Breathing 100% oxygen during water immersion improves postimmersion cardiovascular responses to orthostatic stress

Abstract Physiological compensation to postural stress is weakened after long‐duration water immersion (WI), thus predisposing individuals to orthostatic intolerance. This study was conducted to compare hemodynamic responses to postural stress following exposure to WI alone (Air WI), hyperbaric oxygen alone in a hyperbaric chamber (O2 HC), and WI combined with hyperbaric oxygen (O2 WI), all at a depth of 1.35 ATA, and to determine whether hyperbaric oxygen is protective of orthostatic tolerance. Thirty‐two healthy men underwent up to 15 min of 70° head‐up tilt (HUT) testing before and after a single 6‐h resting exposure to Air WI (N = 10), O2 HC (N = 12), or O2 WI (N = 10). Heart rate (HR), blood pressure (BP), cardiac output (Q), stroke volume (SV), forearm blood flow (FBF), and systemic and forearm vascular resistance (SVR and FVR) were measured. Although all subjects completed HUT before Air WI, three subjects reached presyncope after Air WI exposure at 10.4, 9.4, and 6.9 min. HUT time did not change after O2 WI or O2 HC exposures. Compared to preexposure responses, HR increased (+10 and +17%) and systolic BP (−13 and −8%), and SV (−16 and −23%) decreased during HUT after Air WI and O2 WI, respectively. In contrast, HR and SV did not change, and systolic (+5%) and diastolic BP (+10%) increased after O2 HC. Q decreased (−13 and −7%) and SVR increased (+12 and +20%) after O2 WI and O2 HC, respectively, whereas SVR decreased (−9%) after Air WI. Opposite patterns were evident following Air WI and O2 HC for FBF (−26 and +52%) and FVR (+28 and −30%). Therefore, breathing hyperbaric oxygen during WI may enhance post‐WI cardiovascular compensatory responses to orthostatic stress.


Introduction
Physiological compensation to postural stress is altered after long-duration water immersion (WI), thus predisposing individuals to blood pressure instability and orthostatic intolerance (Stegemann et al. 1975;Faes et al. 2013;Florian et al. 2013). During WI, central blood volume increases due to fluid shifts from the intracellular and interstitial fluid compartments to extracellular compartment (Epstein 1992) and redistribution of blood volume from the lower extremities to the thorax (Greenleaf et al. 1983;Greenleaf 1984;Pendergast et al. 2015). Consequently, there is a shift toward increased parasympathetic and decreased sympathetic tone (Mano et al. 1991;Tripathi 2011;Florian et al. 2013), together with augmented excretion of fluid and electrolytes (Norsk et al. 1985;Epstein 1992). After egress from the water, plasma volume is reduced by 10-15% Boussuges et al. 2009;Florian et al. 2013), and with postural stress, the resulting deterioration of baroreflex sensitivity/coupling  and an inability to augment systemic vascular resistance may contribute to post-WI orthostatic intolerance (Iwase et al. 2000;Florian et al. 2013).
Exposure to high partial pressures of inspired oxygen (hyperoxia) may, through direct and indirect effects on the vasculature and autonomic nervous system, augment post-WI cardiovascular compensatory responses to orthostatic stress. Acute exposure to hyperoxia increases systemic vascular resistance via a local (nonsympathetic) mechanism (Seals and Victor 1991;Lund et al. 1999) and decreases resting heart rate (HR), stroke volume (SV), cardiac output (Q), and regional blood flow (Torbati et al. 1979;Gole et al. 2011). High-frequency HR variability (HF HRV ) is augmented and the low-frequency/high-frequency ratio (LF/ HF HRV ) is diminished with increasing partial pressure of O 2 (Lund et al. 1999(Lund et al. , 2003Shibata et al. 2005). The change in cardiac regulation may be baroreflex-mediated (Shibata et al. 2005;Demchenko et al. 2013) or due to reductions in cardiac sympathetic activity through unloading of chemoreceptor tonic reflex drive (Fukuda et al. 1989).
Although several studies have shown that vascular and cardiac changes may persist for a short duration after hyperoxia (Torbati et al. 1979;Gole et al. 2011), the physiological effects and impact of long-duration hyperoxia on post-WI orthostatic responses have not been evaluated. Therefore, the purpose of this study was to compare hemodynamic responses to postural stress following WI alone (Air WI), hyperbaric oxygen alone in a hyperbaric chamber (O 2 HC), and WI combined with hyperbaric oxygen (O 2 WI). Previously published data ) from the Air WI phase are included in this manuscript for comparison purposes. WI was at the bottom a 15-ft pool, with the lung centroid at approximately 12 ft (1.35 atmospheres absolute [ATA]), and pressure was matched during O 2 HC. During supine rest and 15 min of 70°head-up tilt (HUT), we measured hemodynamic variables, spontaneous baroreflex sensitivity, and time-and frequency-domain measures of HRV. We hypothesized that cardiovascular stability during orthostasis would be diminished after a 6-h Air WI, but that breathing 100% O 2 during a 6-h WI would enhance post-WI orthostatic cardiovascular compensatory responses through blood volume protective mechanisms during WI and sustained vasoconstriction after WI.

