Effects of Oxygen and Carbogen Breathing on Choroidal Hemodynamics in Humans

Abstract

purpose. Previous studies in humans have demonstrated that oxygen (O₂) reduces retinal vessel caliber and blood flow, whereas carbon dioxide (CO₂) usually has opposite effects. The influence of O₂ and CO₂ on choroidal circulation is not fully understood, however. This study was conducted to determine the effects of systemic hyperoxia and hypercapnia on global choroidal hemodynamics, as evaluated by pulsatile ocular blood flow (POBF) tonography.

methods. In this experiment, 16 and 22 healthy volunteers breathed 100% O₂ and carbogen, respectively. POBF and intraocular pressure (IOP) were measured twice with a UK OBF tonograph for each of the following conditions: in ambient room air breathing, after breathing pure O₂ or carbogen through a face mask, and in ambient room air 10 minutes after mask removal. Heart rate (HR), hemoglobin oxygen saturation level (SaO₂), and systemic arterial blood pressure (BP) were monitored throughout testing. The end-tidal CO₂ (EtCO₂) level and respiratory rate (RR) were also recorded during carbogen breathing.

results. Results revealed that HR was reduced (P < 0.004) and SaO₂ was increased (P = 0.0001) by both oxygen and carbogen breathing. Systemic arterial BP remained stable throughout the experiment. EtCO₂ was increased during carbogen breathing (P = 0.0001), whereas RR was reduced (P = 0.0175). IOP was significantly decreased during both phases of the experiment (P = 0.0001). Finally, POBF was not altered by pure O₂ breathing, but it increased on average by 7.7% during carbogen breathing (P = 0.0222).

conclusions. The data obtained with POBF tonography indicate that the choroid reacts to increased blood CO₂ concentration, but not to systemic hyperoxia, in a manner similar to that in retinal and brain vessels.

The ocular circulation can be divided into two vascular systems: the central retinal artery circulation and the uveal vessels distributed within the iris, ciliary body, and choroid. The former circulatory system nourishes the inner layers of the retina, and the choroidal component of the latter supplies the outer retinal layers. 

The response of the retinal circulation to variations in inspired O₂ and CO₂ concentrations has been widely studied in the past decades in both animals and humans. Hyperoxia is known to constrict retinal vessels ^{1 2 3 4} and to reduce blood flow, ^{2 3 4 5} whereas hypoxia ² and hypercapnia ^{5 6 7 8} dilate retinal vessels and increase blood flow. Carbogen, a gas mixture usually containing 4% to 7% CO₂ and 93% to 96% O₂, has frequently been used in humans for the treatment of central retinal artery occlusion. The assumption is that CO₂ will prevent oxygen-induced vasoconstriction and, therefore, maintain or even increase blood flow while providing the retina with increased oxygenation. ^{1 3 5} In fact, recent studies have shown that carbogen breathing oxygenates the retina better than O₂ alone. ^{9 10} However, retinal hemodynamic results obtained with carbogen are still controversial. Specifically, in some studies carbogen did not reduce the vasoconstrictive effect of pure O₂ breathing, ^{1 3 11} whereas results from other investigations indicate that carbogen decreases retinal blood flow to a lesser extent than 100% O₂ ³ and increases perimacular leukocyte velocity ⁵ and retinal blood velocity. ¹¹  

The control of the choroidal circulation is quite different from that of the retinal vasculature. Choroidal blood flow (ChBF) is very high compared with other vascular beds in the body ⁶ and has long been considered to exceed local metabolic requirements. This high flowrate has, therefore, been attributed a major role in stabilizing the temperature of the eye according to changes in the environment, such as is the case during intense light stimulation. ^{12 13} The work of Linsenmeier ¹⁴ demonstrates, however, that the choroid also plays a vital role in ensuring the proper oxygenation of the distal retina. Linsenmeier ¹⁴ has shown that the choroid must have a very high flow rate to maintain the high venous partial pressure of oxygen that is required to properly nourish the photoreceptors. Several researchers have suggested that variations in choroidal circulation are primarily due to autonomic innervation, as opposed to vascular myogenic factors or metabolism within the outer retina. ^{12 13 15} However, more recent investigations in rabbits demonstrate that myogenic factors may regulate the choroid to minimize arterial pressure changes in choroidal blood volume. ^{16 17 18}  

