Divers and hyperbaric chamber attendants breathe hyperbaric air routinely. Hyperbaric oxygen (HBO2) is used therapeutically frequently. Although much is understood about the hemodynamic physiology and gas exchange effects during hyperbaric air and HBO2 exposure, arterial and pulmonary arterial (PA) catheter data, including blood gas values during hyperbaric air and HBO2 exposure of normal humans, have not been reported. We exposed 10 healthy volunteers instrumented with arterial and PA catheters to air at 0.85, 3.0, 2.5, 2.0, 1.3 (decompression stop), 1.12 (decompression stop), and 0.85 atm abs (our altitude) and then at identical pressures breathing O2 followed by atmospheric pressure air while we measured arterial and PA pressures (PAP), cardiac output (Q̇), and blood gas measurements from both arterial and PA catheters. Although hemodynamic changes occurred during exposure to both hyperbaric air and HBO2, we observed a greater magnitude of change under HBO2 conditions: heart rate changes ranged from −9 to −19% (air to O2), respiratory rate from −12 to −17%, Q̇ from −7 to −18%, PAP from −18 to −19%, pulmonary vascular resistance from −38 to −48%, and right-to-left shunt fraction from −87 to −107%. Mixed venous CO2 fell 8% from baseline during HBO2 despite mixed venous O2 tensions of several hundred Torr. The stroke volume, O2 delivery, and O2 consumption did not change across exposures. The arterial and mixed venous partial pressures of O2 and contents were elevated, as predicted. O2 extraction increased 37% during HBO2.
- hyperbaric oxygen
- pulmonary arterial catheter
- cardiac output
- oxygen extraction
thousands of patients have been treated with hyperbaric oxygen (HBO2); however, the precise arterial and pulmonary arterial (PA) hemodynamic and gas exchange responses that occur during HBO2 are not clearly known. Determining precise hemodynamic and gas exchange information requires the subject or patient to be instrumented with arterial and PA (Swan-Ganz) catheters, permitting measurement of intravascular pressures and blood gases, and PA catheter data collected in normal humans exposed to HBO2 has been reported only once (24).
It has long been established that heart rate (HR) falls during HBO2 (3). Forty years ago, investigators reported that this reduction in HR was associated with decreased cardiac output (Q̇) (48), an observation supported by subsequent research (Table 1) (21, 24, 27, 29, 48). Generally accepted hemodynamic effects of HBO2 now “include mild bradycardia, leading to a proportional decline in Q̇ and a small increase in systemic vascular resistance” (28).
As Q̇ falls, O2 consumption (V̇o2) may remain constant if O2 extraction increases proportionally. No report of V̇o2 or O2 extraction under hyperbaric conditions has been published.
HBO2 can precipitate acute pulmonary edema in patients with reduced left ventricular function (43), possibly due to high O2 concentrations reducing myocardial relaxation factor (nitric oxide), thereby making the heart stiffer (23), or by an increase in afterload from increased systemic vascular resistance (48). The systematic recording of intracardiac filling pressures of humans during HBO2 exposure is limited to measurements at 3.0 atmospheres absolute (atm abs; 304 kPa) (24).
A wider alveolar-to-arterial Po2 difference (A-aDo2) than predicted has been observed in normal humans exposed to HBO2 (45). These data suggest an unexplained increase in right-to-left intrapulmonary shunt fraction (Q̇s/Q̇t), or venous admixture, due to HBO2, but Q̇s/Q̇t measurements during HBO2 have not been published.
A description of the technique of using a PA catheter in patients treated in the monoplace hyperbaric chamber is available (41), and hemodynamic data obtained during HBO2 therapy in critically ill patients have been presented (40), but these data from critically ill patients may not extrapolate to hemodynamic effects due to HBO2 in normal humans.
The purpose of the present study was to measure the hemodynamic and blood gas responses of normal humans instrumented with arterial and PA catheters under hyperbaric air and HBO2 conditions.