Methods Subjects
Thirty-two healthy men (10 Air WI, 10 O 2 WI, and 12 O 2 HC), whose physical characteristics at screening are presented in Table 1, participated in the study. All participants were experienced military divers with an average of 9 years diving experience. They were healthy, active, normotensive nonsmokers who were not taking any medications that would affect responses in the study. Each subject underwent medical screening that included complete blood count, complete metabolic panel, lipid profile evaluation, urinalysis, physical examination, skinfold body fat measurement, and determination of maximal oxygen uptake ( _ VO 2 max ). Approval was obtained from the Institutional Review Board of the Navy Experimental Diving Unit. Each subject gave written informed consent, and all procedures conformed to the Declaration of Helsinki.

Study design
Subjects abstained from alcohol for 2 days, caffeine and strenuous exercise for 1 day, and food and drink (except water) for 2 h before reporting to the laboratory in the morning. Subjects wore running shorts and T-shirts for all visits. Each subject underwent physiological testing before and after a 6-h WI or HC trial. All physiological testing was completed in a laboratory (air temperature 22-24°C) adjacent to the immersion tank and HC. After completing pre-WI or -HC testing (see Protocol for preand posttrial testing), each subject received a standardized snack, submitted a urine sample for measurement of urine specific gravity, emptied his bladder, and was weighed. For WI trials, a condom catheter was applied to collect urine during the dive. Subsequently, for WI trials each subject was immersed in the tank, surfaced after the third hour for a 10-min lunch break while still immersed to midchest, and then returned to complete the WI. For HC trials, each subject entered the HC, was pressed to 1.35 ATA, and donned a hood supplying 100% O 2 (PO 2 = 1.35 atm). After the third hour, each subject removed his hood for a 10-min lunch break while  . Each subject was then tilted 70°headup from supine for 15 min or until symptoms associated with presyncope occurred or the subject requested termination of the test. Presyncope was defined as a rapid decrease in a systolic pressure to below 80 mmHg or a sustained systolic pressure below 90 mmHg associated with symptoms of light-headedness, nausea, or diaphoresis. Subjects were tilted back down to the horizontal position at the end of 15 min or, if presyncope occurred, to the Trendelenburg position (À10°) until hemodynamic stability was reached. Ten minutes of recovery were recorded in the supine position. Tilt time was limited to 15 min due to schedule constraints. Only data segments from periods of hemodynamic stability (i.e., excluding presyncope and transients just after HUT) were analyzed.

Water immersion
Twenty participants underwent 6-h WI at the bottom of a 15-ft pool filled with thermoneutral water (32-33°C). As noted previously, thermoneutral conditions varied between the individuals (Conaty et al. 2015;Pendergast et al. 2015).
Since the water temperature provided in this study was optimized for subject comfort and is close to the generally accepted "thermoneutral" range, it will be referred to as "thermoneutral." Ten of these subjects breathed humidified 100% O 2 for an oxygen partial pressure (PO 2 ) of approximately 1.35 atm (135 kPa), and the other 10 breathed humidified air for a PO 2 of approximately 0.3 atm (30 kPa). They wore T-shirts and shorts, and weights were provided to maintain negative buoyancy. While sitting upright in a chair, thus with the chest at a depth of approximately 12 ft, each participant breathed surfaced-supplied O 2 or air delivered with a MK20 breathing apparatus (Aga mask; Interspiro Inc., Pleasant Prairie, WI) . After 3 h of WI, each subject returned to the surface to stand on a platform with head and shoulders out of the water for 10 min while consuming a small lunch consisting of 2.2 MJ energy content (24% fat, 64% carbohydrate, 12% protein) and 500 mL total of a meal replacement and a sports drink.

