnia repeat protocol, the first hyperoxic provocation resulted in a significant decrease (P < 0.0001) in velocity relative to baseline and recovery 1 for both sites (Fig. 5 , top). Hypercapnia resulted in a significant increase (P ≤ 0.0020) in velocity relative to recovery 1 and 2. The repeat hyperoxic provocation resulted in a significant decrease (P < 0.0001) in diameter relative to recovery 2 and 3. 

Hypercapnia Repeat Gas Provocation

The first hypercapnic provocation resulted in an increase in flow relative to baseline (Tukey HSD; narrow, P = 0.0026; wide, P < 0.0001) and relative to recovery 1 (narrow, P = 0.0050; wide, P < 0.0001) for both measurement sites (Fig. 3) . Hyperoxia resulted in a reduction in flow for both sites relative to recovery 1 (P < 0.0001) and recovery 2 (narrow, P = 0.0004; wide, P < 0.0001). The repeat hypercapnic provocation resulted in an increase in flow relative to recovery 2 (narrow, P = 0.0003; wide, P < 0.0001) and relative to recovery 3 (narrow, P = 0.0226; wide, P = 0.0039) for both sites. 

The first hypercapnic provocation resulted in a significant increase (P < 0.0444) in diameter relative to baseline and recovery 1 (Fig. 4 , bottom). Hyperoxia resulted in a significant reduction (P < 0.0001) in diameter relative to recovery 1 and 2. The repeat hypercapnic provocation was not different relative to recovery 2 and 3. 

For hypercapnia repeat protocol, the first hypercapnic provocation resulted in a significant increase (P < 0.0070) in velocity relative to baseline and recovery 1 for both sites (Fig. 5 , bottom). Hyperoxia resulted in a significant reduction (P < 0.0020) in velocity relative to recovery 1 and 2 for both sites. The repeat hypercapnic provocation was not different relative to recovery 2 and 3 for both sites. 

Gas Parameters and Blood Pressure

Group mean FiO₂ (± SD) and F_ETCO₂ (± SD) values for each condition and protocol are shown in Tables 1 and 2 . FiO₂ showed a group mean (i.e., mean of the three provocations across both protocols) increase from 19.8% to 93.1% (paired Student’s t-test, P < 0.0001). Importantly, F_ETCO₂ did not change relative to baseline or recovery during hyperoxic provocation. During the hypercapnic phases, F_ETCO₂ showed a group mean increase of 24.3% relative to the group mean of the baselines (from 37.6 mm Hg to 46.8 mm Hg; paired Student’s t-test, P < 0.0001). Group mean arterial blood pressure did not change over the course of each gas provocation. 

Discussion

In agreement with the hypothesis, this study found that retinal arterioles with wider diameters show greater response in terms of absolute TVRC expressed in flow units than do those with narrower arterioles. In other words, the magnitude of vascular reactivity, measured as the difference in flow between CO₂ and O₂ provocation, was greater for arterioles with wider diameters. However, when the absolute reactivity was related to initial vessel diameter to calculate percentage change in flow, both the wide- and the narrow-diameter arterioles were found to be the same, in agreement with most of the previous work on this topic. ^{23 24 25} Unlike previous studies, vascular reactivity was systematically quantified in this study by calculating TVRC. TVRC represents the absolute change in blood flow that occurred between standardized hypercapnia and isocapnic hyperoxia, a combined gas challenge that resulted in substantial vasoconstriction and vasodilation. Arteriolar diameter and centerline blood velocity, which determine retinal blood flow, demonstrated results similar to those of blood flow, including the finding that hypercapnia resulted in a less pronounced vascular reactivity response than hyperoxia. 

Smaller diameter vessels are thought to demonstrate a greater magnitude of vascular reactivity than larger diameter vessels in the cerebral vasculature. ²⁰ However, this is based on the comparison of large conducting vessels with the downstream smaller resistance arterioles. Theoretically, the large conducting vessels require greater tension on the walls to induce a given diameter change because they have greater diameters and thicker walls (La Place’s Law) ³⁷ and therefore can be expected to manifest a blunted response to vasoactive stimuli. Indeed, intravascular measurements suggest that vascular resistance upstream of the capillary bed is equally divided between the small resistance arteries (approximately 500 μm diameter) and arterioles (approximately 100 μm diameter), with little contribution from the conducting vessels. ³⁸ Conversely, the retinal arterioles have relatively thin walls and are endowed with layers of contractile smooth muscle cells, controlling blood flow to the downstream capillary bed. The numbers of smooth muscle cells in the wall of the retinal arterioles progressively reduce from the optic nerve head along the course of the arterioles, ³⁹ suggesting that the wider diameter retinal arterioles exhibit greater vascular reactivity reserve than do narrower arterioles. However, the percentage magnitude of vascular reactivity might be anticipated to be equivalent in retinal arterioles of different diameters because the retinal circulation in healthy persons is a closed system and any percentage change in flow in the upstream arteriole would be expected to be reflected in the downstream arteriole. 