The Institutional Review Board at the LDS Hospital approved the study protocol. After signing informed consent and completing a prehyperbaric exposure history and physical examination, 10 normal, healthy subjects (5 males and 5 females) between the ages of 18 and 41 yr were selected for participation in this study. None of the subjects had evidence of cardiac or pulmonary disease. None of the subjects smoked or were taking medications, including birth control pills. At least 24 h before undergoing catheter cannulation, each subject was compressed in a monoplace hyperbaric chamber (model 2500B; Sechrist Industries, Anaheim, CA) to 3.0 atm abs to be confident the subject could equalize middle ear pressure and could tolerate the confines of the chamber.
The subjects were encouraged to eat breakfast on the day of the experiment. No further nutrition was allowed during the data collection period. Water intake was permitted and was quantified.
The subject was placed supine and had five-lead electrocardiography (ECG) monitoring (GE Healthcare, Little Chalfont, UK). One percent lidocaine without epinephrine (1 ml) was injected as a local anesthetic in the dermis and subcutaneous space of the right subclavian cannulation site. A PA catheter (8 Fr, continuous cardiac output and mixed venous Oximetrix with extra infusion port; Baxter Healthcare, Irvine, CA) was placed through a 9-Fr Introducer sheath (Arrow International, Reading, PA) inserted in the right subclavian vein following standard aseptic techniques. The PA catheter distal port was connected to a pressure transducer (Transpac II disposable transducer; Abbott Laboratories, North Chicago, IL) permitting measurement and monitor display of the distal port pressure waveforms. Appropriate PA catheter position was confirmed by demonstrating a PA occlusion pressure waveform (wedge pressure) with a concomitant >10% rise in mixed venous pulmonary arterial O2 saturation (Sv̄O2) (i.e., capillary gas) (14). Anterior-to-posterior chest radiographs performed on all subjects demonstrated acceptable PA catheter positions and no evidence of pneumothorax or hemorrhage. During chamber pressurization, pressure transducers were located inside the chamber.
Thermal dilution Q̇ was measured by using a closed-loop injectable delivery system, using iced saline as injectate, with in-line temperature sensor (Baxter Healthcare). Continuous Q̇ was measured using a continuous cardiac output catheter system (Baxter Healthcare).
A radial arterial catheter (A-Line RA-04020; Arrow International) was placed aseptically following standard hospital procedures. In seven subjects, an intra-arterial continuous O2 sensor (Continucath 1000; Biomedical Sensors, Malvern, PA) was inserted through the arterial cannulas to permit continuous measurement of arterial O2 tension (PaO2).
After placement of the invasive catheters, the subject rested supine within the monoplace hyperbaric chamber. Detailed methods describing how the catheters were passed out of the chamber, including how the transducers were connected, were presented in a prior report (41). After passing the appropriate leads, tubes, and wires through special pass-throughs in the monoplace chamber hatch (39, 41), the continuous Q̇ system was activated and the Sv̄O2 and Continucath were calibrated, following the manufacturer's recommendations. The arterial catheter was continually flushed with saline containing heparin (10 U/ml saline) at 3 ml/h by the continuous flush device (Abbott Laboratories). A special bracket was constructed to hold the right atrial (RA) and PA pressure transducers in a fixed position relative to the subject. All transducers were zeroed to the RA reference level for each subject.
After all equipment was set up and functioning properly with the subject inside the chamber and the chamber hatch slightly open, the lights in the room were dimmed. The subjects were kept in a dim room in a resting state with as little external activity and conversation as possible to minimize external stimuli. The respiratory quotient (RQ) was measured by collecting exhaled gas from each subject after he/she quietly rested inside the chamber for at least 10 min (38).
Pressure measurements from the arterial catheter and RA and PA ports of the PA catheter were recorded with a strip chart (model 2200S; Gould, Cleveland, OH). Adequate frequency response was confirmed with both the arterial and PA catheters before data collection for all subjects (15).
All thermal dilution saline injection Q̇ measurements (10 ml injected with in-line temperature measurements) were performed by one individual with extensive bedside critical care nursing experience using iced normal saline for bolus Q̇ measurements injected through the injection port of the PA catheter, located outside the chamber (41). To verify reproducible Q̇ measurements, we plotted the injection pressure vs. time of injection. A pressure transducer was connected with a four-foot luer-lock neonatal high-pressure monitoring line (Argon Medical Devices, Athens, TX) to a four-way stopcock (Baxter Healthcare), in-line with the Q̇ syringe. A computer displayed a plot in real time of injection pressure vs. time. Successive Q̇ measurements overlaid the previous displays on the computer screen. Five Q̇ injections were performed at each sampling time. The three that had virtually identical pressure vs. time curves were selected for averaging as the thermal dilution Q̇ value. The Baxter monitor displayed a conventional change in temperature vs. time with each of the injections of iced saline, and for selection, these individual Q̇ measures also fell within ±10% of one another.