Hyperbaric chamber
Ten participants underwent a 6-h HC trial in one of the dry chambers of the Navy Experimental Diving Unit Ocean Simulation Facility. Subjects wearing T-shirts and shorts sat upright on a bench in the HC and were pressed to an equivalent pressure to the WI, that is, 1.35 ATA.
Once the HC reached the appropriate depth, subjects donned oxygen treatment hoods (Amron Model 8891; Amron International, Vista, CA) and started their 6 h of breathing humidified 100% oxygen. After 3 h, subjects removed their hoods for 10 min while at depth and ate a lunch identical to that provided during WI exposures. Absolute pressure was kept constant for all phases so that the effects of WI and differences between hyperoxia and approximate normoxia (that is, with a slightly elevated PO 2 ) on physiological function could be determined.

Hemodynamic measurements
Resting brachial BP was measured oscillometrically (HEM 007XL; Omron Healthcare, Inc., Lake Forest, IL). Continuous BP was measured by photoplethysmography (Finometer) on a finger of the right hand. Finger pressure was calibrated to brachial artery pressure using the manufacturer's return-to-flow system. A 5-lead surface electrocardiogram (Dash 3000; General Electric Company, Fairfield, CT) was used to determine HR. Inc., Malvern, PA). Data were recorded at baseline, every 5 min during HUT, and during recovery. During each data recording period, blood flow was acquired from three to four measurement cycles in succession. Forearm vascular resistance (FVR) was estimated as corresponding brachial MAP/FBF.

Cardiac output
_ Q was assessed using transthoracic echocardiography (Acuson Cypress; Siemens Medical Solutions USA Inc., Malvern, PA). Stroke volume (SV) was determined from the flow velocity across the aortic valve (apical approach) and the diameter of the aortic orifice during systole (parasternal long axis). _ Q was calculated as SVÁHR and expressed in L/min. Systemic vascular resistance (SVR) was calculated as MAP/ _ Q. Orthostatic tolerance was estimated by the maximum increase in HR (+DHR max ) during HUT and by the orthostatic index (Stegemann et al. 1975;Florian et al. 2013) calculated from the change in HR and BP during HUT.

Time-domain HRV and BPV
Mean HR, root-mean square of successive differences (RMSSD) of RR intervals, and the standard deviation of normal-to-normal R waves (SDNN) were calculated. RMSSD mainly reflects the modulation of the parasympathetic system and SDNN is an indicator of overall ANS activity. BPV was assessed using the standard deviation of mean SBP (SBP-SD) (1996).

Approximate entropy (ApEn)
ApEn, a nonlinear statistical method used to assess the complexity of data, has been used to measure the loss of complex nonlinear HRV in a variety of pathological and physiological conditions (Pincus and Goldberger 1994;Florian et al. 2013). ApEn values were calculated from consecutive RR intervals using embedding dimension m = 2 and automatically selected threshold value r (Chon et al. 2009).

Frequency-domain analyses
A time-domain HRV signal was generated from the instantaneous RR interval series at a uniform sampling rate of 4 Hz using cubic spline interpolation. The HRV signal was down-sampled to 2 Hz, mean and linear trends were removed, and the signal was transformed into the frequency domain. For each subject, HRV time-domain signal segments containing 360 points (3 min) were analyzed with both power spectral density (PSD) (1996) and principal dynamic modes (PDM) (Zhong et al. 2004) methods.

Power spectral density analysis of HRV
Power spectral densitys of HRV data were obtained using the Welch periodogram method (Matlab â 7.9, Natick, MA). A 512 point Fast Fourier Transform (FFT), giving a frequency resolution of 0.004 Hz, was applied to data filtered with a 360-point Hamming window and no overlapping segments. Mean spectral power in the low-(0.04-0.15 Hz) and high-frequency (0.15-0.4 Hz) bands and the ratio between them were calculated (LF HRV , HF HRV , and LF/HF HRV , respectively). For LF HRV and HF HRV , power was also represented normalized by total power (LFn HRV and HFn HRV ).