Studies evaluating the effects of inspired O₂ and CO₂ concentrations on choroidal circulation are limited, especially in humans. Some authors have reported that systemic hyperoxia decreases ChBF in albino rabbits ¹⁹ and cats, ²⁰ and hypercapnia increases ChBF and choroidal blood volume. ^{6 19 20} Carbogen breathing also increases choroidal blood volume and ChBF and reduces choroidal peripheral resistance in albino rabbits. ¹⁹ In the monkey, hyperoxia slightly decreases blood circulation within the choriocapillaris at the level of the posterior pole, whereas CO₂ breathing increases blood flow throughout the choriocapillaris. ²¹ These modifications are likely due to altered vessel resistance within the choroidal arteries during blood gas perturbation. ²²  

In humans, laser Doppler flowmetry measurements indicate that hyperoxia does not affect the foveolar ChBF ²³ but that hypercapnia increases ChBF in the fovea. Other studies measuring fundus pulsations in the macula by laser interferometry have shown that hyperoxia slightly reduces ChBF ^{24 25} but that CO₂, combined with either air ²⁴ or oxygen, ²⁵ increases ChBF in the macula. 

Few studies have investigated more global variations in ChBF in response to changes in the environment. The purpose of this study was to evaluate the global response of the choroidal vasculature to 100% O₂ and carbogen breathing in humans with noninvasive recordings of pulsatile ocular blood flow (POBF). 

Materials and Methods

Subjects

Sixteen healthy subjects (7 men, 9 women) volunteered for the study with pure O₂, and 22 subjects (9 men, 13 women) participated in the experiment with carbogen. The subjects were between 18 and 35 years of age and had no ocular or systemic disease. All subjects had 20/20 or better monocular visual acuity. All aspects of testing conformed to the tenets of the Declaration of Helsinki for the use of humans in research and complied with the ethical guidelines for human experimentation at our institution. Written informed consent was obtained from each subject after the experimental procedures had been clearly explained. 

Pulsatile Ocular Blood Flow and Intraocular Pressure

The POBF was measured with a handheld ocular blood flow pneumatic tonometer (OBF laboratories UK Ltd., Malmesbury, Wiltshire, UK). This instrument allows repeated tonometry recordings at a rate of 200/sec, and it measures dynamic changes in intraocular pressure (IOP) over time. The computerized tonometer system analyzes the various characteristics of each IOP pulse wave and compares them to one another as they are recorded. The pulses are compared for shape, amplitude, systolic and diastolic times, and the base level of IOP. The five most representative pulses are selected and processed to mathematically derive POBF values from the IOP readings. This method provides a high-fidelity measure of the IOP and its time variation. ²⁶  

Before the measurements, each subject was asked to sit on a comfortable chair and rest for 10 to 15 minutes to stabilize the various pressures within the body. The cornea of the right eye was anesthetized with one or two drops of proparacaine HCl 0.5%. Care was taken to ensure that the head was resting straight up against the chair, without any bending of the neck, to avoid any perturbations of blood flow in the vessels of the neck. The head was kept in the same position throughout testing. Subjects were asked to fixate either the red light-emitting diode housed within the tonometer probe or a fixation target on the wall. Two POBF measurements were taken for each of 3 conditions: while subjects were breathing ambient room air (baseline), at the end of a 5-minute period of pure O₂ breathing (O₂ experiment) or at the end of a 10-minute period of carbogen breathing (carbogen experiment) through a face mask, and while the subject was breathing room air, 10 minutes after removing the mask. One or two drops of artificial tears were administered on the surface of the eye before the second measurement for each test condition. This precaution was found to optimize the quality of the results by improving the contact between the probe and the cornea as well as by better preserving the integrity of the corneal epithelium. All manipulations were performed by the same experienced person. 