Collected baseline data included body surface area (BSA), subject core temperature as measured by PA catheter (T), HR, respiratory rate (RR), systemic blood pressure (BP), mean systemic blood pressure (MBP), PA pressure (PAP), mean PA pressure (MPAP), RA pressure (RAP), PA wedge pressure (balloon occlusion pressure; PAWP), RQ, arterial blood gas data including arterial pH (pHa), arterial partial pressure of CO2 (PaCO2), PaO2, arterial O2 saturation (SaO2), arterial hemoglobin (Hba), mixed venous pH (pHv̄), mixed venous partial pressure of CO2 (Pv̄CO2), mixed venous partial pressure of O2 (Pv̄O2), Sv̄O2, mixed venous hemoglobin (Hbv̄), Q̇ by thermal dilution, and cardiac index (CI). Calculated baseline data include stroke volume (SV), right-to-left shunt fraction (Q̇s/Q̇t) (47), systemic vascular resistance (SVR), and pulmonary vascular resistance (PVR) (25), O2 content by arterial (CaO2) and mixed venous (Cv̄O2) measurement, O2 extraction (CaO2 − Cv̄O2), O2 delivery (Q̇O2), and V̇o2.
The subjects were compressed on air in the monoplace hyperbaric chamber to 3.0 atm abs (304 kPa), 2.5 atm abs (253 kPa), 2.0 atm abs (203 kPa), 1.3 atm abs (132 kPa), 1.12 atm abs (113 kPa), and then back to 0.85 atm abs (86 kPa) (barometric pressure at our altitude of 1,500 m above sea level) (Fig. 1). Upon completion of the first compression sequence with air and data gathering, the subject breathed 100% O2 via a 100% O2 reservoir bag through a one-way flow-directed valve (model 2600; Hans Rudolph, Kansas City, MO) fitted with a SCUBA mouthpiece while wearing a nose clip (chamber hatch slightly open). After the chamber hatch was closed, each subject continued to breathe 100% O2 by demand regulator (model L451; Life Support Products, St. Louis, MO) supplied with 100% O2 inside the chamber until we verified that the chamber gas concentration was ≥98% O2. The chamber gas concentration was measured by placing a gas sampling line just above the subject's face, passing chamber gas out through a pass-through in the chamber hatch while measuring O2 concentration with an O2 analyzer, calibrated according to the manufacturer's recommendations. Once the chamber O2 concentration measured ≥98%, the subject discontinued use of the demand regulator and breathed chamber O2. The compression and data collection sequence was repeated in the same fashion on 100% O2 (Fig. 1).
This compression sequence maximized the hemodynamic and blood gas data obtained while minimizing risk of decompression sickness to the subjects. The air compression was required to separate effects due to hyperbaric pressure from those due to 100% O2. The decompression stops and times (1.3 and 1.12 atm abs, respectively) for the air compression were derived from the US Navy Dive Manual (9) using cross-corrections for altitude “diving” (1,500 m above sea level) (4).
Hemodynamic data were obtained at all pressure levels except 1.3 atm abs pressure, after 10 min at each pressure exposure and 30 min postcompression for both the air and O2 exposures (Fig. 1). Arterial and PA mixed venous blood were simultaneously sampled at each data point. The blood gases were measured with an ABL 330 (Radiometer, Copenhagen, Denmark) located at the chamber side using methodology and instrumentation previously demonstrated to be valid (45, 46). The O2 values were adjusted for blood using methods previously described (45, 46). Hemoglobin, carboxyhemoglobin, methemoglobin, and oxyhemoglobin were measured with an OSM3 (Radiometer). Data collected included T, HR, RR, BP, MBP, PAP, MPAP, RAP, PAWP, pHa, PaCO2, PaO2, SaO2, Hba, pHv̄, Pv̄CO2, Pv̄O2, Sv̄O2, Hbv̄, and Q̇ by bolus injection thermal dilution, CI, and Q̇ and PaO2 by continuous measurement. SV, Q̇s/Q̇t, SVR, PVR, CaO2, Cv̄O2, CaO2 − Cv̄O2, Q̇O2, and V̇o2 were calculated for each data collection point.