Principal dynamic modes (PDM) analysis of HRV
Principal dynamic modes analysis was used in addition to PSD to assess sympathetic and parasympathetic dynamics during HUT. Unlike PSD, PDM accounts for the inherent nonlinear dynamics of HR control. The two most dominant PDMs are considered to represent sympathetic (SNS) and parasympathetic (PNS) nervous activity. Methodological details have been described previously (Zhong et al. 2004). In this study, the optimal estimation error was found with nine Laguerre functions and a memory length of 60.

Baroreflex sensitivity
The complex-valued transfer function between RR and SBP was evaluated as the ratio of the cross-spectral density function of the two series and the PSD of the SBP series. The BRS gain (transfer function modulus) was determined by averaging the gain in the whole LF band (GainLF) regardless of the value of coherence (Pinna et al. 2002) within the LF.

Blood samples
Glucose, Hb, and Hct levels were determined immediately after blood collection (Rapidpoint 400; Siemens Medical Solutions USA Inc). Blood for all other analyses was centrifuged at 4°C and stored at À80°C until assay. Samples for norepinephrine, ANP, AVP, and 8-isoprostane were drawn into prechilled tubes containing EDTA. Blood for aldosterone was allowed to clot at room temperature for 30 min before centrifugation. Glucose was measured by the oxidase method; norepinephrine by HPLC; and aldosterone, ANP, AVP, and 8-isoprostane by immunoassay. For calculation of PV, blood was drawn in a 2-mL sodium heparin tube for measurements of Hb and Hct using Rapidpoint 400 (Siemens Medical Solutions USA Inc.). The relative change in PV (DPV) following WI was calculated from changes in Hb and Hct concentrations according to the Harrison modification of the Dill and Costill equation (Harrison et al. 1982).

Statistical analysis
A mixed model repeated measures analysis of variance (ANOVA) was used to determine the effect of WI and hyperbaric oxygen on neural, hormonal, and hemodynamic variables. The within-subject factors were pre/post and time, and the between-subject factor was exposure, or phase. When appropriate, differences between factors were identified using the Bonferroni-Holm correction. All statistical analyses were performed using SAS 9.2 (SAS Institute, Inc., Cary, NC). The level of significance was set at a = 0.05, and values are presented as mean AE SEM. When significant, P-values for main effects and interactions are noted on the figures. Table 2 shows mean weight loss after each exposure, adjusted for food and fluid intake during the exposure, as well as at 0-3 h and at 3-6 h, total urine production, and and DPV after each exposure. Weight loss, DPV, and 0-3 h and total urine production were greater for Air and O 2 WI compared with O 2 HC (P < 0.05).

Supine resting measurements
Supine pre-and postexposure hormone and electrolyte concentrations, as well as resting hemodynamic data, are displayed in Table 3. Due to technical issues, no AVP data are available for O 2 HC exposures. No changes were noted for urine specific gravity, ANP, AVP, 8-isoprostane, glucose, HR, and FVR for any of the phases. Significant phase 9 pre/postinteractions were identified for weight, NE, Hb, Hct, SBP, DBP, _ Q, SV, and FBF. Specifically, in accord with the weight loss and decrease in PV following WI, Hb and Hct concentrations were also increased following WI only. SBP and FBF decreased after Air WI, but did not change after O 2 WI, suggesting a compensatory response of hyperoxia to these postimmersion supine variations. Breathing O 2 during WI also caused a decrease in _ Q and SV and an increase in NE and SVR that was not observed after Air WI. In contrast, DBP and FBF showed a significant increase after the O 2 HC exposure that was not observed after O 2 WI.

HUT testing
Divergent orthostatic responses (Tables 4, 5, Figs 1-4) were evident for the three different exposures. Although all subjects completed HUT before Air WI, three subjects reached presyncope after Air WI exposure at 10.4, 9.4, and 6.9 min. One subject became presyncopal at 14.3 min before O 2 WI and at 14.9 min after O 2 WI exposure. One subject reached presyncope at 11 min before O 2 HC but completed the full 15 min after O 2 HC. Both ÀDHUT SBP max and ÀDHUT DBP max significantly increased after Air WI, and ÀD HUT SBP max but not ÀD HUT DBP max increased after O 2 WI. Significant phase 9 pre/postinteractions were present for all orthostatic tolerance variables, including HUT time.
Hemodynamic and autonomic measurements before, during, and after HUT for all phases are presented in Tables 4, 5 and Figures 1-4. Table 5 shows pre-topostdive directional changes in the effects of HUT and % values when significant. A ↑ or ↓ is shown for Pvalue <0.05, and ↑↑ or ↓↓ is shown for P-value <0.01.   Statistical comparisons among the postexposure HUT changes are shown.