The results of this study agreed with most of the previous work on this topic ^{23 24 25} because the percentage change in vascular reactivity of the retinal arterioles was found to be the same irrespective of initial vessel diameter. In contrast to previous work, however, this study specifically and systematically addressed the relationship between initial retinal vessel diameter and vascular reactivity, in terms of absolute change in blood flow (derived by the simultaneous measurement of vessel diameter and blood velocity) to inhaled gas stimuli. Indeed, the differences between the results of this study and previous work ^{21 26} may also be explained by differences in the techniques used to assess hemodynamics and to provoke vascular reactivity. For example, blood flow estimation techniques that measure surrogate parameters to indicate flow often present data in terms of percentage change relative to baseline, which will reveal a different result when expressed as absolute flow, as demonstrated in this study. Kiss et al. ²¹ showed an inverse relationship between magnitude of vascular reactivity and vessel diameter in response to 100% O₂ breathing using the blue field entoptic phenomenon and the retinal vessel analyzer. The subjective nature, however, of the blue field entoptic phenomenon results in high measurement variability. Furthermore, Jeppesen et al. ²² reported that elevation of blood pressure induces a retinal arteriolar vasoconstriction that increases with decreasing arteriolar diameter. However, the physiological vascular response to change in blood pressure in the retinal arterioles may be different from the response to inhaled gas stimuli. A more recent study ²⁵ from our group found that the percentage increase in flow in response to an isoxic hypercapnic stimulus was the same in retinal arterioles and capillaries, consistent with the findings of this study. 

In answering the second aim of the study, we found a more favorable CV for the narrower measurement site than for the wider site (i.e., 31% vs. 57%), which suggests that improved signal detection may be obtained from narrower arterioles. In other words, the greater ratio of vascular reactivity to measurement variability in the smaller diameter retinal vessels indicates that the study of these vessels will provide the more accurate measures of absolute blood flow. However, previous work found that vessel wall pulsatility is greater in smaller diameter retinal vessels ²⁶ which, in turn, might be anticipated to generate a higher CV. The answer to this question, which requires further investigation, has important implications in terms of sensitivity to detect change of blood flow in retinal disease. 

A limitation of this study was that the arterial gas concentration transition from one phase to the next was not square wave because a manual gas delivery technique was used. However, the slope of the gas transition was expected to be similar between phases. To further negate this potential limitation, a minimum of 3 minutes was allowed after the initiation of each phase before blood flow measurement. Another limitation was that isoxic control of end-tidal gases was not attempted during the hypercapnic stimulus. However, the use of a sequential rebreathing circuit minimizes concomitant change in Po ₂ during hypercapnia. ²⁸ The design of this study used two different locations along the same arteriole. Retinal vessels have been shown to respond uniformly in terms of change in diameter to hyperoxia across quadrants in healthy adults. ⁴⁰ From a practical perspective, the lower range of retinal arteriolar diameters that could be investigated was limited by the ability of the CLBF eye tracking system to repeatedly and reliably stabilize the narrower measurement site. In turn, this limited the overall difference in diameter values between the narrow and wide measurement sites. Although definitive results and conclusions arose from the study, we were unable to exclude the possibility that smaller diameter retinal arterioles (i.e., <60 μm) might have an enhanced TVRC, as found in coronary arteries. ⁴¹  

In summary, this study demonstrated that the magnitude of retinal vascular reactivity in response to metabolic provocation was greater for arteriolar measurement sites with wider baseline vessel diameters; however, in terms of percentage change in blood flow, the responses of the retinal arterioles were the same at the narrow and wide measurement sites. In addition, a more favorable normalized variation of absolute vascular reactivity measurement was found for the narrow, compared with the wide, site. 

 Supported by the Canadian Institutes of Health Research Operating Grant (CH) and New Investigator Award (CH), the Canada Foundation for Innovation, and the Ontario Innovation Trust New Opportunities Infrastructure Grant (CH).