Data for each hemodynamic parameter were tallied on a universal flow sheet and subsequently entered into a computerized database, with double checking for accuracy. The chamber T, HR, BP, RAP, and PAP values were continuously measured and stored every 5 min automatically by the HELP system at LDS Hospital (16, 17).
There was an obligate blood loss of ∼150–200 ml per subject throughout the entire investigation. The subjects received ∼800–1,200 ml of intravenous normal saline due to the number of bolus injection Q̇ measurements (5 injections of 10 ml each at each data-gathering condition) during the experiment. The subjects were weighed before and after the experiment. Urine output was quantified.
After compression with O2, data were recorded 30 min after atmospheric pressure was reached with the subject breathing 100% O2 and, finally, after 30 min with the subject breathing air. At this point in time, the experiment was concluded. The PA catheter and the arterial catheters were withdrawn, and firm pressure was held for 10 min with suitable dressings applied. The subjects were encouraged to eat as needed and not to ascend in altitude for at least 24 h. The subjects also were encouraged to notify us if they had complaints referable to the study, especially if they had manifestations of decompression sickness. The subjects were evaluated the following day to ensure adequate hemostasis at the catheter sites and to discuss any worries or concerns.
The data collected by the continuous cardiac output catheter system and intra-arterial continuous O2 sensor are not presented in this report. Transcutaneous O2 and CO2 data collected at this time, compared with PaO2 and PaCO2, are available elsewhere (42).
For the Q̇s/Q̇t or venous admixture calculations (47) and for all arterial content of O2 calculations, we averaged all Hba values. Similarly, for the contents of O2 in PA blood, all Hbv̄ values were averaged. This averaging was done to minimize possible confounding errors from minor differences in hemoglobin concentrations from across conditions. We did not average hemoglobin when reporting contents of arterial or pulmonary arterial O2 elsewhere.
Data were analyzed using mixed-effects regression models where each subject was allowed to have a different random intercept, but the overall pattern was consistent between subjects. Each response was modeled separately using a guided stepwise procedure starting with the model that the response is predicted by gas and pressure (with a possibility of different slopes on pressure for air and O2). Additional models were fit to see whether simpler models fit or if a more complex model was needed to best fit the data (curved relation with pressure, individual time points needing an offset from the base model). All tests were done using α = 0.05 (2-sided). Descriptive data are expressed as means (SD).
Baseline characteristics of individual research subjects are shown in Supplemental Table 1. (Supplemental data for this article is available online at the Journal of Applied Physiology website.) One-half of the subjects were male. The subjects had no identified comorbid conditions or influences. All subjects completed the study protocol.
The baseline condition RQ was 0.87 (0.14). Average Hba across all experimental conditions was 14.2 (1.4) mg/dl. The average total blood loss was 185.3 (21.0) ml. The baseline Hba was 14.6 (1.4), and the final Hba was 14.2 (1.5). The baseline weight was 70.4 (13.2) kg, and the final weight was 69.8 (14.0) kg. The normal saline administered was 1,038 (352) ml. The liquid oral intake was 97 (77) ml. The urine output was 630 (299) ml. No complications were noted. The inside chamber temperature increased 3°C with chamber compression and then sloped downward to baseline as the chamber was decompressed (for chamber and subject temperature under experimental conditions, see Supplemental Fig. 1).
Results from this study are depicted in Figs. 2–9. Supplemental Table 2 reports mean change from baseline for each data point, with a 95% confidence interval and 25th and 75th percentiles.
Heart rate, respiratory rate, and blood pressure.