Phase 1: Oxygen Breathing

Pure O₂ from a cylinder tank was administered through a disposable high-concentration O₂ mask with a one-way valve and a 1-liter reservoir bag (Inspiron, Intertech Resources, Bannockburn, IL). The mask was adjusted as tight and comfortable as possible, and the subject was asked to breathe normally inside the mask. The flow rate of O₂ was adjusted to 5 l/min, a level found to provide adequate ventilation for all subjects. 

Phase 2: Carbogen Breathing

The carbogen used contained a mixture of 5% CO₂ and 95% O₂. High-concentration disposable O₂ masks with a one-way valve and a 1-liter reservoir bag (Inspiron; Intertech Resources) were modified slightly for the carbogen study to further increase their airtightness. The flow rate of carbogen was started at 8.5 l/min and adjusted for each subject according to the speed at which the reservoir bag deflated and inflated during the respiratory cycle. 

End-Tidal CO₂ and Respiratory Rate Monitoring

End-tidal CO₂ (EtCO₂) and respiratory rate (RR) were monitored during the carbogen experiment. A disposable nasal cannula (model No. 1606; Salter Laboratories, Arvin, CA) was connected to a capnograph (EtCO₂/SaO₂ monitor, model 7100, CO₂SMO; Novametrix, Medical Systems, Trudell Medical, London, Ontario, Canada) that monitored and recorded the EtCO₂ levels and RR every 8 seconds. Subjects were asked to breathe as normally as possible by the nose only, to enable a portion of the expired gas to be evacuated by the cannula and then collected and analyzed by the capnograph. The excess in expired gas was evacuated through the one-way valve on the side of the face mask. 

Arterial Blood Pressure, Oxygen Saturation and Heart Rate Monitoring

The systemic arterial blood pressure (BP) was monitored with a 90601 SpaceLabs monitor (Redmond, WA). Two consecutive measurements of systolic, diastolic, and mean arterial BPs were taken 2 to 3 minutes before the POBF measurement within each phase of the experiment. The heart rate (HR) and oxygen saturation level (SaO₂) level were continuously monitored by finger pulse oximetry. 

Time Control Study

To ensure that any changes in the variables measured in the two phases of the experiment were directly related to either O₂ or carbogen breathing, we conducted a time control study on 10 subjects (mean age, 23.1 years; SD, 4.1 years). Each participant underwent the same experimental protocol described above, where POBF (and therefore IOP) was measured (1) while breathing ambient room air, (2) immediately after 10 minutes of wearing the mask, and (3) while breathing ambient air after 10 minutes of mask removal. For this control study, however, no gas was delivered into the face mask, which was slightly modified with an opening in the front to allow subjects to breathe ambient room air. For each of the three phases of this control study, the systemic BP was measured using automated sphygmomanometry just before the POBF/IOP recordings. Two measurements of both the systemic BP and the POBF/IOP were averaged to derive each data point. Throughout the time control evaluation, EtCO₂, RR, SaO₂, and HR were monitored by capnography involving a nasal cannula (for EtCO₂ and RR values) and a finger probe (for SaO₂ and HR recordings). For all four variables measured, an average of the recordings obtained within the last 5 minutes of the evaluation (in all three phases of this control study) was considered representative of the individual variables. By closely monitoring the digital readings on the capnograph, we ensured that the subjects did not breathe their expired air; a digital indicator on the capnograph warned the experimenter each time this happened. For a few subjects, the mask had to be slightly repositioned on a few occasions during the 10-minute breathing period because the indicator signaled that CO₂ was being inspired. The repositioning of the mask cleared that signal, which was on for a few seconds only. 