Hemodynamic measurements
Water immersion caused changes in the HUT values of HR, SBP, PP, and SV, changes that were not compensated by breathing O 2 . HR during HUT was greater after than before exposure (+10 and +17%), while SBP (À13 and À8%), PP (À10 and À14%), and SV (À16 and À23%) during HUT were lower after exposure to Air WI and O 2 WI, respectively. These changes seem related only to the WI; HR, PP, and SV during HUT were not different after O 2 HC, while SBP and DBP during HUT were greater (+5 and +10%) after than before O 2 HC. On the other hand, hyperoxia induced different responses in the HUT values of _ Q and SVR. _ Q during HUT was unchanged from the preimmersion value after Air WI but was lower after than before exposures to both O 2 WI (À13%) and O 2 HC (À7%). Similarly, SVR during HUT was lower after than before exposure to Air WI (À9%) but higher after than before exposures to both O 2 WI (+12%) and O 2 HC (+20%). DBP during HUT was lower (À6%) after Air WI than preexposure, was unchanged relative to preexposure after O 2 WI, and was greater (+10%) after O 2 HC than before the exposure. Different patterns were also evident for the HUT values of FBF, which was lower (À26%) after than before Air WI, was unchanged relative to preimmersion after O 2 WI, and was greater (+52%) after than before O 2 HC, and of FVR, which, in an inverse pattern to that seen for FBF, was greater (+28%) after than before Air WI, was unchanged relative to preexposure after O 2 WI, and was lower (À30%) after than before O 2 HC. In Figures 1 and 2, a general pattern can be seen: Air WI and O 2 HC are the two extremes of preto-post D values, and O 2 WI falls between them.
When the pre-to-postexposure D values during HUT were compared for all exposures (phases), the main effect of phase was significant for DBP, SV, and SVR, and both the main effect of phase and the phase 9 time interaction were significant for HR, SBP, and FBF. Overall, the orthostatic variables and hemodynamic responses to HUT suggest an overall reduction in orthostatic tolerance after Air WI, a blunted reduction in orthostatic tolerance after O 2 WI, and improved orthostatic tolerance after O 2 HC.
Heart rate variability and approximate entropy HUT LFn HRV and HFn HRV increased and decreased, respectively, following WI, but did not change after O 2 HC. However, the pre-to postexposure D was not significantly different among phases. LF/HF increased only after Air WI, and RMSSD decreased only after O 2 WI. Compared to pre-WI exposures, SNS increased after the both WI exposures, but not after O 2 HC, and PNS and SNS/ PNS were not different after any exposure. The expected decrease in ApEn during HUT was larger after than before WI exposures, but the pre-to-post ApEn D during HUT did not change after O 2 HC. The differing responses between phases (i.e., WI vs. HC) was also evident as a significant phase 9 time interaction (P < 0.03).

Baroreflex sensitivity and blood pressure variability
As expected, preexposure GainLF for all phases decreased during the beginning of HUT, plateaued through the rest of HUT, and returned to baseline values during recovery. Postexposure GainLF for Air WI and O 2 WI decreased further than it had preexposure, but the change after exposure was not significantly different among phases. Compared to preexposure values, postexposure SBP-SD increased significantly during HUT after Air WI and O 2 WI, but did not change following O 2 HC.