Results are depicted in Fig. 2. Mean HR decreased 3.19 beats/min for each 1 atm abs pressure increase during exposure to air (P < 0.001). Mean HR was 1.95 beats/min lower while subjects breathed O2 than during exposure to air (P < 0.001). Mean RR decreased 0.42 breaths/min for each 1 atm abs pressure increase while subjects were exposed to air (P = 0.045). Mean RR was 0.4 breaths/min lower on O2 than on air (P = 0.03). Average systolic BP was ∼3 Torr higher with O2 than air breathing (P = 0.009). Diastolic BP was not significantly different while subjects were breathing air compared with that while breathing O2 (P = 0.105). MBP for subjects breathing air was 88 Torr. MBP for subjects breathing O2 was marginally higher, 89 Torr.
Right atrial, pulmonary arterial, and balloon occlusion pressures.
Results are depicted in Fig. 3. RAP did not change across test conditions with the exception of data from two subjects with elevated RAP for test condition 11 (100% O2 at 1.12 atm abs).
Average systolic PAP was related to hyperbaric pressure, both on air and on O2 (P < 0.001). Systolic PAP dropped 1.52 Torr for each 1 atm abs pressure increase (P < 0.001). With the first O2 breathing interval at 0.85 atm abs (condition 7) and at the conclusion of the experiment (condition 13, breathing air at 0.85 atm abs), systolic PAP was 1.52 Torr lower for condition 7 and 2.15 Torr higher for condition 13 than predicted based on pressure measurements alone. Average diastolic PAP was 0.54 Torr lower per 1 atm abs pressure increase while subjects were exposed to air (P = 0.005). Average MPAP decreased 0.85 Torr per 1 atm abs increase in pressure (P < 0.001) and was 0.9 Torr lower during O2 exposure (P = 0.002). Average PAWP increased 0.67 Torr while subjects were breathing HBO2 (P = 0.005).
The thermal dilution Q̇ decreased from baseline 0.19 l/min per 1 atm abs increase in pressure (P = 0.002). Q̇ was 0.46 l/min lower while subjects were breathing O2 than while breathing air (P = 0.001). The CI dropped 0.11 l·min−1·m−2 for each 1 atm abs increase in pressure (P = 0.001). The CI was 0.25 l·min−1·m−2 lower while subjects were breathing O2 than while breathing air (P = 0.002) (Figs. 4 and 5).
Arterial blood gases.
Results are depicted in Fig. 6. pHa dropped during hyperbaric air and was higher than baseline during HBO2 exposure. PaCO2 was lower during HBO2 exposure compared with the initial baseline or hyperbaric air exposures. From the mixed-effects model, PaCO2 sloped upward as HBO2 pressure decreased (P < 0.001) (Fig. 5). The nadir for PaCO2 occurred at 100% O2 at 3 atm abs and increased as HBO2 pressure decreased. The PaO2 measurements were in the range predicted based on the subjects' inhaled partial O2 pressures.
Pulmonary arterial (mixed venous) blood gases.
Results are depicted in Fig. 7. pHv̄ was essentially constant across test conditions. Pv̄CO2 was lower during HBO2 exposure at 3.0 and 2.0 atm abs (P = 0.02). Pv̄O2 and Sv̄O2 were elevated during hyperbaric air and O2 exposures.
Calculations from arterial and pulmonary arterial measurements.
CaO2 and Cv̄O2 increased as the partial pressures of inhaled O2 increased, during both hyperbaric air and HBO2 exposures (Fig. 8). From the mixed-effects model, CaO2 − Cv̄O2 decreased 0.18 Torr with each 1 atm abs increase in air pressure (P = 0.006). CaO2 − Cv̄O2 increased 0.61 Torr for each 1 atm abs pressure increase with O2 breathing (P < 0.001). CaO2 − Cv̄O2 was higher during HBO2 breathing at 3.0, 2.5, and 2.0 atm abs (Fig. 8). The Q̇O2 delivery (Q̇ × PaO2) and V̇o2 [Q̇ × (CaO2 − Cv̄O2)] were relatively constant across test conditions (Fig. 8).