Data Analyses

The mean of two consecutive measurements of POBF, IOP, and BP was calculated for each subject, and repeated-measures ANOVA was performed across the mean values taken before, during, and after gas inhalation. For the carbogen experiment, data acquired within the last 5 minutes of each condition were averaged and taken as representative values for EtCO₂, RR, SaO₂, and HR. This corresponded to the phase within each test condition in which the EtCO₂ levels were most stable. The mean values were calculated, and repeated-measures ANOVA was performed to compare the results between test conditions. For all variables, the level of significance was set to P < 0.05. 

Results

Effects of Oxygen Breathing

The mean SaO₂ level was 98.0% at baseline, which increased to 99.1% during pure O₂ breathing and returned to baseline values 10 minutes after mask removal (P = 0.0001). The mean HR decreased from 76.0 beats per minute (bpm) at baseline to 73.1 bpm during O₂ breathing (−3.8%) and recovered to baseline within 10 minutes of room air breathing (P = 0.0032). All subjects had normal baseline systolic (average, 121.14 mm Hg), diastolic (average, 74.95 mm Hg), and mean (average, 89.57 mm Hg) arterial BPs, which remained stable throughout the experiment. The IOP decreased from 16 to 14.5 mm Hg during O₂ breathing and was lowered again to 13.4 mm Hg after 10 minutes in room air, after the removal of the mask (P = 0.0001; Fig. 1 ). The mean POBF was 821 μl/min at baseline and did not vary across test conditions (P = 0.5081; Fig. 2 ). 

Effects of Carbogen Breathing

Carbogen breathing was well tolerated by all subjects. The mean SaO₂ level increased from 97.3% to 98.5% with carbogen breathing and recovered to baseline values within 10 minutes of ambient room air breathing (P = 0.0001). The EtCO₂ level increased by a mean of 10.1% during carbogen breathing and immediately decreased when the mask was removed, to return to its baseline level after 10 minutes in room air (P = 0.0001). 

The RR was reduced from 15.7 to 14.2 breaths/min (−9.6%) with carbogen and recovered to 15.5 breaths/min after 10 minutes in ambient air (P = 0.0175). The mean HR was reduced from 75.4 to 72.8 bpm (−3.5%) during carbogen inhalation and returned to its baseline value 10 minutes after mask removal (P = 0.0005). All subjects had normal baseline systolic (average, 120.66 mm Hg), diastolic (average, 71.8 mHg), and mean (average, 86.84 mm Hg) arterial BP levels that remained stable throughout the experiment. 

The mean IOP was reduced from 15.2 mm Hg in ambient air to 14.0 mm Hg during carbogen inhalation, and it was still low (13.3 mm Hg) 10 minutes after the end of carbogen breathing (P = 0.0001; Fig. 1 ). The mean POBF at baseline was 822 μl/min. The POBF increased on average by 7.7% with carbogen breathing (P = 0.0222; Fig. 2 ). The POBF values at the end of the experiment were not different from those recorded during baseline and carbogen breathing conditions. 

Time Control Study

The averaged data (n = 10) for POBF; EtCO₂; RR; systemic, diastolic, and mean BP; SaO₂; and HR monitored throughout the time control study did not vary (P ≥ 0.08). The systemic systolic BP decreased (P = 0.0174) from 119.0 to 116.15 mm Hg during the period when the face mask was worn, and it returned to baseline (119.4 mm Hg) 10 minutes after mask removal. The IOP changed (P = 0.0001) from 14.8 mm Hg (at baseline) to 13.1 mm Hg (with face mask), and it was 12.4 mm Hg at the end of 10 minutes of ambient room air breathing after the mask was removed. 

Discussion

In the present study, we found that SaO₂ levels increased with both O₂ and carbogen breathing, indicating that systemic hyperoxia was achieved in both experimental conditions. Our results also revealed that increases in EtCO₂ concentration during carbogen breathing differed among subjects, which is likely due to intersubject variations in response to this specific gas. ²⁷ Typically, EtCO₂ levels rose quickly at the onset of carbogen inhalation, then declined slightly during the remaining 10-minute period of carbogen breathing. This may be due to the Haldane effect, whereby an increase in blood oxygen partial pressure (Po ₂) decreases the affinity of hemoglobin for CO₂. ²⁸  