 Submitted for publication January 1, 2009; revised February 16 and March 11, 2009; accepted August 14, 2009.

 Disclosure: F. Tayyari, None; S.T. Venkataraman, None; E.D. Gilmore, None; T. Wong, None; J. Fisher, None; C. Hudson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Hyperoxia repeat protocol (top) and hypercapnia repeat protocol (bottom) used at visits 2 and 3. One narrow- and one relatively wide-diameter measurement site along the superior temporal arteriole were selected for each individual. At least 3 minutes were allowed to ensure stabilization of tidal gas values after initiation of each condition before data acquisition, which occurred in the final 7 minutes of each gas phase. O₂ and CO₂ order was alternated between volunteers and across visits. The hyperoxic gas challenge resulted in a persistent vasoconstrictive effect. ³⁴ As a result, TVRC was assessed as the difference in response to CO₂ and subsequently O₂ (i.e., “b” for the hyperoxia repeat and “a” for the hypercapnia repeat protocols, respectively). F_ETCO₂, fractional end-tidal CO₂ relative to baseline; FiO₂, fractional percentage inspired O₂; O₂, isocapnic hyperoxia; CO₂, hypercapnia; Recov, recovery.

Figure 2.

 

Baseline arteriolar diameters as a function of absolute TVRC (top) and as a function of percentage TVRC (bottom; i.e., difference in flow between CO₂ and subsequently O₂ provocation) for the narrow and wide arteriolar measurement sites. C losed circles: wide measurement site; open squares: arrow measurement site.

Figure 3.

 

Box plots illustrating group change in retinal arteriolar blood flow during the hyperoxia (top) and hypercapnia (bottom) repeat protocol for the wide (left) and narrow (right) measurement sites. The legend within the box represents the median value, the upper and lower extremes of the box represent the 25th and 75th percentiles, and the error bars represent the nonoutlier range. C ircle: outlier.

Figure 4.

 

Box plots illustrating group change in arteriolar diameter during the hyperoxia (top) and hypercapnia (bottom) repeat protocol for the wide (left) and narrow (right) measurement sites. The legend within the box represents the median value, the upper and lower extremes of the box represent the 25th and 75th percentiles, and the error bars represent the nonoutlier range.

Figure 5.

 

Box plots illustrating group change in centerline blood velocity during the hyperoxia (top) and hypercapnia (bottom) repeat protocol for the wide (left) and narrow (right) measurement sites. The legend within the box represents the median value, the upper and lower extremes of the box represent the 25th and 75th percentiles, and the error bars represent the nonoutlier range. C ircles: outliers.

Table 1.

 

Fractional End-Tidal CO₂ and Fractional Inspired O₂ as a Function of Gas Phase for the Hyperoxia Repeat Protocol

	Baseline	O2	Recovery	CO2	Recovery	O2	Recovery
FETCO2 (%)	5.01 (0.37)	5.00 (0.12)	4.90 (0.20)	6.12 (0.40)	5.01 (0.13)	5.01 (0.12)	4.99 (0.15)
FiO2 (%)	20.02 (0.16)	93.39 (1.14)	19.96 (0.23)	18.90 (0.87)	19.76 (0.32)	92.43 (2.04)	19.89 (0.32)

 F_ETCO₂ is the partial pressure or maximal concentration of carbon dioxide (CO₂) at the end of an exhaled breath expressed as a percentage of CO₂. FiO₂ is the fractional concentration of inspired O₂ expressed as a percentage of O₂.

Table 2.

 

Fractional End-Tidal CO₂ and Fractional Inspired O₂ as a Function of Gas Phase for the Hypercapnia Repeat Protocol

	Baseline	CO2	Recovery	O2	Recovery	CO2	Recovery
FETCO2 (%)	4.89 (0.18)	6.11 (0.35)	5.08 (0.15)	5.00 (0.12)	4.92 (0.17)	6.06 (0.34)	5.12 (0.24)
FiO2 (%)	19.79 (0.45)	18.18 (1.12)	19.58 (0.84)	93.49 (1.30)	19.79 (0.24)	18.19 (1.41)	19.75 (0.41)

 F_ETCO₂ and FiO₂ as defined in Table 1 .

The authors thank Erin Harvey (Department of Statistics and Actuarial Science, University of Waterloo) for assistance in statistical analysis and Edmund Tsui for assistance with the figures.