There was no significant change in SV across conditions (Fig. 9). From the mixed-effects model, SVR increased 58.7 dyn·s·cm−5 with each 1 atm abs increase in air pressure (P = 0.003). SVR was higher during HBO2 breathing at 2.5 and 2.0 atm abs compared with baseline (Fig. 9). From the mixed-effects model, PVR decreased 15 dyn·s·cm−5 with each 1 atm abs increase in pressure (P < 0.001) and was 39.4 dyn·s·cm−5 lower with HBO2 than air exposure (P < 0.001). Q̇s/Q̇t was 15% during all air breathing periods at 0.85 atm abs. During both hyperbaric air and HBO2 breathing, Q̇s/Q̇t decreased. This change was greater during HBO2 than hyperbaric air breathing (Fig. 9).
This is the only study that reports arterial and PA hemodynamic measurements with blood gases of healthy volunteers exposed to both hyperbaric air and HBO2. Limited observations in prior human studies measuring Q̇ suggest that HR and Q̇ decrease ∼20% during HBO2 exposure (21, 24, 27, 29, 48). In our subjects, we noted a decrease in HR with exposure to hyperbaric air and HBO2. The Q̇ value also dropped with HBO2 exposure, but the magnitude of change was <20% and could have been influenced by noise in thermal dilution measurements (34), despite making these measurements carefully. Although not using HBO2 in their experimental protocol, some studies noted a decrease in Q̇ on exposure to hyperoxia (1, 30). However, other studies found that Q̇ did not change with exposure to hyperoxia (2, 23). Only a few studies have measured cardiac output during HBO2 exposure (Table 1). In a prior report (24), investigators used flow-directed PA catheters in 11–12 healthy volunteers exposed to air at 0.56 atm abs, then air at 1 atm abs, and then 100% O2 at 3 atm abs and found that exposure to HBO2 resulted in decreases in Q̇, PAP, and PVR. The fall in HR and Q̇ during HBO2 exposure may be due to increased vagal stimulation, since atropine administration reverses the fall in HR and Q̇ with 100% O2 at 1 atm abs (14). However, in an animal study, dogs pretreated with propranolol, phentolamine, and atropine demonstrated decreased Q̇ during HBO2 exposure (35), presumably due to physiological autoregulation of the myocardium (35) and vasculature (5) due to increased PaO2.
Hyperoxic breathing at 1 atm abs may increase left ventricular end-diastolic pressure (LVEDP) (23). Explanations for this finding are reductions in nitric oxide myocardial relaxant factor attributable to increased PaO2 (28, 33). LVEDP was not measured in this study, but PAWP approximates LVEDP closely in normal individuals (11, 19). The PAWP values in this study were slightly lower than baseline with inhalation of 100% O2 at 0.85 atm abs, and LVEDP should have been similarly reduced from baseline under these conditions. During HBO2 exposure, the PAWP was higher than baseline or air breathing values, and therefore the LVEDP should have been higher. This finding supports other evidence that HBO2 may increase left ventricular wall stiffness (23, 43). Hyperoxia may reduce myocardial relaxant factor and increase left ventricular myocardial stiffness, resulting in increases in LVEDP, which could contribute to acute lung edema in heart failure patients treated with HBO2 (43). However, during HBO2 exposure, RAP and PAP did not increase compared with baseline values: their values were lower than baseline. It is possible that only the left ventricle increases filling pressures during HBO2 exposure. It also is possible that the hyperbaric air exposure before HBO2 exposure modified the RAP and PAP responses to HBO2 with the increased partial pressures of nitrogen modulating a subsequent HBO2 effect. It also is possible that the duration of O2 inhalation was insufficient to alter RAP or PAP. Differences in intracardiac pressure measurement results also may be accounted for by different dosing of O2 inhalation [Po2 ∼0.7 atm abs (23) compared with 0.85 to 3.0 atm abs in this study].
The pHa increased and PaCO2 decreased during HBO2 exposure, a confirmatory finding (22). A previous study reported hypocapnea during hyperbaric oxygen inhalation caused by a central accumulation of CO2, contributing to hyperventilation (22).