The RR was not monitored in the first experiment involving pure O₂ breathing, but it was found to be reduced during the carbogen breathing condition. This finding is probably related to a high O₂ content in the inspired gas. Hyperoxia is known to reduce the RR and tidal volume, whereas CO₂ normally increases these two variables to increase the minute ventilation and better eliminate the excess of CO₂. ²⁸ In the present study, although carbogen produced a significant decrease in RR, eight subjects reported that their breaths seemed to be amplified during carbogen inhalation. The minute ventilation was not monitored, but it may be that the decrease in RR was compensated for by an increase in tidal volume during carbogen breathing. 

Consistent with our findings, a decrease in HR during O₂ breathing has been reported by investigators previously, ²⁵ but it is not a universal finding. ^{29 30} Some studies have also indicated that an increase in CO₂ concentration in inspired air is accompanied by an increase in HR, ^{7 30 31} although other reports have shown that changes in CO₂ levels do not affect HR. ⁸ These previous experiments were not conducted with carbogen, however, which means that there was a failure to induce hyperoxia and may thus explain the discrepant findings. Moreover, in the present study, the stimulating effect of CO₂ on HR may likely have been inhibited by the carbogen-induced increase in SaO₂ concentration. 

In agreement with previous findings, our data show that arterial BP did not vary with O₂ breathing. ^{29 30} Several investigations in which BP levels have been monitored during hypercapnia not accompanied by hyperoxia have revealed that BP remains unchanged, ^{8 30} or increases slightly, ^{7 24} during this manipulation. Data on BP variations with carbogen breathing are limited, however. Schmetterer et al. ²⁵ found an increase in the mean arterial BP, and in a parallel experiment, ³² we found that 10 minutes of carbogen breathing increased diastolic, but not systolic, BP. 

The presently observed decrease in IOP is probably due to both O₂ breathing and repeated tonometry measurements. A reduction in IOP with O₂ breathing has been demonstrated in humans ^{33 34} and rabbits. ³⁴ However, similar results have not been reported by other investigators. ^{24 29} In a control study (n = 20) using a single Perkins measurement and, therefore, minimizing the ocular massaging effect of repeated tonometry, we also found a 1–mm Hg reduction in IOP after a 5-minute period of 100% O₂ breathing (P = 0.007), suggesting that at least a small portion of the IOP changes are related to O₂ levels. Furthermore, the results obtained in our time control study in which only ambient air was used reveal a decrease in IOP after repeated tonometry. This clearly demonstrates the importance of the massaging effect of repeated tonometry on the reduction of baseline IOP. 

The mean POBF values for our baseline conditions (experiment 1: 821μ l/min; experiment 2: 822 μl/min) were in agreement with published normative data obtained with the same system. ³⁵ Our data indicate that systemic hyperoxia does not alter POBF. This parallels previous results in cats ³⁶ but is in disaccord with other data showing that hyperoxia decreases ChBF. ^{19 20} Studies in humans that investigated the effect of systemic hyperoxia on a more focal aspect of choroidal circulation have demonstrated either no change ²³ or a slight reduction ^{24 25} in ChBF within the macular area. In the monkey, hyperoxygenation has been found to alter choriocapillary blood circulation at the posterior pole only. ²¹ The POBF technique used in the current investigation evaluates the global response of ocular blood flow and would likely not detect any perturbations restricted to a small area of the fundus. 

The response of POBF to increased CO₂ concentrations in inspired air that we have demonstrated in this experiment parallels the findings of previous studies investigating ChBF levels within the macula in humans ^{23 24 25} as well as most studies evaluating choroidal circulation in animals, ^{6 19 20 21 22} with the exception of an investigation by Goldstick and Ernest ³⁶ showing that CO₂ has little effect on choroidal circulation. Importantly, the systemic systolic and diastolic BPs did not vary throughout the experimental protocol, indicating that the changes in blood flow can be safely attributed to alterations taking place in the ocular circulation. 