Whalen et al. (48) demonstrated that in 10 healthy volunteers breathing 100% O2 at 3.04 atm abs, PaCO2 did not change compared with baseline, whereas Pv̄CO2 increased 4 Torr from baseline. However, we observed that Pv̄CO2 was slightly lower during HBO2 exposure. Explanations these conflicting data include the following. 1) Whalen et al. did not measure true mixed venous blood, since the sampling catheter was located near the RA, not the PA. 2) A reduction in V̇o2 and CO2 production during HBO2 exposure could contribute to reduced Pv̄CO2. In our study, V̇o2 did not change during the experiment, so this explanation is not supported by the data. 3) If PaCO2 fell sufficiently during HBO2 exposure, Pv̄CO2 might not elevate as expected, since the arterial capillary values were lower. 4) There may be ventilatory effects such as hyperventilation (22) acting independently regarding the reduction seen in PaCO2. 5) It is possible the prior hyperbaric air exposure influenced the CO2 measurements during subsequent HBO2 exposure, although we do not have an explanation based on physiological principles. 6) The time course from the baseline value to those obtained during HBO2 exposure in our study was several hours, whereas was likely a shorter interval in research conducted by Whalen et al. 7) Finally, it is possible a difference in measurement technique of CO2 in 1965 compared with that used now has contributed to small differences in observed results.
As expected, PaO2, Pv̄O2, SaO2, and Sv̄O2 were elevated during exposure to hyperbaric air and HBO2, within predicted ranges (6, 8, 24, 41, 45). The SV was constant across the experimental conditions. This finding is expected, since the reduction in Q̇ seems to be proportional to the reduction in HR.
Although Q̇O2 and V̇o2, as determined using PaO2, Pv̄O2, and Q̇ by thermal dilution, did not change across conditions, CaO2 − Cv̄O2 increased during HBO2 exposure. If V̇o2 remains constant (as expected in a normal volunteer while at rest) and if Q̇ falls, CaO2 − Cv̄O2 must increase because V̇o2 is proportional to Q̇ × (CaO2 − Cv̄O2).
As expected, SVR increased during HBO2 exposure. Since SVR is calculated as (MPAP − RAP)/Q̇, if the pressures in the numerator remain constant and Q̇ falls, then SVR must increase. Since vascular tone was not actually measured in our study, we cannot determine whether the actual caliber of arterial blood vessels was reduced during HBO2 exposure.
PVR fell while subjects breathed increased partial pressures of O2, during both hyperbaric air and HBO2 exposures. It is well accepted that increased partial pressures of O2 reduce pulmonary pressures in patients with chronic pulmonary disease or in patients with hypoxic vasoconstriction, but normobaric hyperoxia has no significant effect on PAP (10). However, in our study, hyperbaric oxygen pressures were used. Our findings are in agreement with the limited data presented by McMahon et al. (24).
Q̇s/Q̇t was higher than expected while subjects breathed air at 0.85 atm abs. One explanation for this elevation is the fact that this research took place at 1,500 m above sea level, where the resting baseline PaO2 was 75–80 Torr. In one study, intubated patients with pulmonary problems breathing supplemental oxygen also had increased Q̇s/Q̇t (10), but we could find no other experiment in normal human subjects where Q̇s/Q̇t was measured by incorporating data from PA catheters at increased altitude. Other investigators have demonstrated that ventilation-perfusion (V̇A/Q̇) mismatch does not increase while breathing at increased altitude at rest (13, 18, 37). If V̇A/Q̇ does not change, it is reasonable to assume that Q̇s/Q̇t will not change. An increased Q̇t due to stress or exercise could increase Q̇s/Q̇t while breathing air at our altitude, but all subjects were at rest and simultaneous measurements of Q̇t were normal. Atelectasis could contribute to an increased Q̇s/Q̇t value. All subjects in this study were normal, and a chest radiograph taken only a short period before measurement of the initial Q̇s/Q̇t value was normal in all 10 subjects, without evidence of atelectasis. However, subjects were supine for up to 2 h before initial Q̇s/Q̇t measurement, so it is possible undiscovered atelectasis played a role in increasing Q̇s/Q̇t in these subjects.
During breathing of 100% O2 at 1.12 atm abs, a pressure slightly greater than sea level, Q̇s/Q̇t was 7%, a value that is within the expected range. With exposure to HBO2, Q̇s/Q̇t dropped to zero. It is reasonable to assume that this low Q̇s/Q̇t value is attributable to high O2 concentrations in mixed venous blood and in other circulations that contribute to venous admixture (e.g., Thebesian and bronchial circulation) such that venous admixture (physiological shunt) do not significantly contribute to increasing the shunt fraction.
In addition, intrapulmonary shunts may contribute to Q̇s/Q̇t (20). With high alveolar oxygen tensions during hyperbaric oxygen exposure, these intrapulmonary shunts may participate in gas exchange because of high oxygen diffusion gradients. In patients with abnormal lung function, PaO2 was greater than predicted during HBO2 compared with 1.0 atm abs measurements, suggesting their Q̇s/Q̇t fell during HBO2 exposure (44). Mathematical modeling demonstrates that Q̇s/Q̇t falls as the ratio of PaO2 to FiO2 (fraction of inspired O2) increases (31), such as during hyperbaric air and HBO2 breathing, which our data support. This is the first published report measuring Q̇s/Q̇t of humans exposed to HBO2, so these results cannot be compared with other work.
In prior research, we found subjects exposed to HBO2 had a wider A-aDo2 (45) compared with those in a prior study (7), similar to a finding in patients following open-heart surgery (32). These findings raise the question of whether the increased A-aDo2 might be due to a worsening Q̇s/Q̇t in association with HBO2. On the contrary, in this present study, we found Q̇s/Q̇t fell below baseline and did not increase.
Our study is limited due to the relatively small sample size of 10 subjects and individual variability. Other limitations include confounding covariables such as the stress of catheter insertion and variation in ambient chamber temperature. The contribution to hemodynamic variables of a change in chamber temperature up to 3°C is unknown but is probably of little significance.
Measurements of V̇o2 in this study may be limited by measurement of V̇o2 by PA catheter. Because of the technical challenges of measuring V̇o2 when a subject breathes 100% O2, compounded by the difficulty performing V̇o2 measurements inside a monoplace hyperbaric chamber, we were not able to overcome this limitation.
The apparent reduction in calculated Q̇s/Q̇t may not reflect an actual reduction in Q̇s/Q̇t because of limitation in the calculations when applied to HBO2 conditions. Inaccuracy can occur in this calculation when very high PaO2 and Pv̄O2 values are present, as well as by low effective solubility of oxygen during HBO2 (36).
It is possible the hyperbaric air exposure modified the subjects' hemodynamic response to 100% O2 inhalation, including HBO2. To maximize the information obtained from each subject, they were compressed with analogous hyperbaric profiles, differing only in the FiO2. Subjects underwent the hyperbaric air profile before the HBO2 profile to reduce the risk of decompression sickness. The study design could have been improved by randomly allocating subjects to receive either hyperbaric air or HBO2 initially, but we did not have a sufficient number of subjects to gain useful information by randomization. Hyperbaric air compression could have been omitted but would have limited the conclusions that could be drawn regarding the effect of hyperbaric pressure vs. the effect of HBO2.
This study examined the effect of hyperbaric air and HBO2 on arterial and PA hemodynamics and blood gas measures. The data presented are in agreement with others that HR and Q̇ fall with HBO2 exposure. The PaO2, Pv̄O2, Sv̄O2, and PaCO2 values were within expected ranges during hyperbaric air and HBO2 exposures. PVR and Pv̄CO2 decreased during HBO2 exposure. Finally, Q̇s/Q̇t was higher than expected with subjects breathing air at atmospheric pressure and fell to zero during HBO2 exposure.
This study was supported by Deseret Foundation Grant 268; LDS Hospital; Baxter Healthcare Corporation, Irvine, CA; Abbott Critical Care Systems, Chicago, IL; Biomedical Sensors, Malvern, PA; and Arrow International, Reading, PA.
We are very grateful to the subjects who volunteered for this study. We thank Jeff Noyes, CV monitoring, for assistance with the arterial catheter placements. The support and critique of study design of Dr. Terry Clemmer, Director of Critical Care, LDS Hospital, is appreciated. We also thank Gary Willoughbee, Baxter Healthcare, for equipment and technical advice. We also appreciate the assistance of Dr. David Kristo, Valerie Larson-Lohr, Shauna Hein, and Diane Haberstock.
Present address of S. Howe: Medical Informatics, LDS Hospital, Salt Lake City, UT 84143.
- Copyright © 2009 the American Physiological Society