The POBF used in this study is believed to reflect the pulsatile component of the total ChBF. The proportion of pulsatile to nonpulsatile flow has not been clearly identified, but the pulsatile component is believed to approximate 50% to 80% of the total flow. ^{37 38 39} Although POBF tonography is a technology that remains controversial for some, ⁴⁰ it provides valid measurements of the pulsatile component of ocular blood flow ²⁶ and can be safely and noninvasively applied in human research. POBF is dependent on arterial pulse pressure amplitude and pulse rate. ⁴¹ In this study, the various systemic physiological variables measured did not change throughout the experimental protocol. Any alterations in POBF that are associated with O₂ or carbogen breathing may, therefore, be attributed to changes in the ocular, rather than the systemic, circulation. 

Within the limitations of the instrumentation and experimental protocol used to measure ocular blood flow, our results pertaining to POBF levels represent new data that confirm that the choroid reacts like the retinal, ^{5 6 7 8 11} optic nerve head, ³⁰ and cerebral blood ^{42 43} vessels, as well as the vasculature of other organs, ⁶ to an increase in blood CO₂ content. The exact mechanism underlying this vasodilation, and consequent increase in blood flow, remains unknown but may likely involve a reduction in arterial pH. ^{6 25 44 45}  

Carbogen-induced vasodilation and increased blood flow in the choroidal vascular bed may increase O₂ diffusion from the choroid through the outer retina. This would provide a better oxygenation to inner retinal layers compared with pure O₂, ^{9 10} which is not efficient in supplying adequate oxygenation to the whole retina when retinal circulation is absent or compromised. ^{10 46 47}  

Animal models investigating retinal Po ₂, arterial and local pH, electrical activity of the retina, and ocular circulation during the inhalation of different CO₂/O₂ concentrations may help to determine the relationships between these factors and improve our knowledge of the influence of O₂ and carbogen on both ocular circulation and oxygenation of the retina. This may lead to improved therapeutic modalities for various vascular/ischemic ocular disorders. One should be careful, however, when comparing results obtained from animal versus human studies. Typically, animal studies are performed under systemic anesthesia, which may introduce a bias in experimental designs, especially when the vascular system is being examined. It has been shown that some anesthetic agents impair blood flow ^{48 49} and O₂ delivery to tissues, ⁴⁹ even when mechanical ventilation is provided. Lee et al. ⁵⁰ have demonstrated that anesthesia also alters the cerebrovascular response to CO₂ inhalation. Therefore, the constant evolution of noninvasive techniques, such as POBF tonography, laser interferometry, and laser Doppler flowmetry, are instrumental in investigating ChBF in humans. Parallel objective and noninvasive evaluations of visual function and ocular structure, such as those provided by electrophysiological recordings or scanning laser tomography, are also important to correlate visual function and ocular structure with altered ocular hemodynamics. 

In summary, our results obtained with POBF tonography indicate that the choroid reacts in a fashion similar to that of retinal and brain vessels to increased blood CO₂ concentration, but not to systemic hyperoxia. 

 Supported by Natural Sciences and Engineering Research Council operating and equipment grants to HK, by a Fonds de la Recherche en Santé du Québec postgraduate fellowship to CF, and by the Canadian Optometric Education Trust Fund to CF and HK.

 Submitted for publication November 11, 1998; revised April 19, 1999; accepted May 27, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Level of IOP for the different testing conditions during both pure O₂ and carbogen breathing. Baseline indicates the first recording in ambient room air, test condition indicates results obtained in either the O₂ or carbogen breathing period, and recovery indicates results obtained in room air 10 minutes later. Error bars represent one SEM; asterisks indicate data differing from baseline.

Figure 2.

 

Level of POBF for each test condition during both pure O₂ and carbogen breathing. Baseline indicates the first recording in ambient room air, test condition represents the results obtained in either the O₂ or carbogen breathing period, and recovery indicates results obtained in room air 10 minutes later. Error bars represent one SEM; asterisks indicate data differing from baseline.

The participation of all subjects is gratefully acknowledged.