Discussion
This study demonstrates divergent effects of WI, breathing 100% oxygen in a HC, and the combination of WI and hyperoxia on hemodynamic responses and HRV during supine rest and 70°HUT. The novel finding is that breathing hyperbaric oxygen during a 6-h WI improves postimmersion cardiovascular compensatory responses to orthostatic stress. This protection persisted for at least 1.5 h after egress from the water and was associated with  -minus-preexposure). HR, heart rate; SBP, systolic blood pressure; DBP, diastolic blood pressure; WI, water immersion; HC, hyperbaric chamber. Values are group means AE SEM. Statistical comparisons between phases are noted on each plot. * denotes differences between phases. See Table 5  higher SBP, DBP, and SVR at rest and during HUT compared with responses after Air WI. The change in hemodynamic responses after hyperbaric oxygen exposure was not accompanied by any changes in HRV or HR complexity measures. Lastly, after O 2 HC exposure, although SVR increased during HUT, FVR decreased, indicating that constriction was not a global phenomenon even though most vascular beds must have been constricted to produce the whole-body increase in SVR. Although the underlying mechanisms cannot be determined from the current study, these results suggest a robust and persistent vasoconstriction related to hyperbaric oxygen.

Baseline responses
As expected and in agreement with previous studies (Mourot et al. 2004;Boussuges et al. 2009;Florian et al. 2013), weight decreased and PV was diminished by 9-11% after WI exposures but not after O 2 HC exposure. In concord, 3-h and total urine production were higher after WI exposures than after O 2 HC exposure. In agreement with previous studies (Epstein 1992;Stadeager et al. 1992;Florian et al. 2013), aldosterone, ANP, and AVP were unchanged 1 h after immersion, independent of breathing gas. However, during supine rest NE was Values are group means AE SEM. Statistical comparisons between phases are noted on each plot. * denotes differences between phases. See Table 5 for pre/post directional responses and statistics. elevated after O 2 WI but not after Air WI or O 2 HC, suggesting that the combination of hyperoxic exposure and direct or indirect effects of WI are required to mount a sympathetic response. In contrast to the reductions in SBP and FBF and increase in CVR after Air WI, after O 2 WI, SBP and FBF were preserved and SVR significantly increased. Interestingly, after O 2 HC, FBF was increased even with an increase in DBP and trend toward increased SVR and reduced FVR, which suggests that other nonmuscular vascular beds may be constricted postexposure to compensate for muscle vasodilation. Similar to information previously reported for Air WI , subjects after exposures to O 2 WI and O 2 HC did not experience a change in resting supine HR and HRV. In spite of the noted vascular effects, it is possible that cardiac autonomic compensation is not necessary for supine individuals with mild hypovolemia (Kimmerly and Shoemaker 2002) or with residual vascular effects of hyperoxia as seen in this study.

Head-up tilt responses and WI
Individuals who maintained adequate arterial pressure during an orthostatic challenge typically mount an appropriate cardiac and arterial vasoconstrictor response to the gravitational blood translocation (Blomqvist 1986). However, post-WI hypovolemia and altered autonomic regulation predispose individuals to orthostatic intolerance. Indeed, orthostatic intolerance has been repeatedly observed after long-duration WI (Stegemann et al. 1975;Faes et al. 2013;Florian et al. 2013) and 3 days of dry immersion (Iwase et al. 2000). Results from this study  and others Florian et al. 2013) suggest that in addition to diminished blood volume, changes in cardiac function, reduced baroreflex sensitivity and coupling, and decreased vascular responsiveness, are responsible for the reduced orthostatic tolerance. However, the clear differences among phases in this study show divergent physiological effects of WI and hyperoxia on cardiovascular function, and in particular, a possible protective effect of hyperoxia on orthostatic tolerance, as evidenced by HUT time and hemodynamic responses during HUT.

Oxygen effects
Normobaric and hyperbaric hyperoxia increased vascular resistance and alter cardiac and baroreflex regulation in patients, healthy participants, and animals. The increased vascular constriction has been shown to be nonsympathetically mediated (Seals and Victor 1991) but is induced rather by changes in reactive O 2 species levels (Jackson 1987), endothelin-1 (Dallinger et al. 2000), prostaglandins (Rousseau et al. 2010), or endothelium-derived vasoactive products (Pasgaard et al. 2007). It is unclear to what extent the constriction is attributed to endotheliumdependent vasoconstriction versus an attenuated endothelium-derived vasorelaxation, but recent evidence suggests that reduced release and availability of nitric oxide (NO) plays a major role in the hyperoxic vasoconstriction (Pasgaard et al. 2007). Hyperoxia also reduces _ Q via baroreflex-mediated vagal activation and diminished cardiac contractility (Demchenko et al. 2013). There is a linear  increase in arterial-cardiac baroreflex function as the fraction of inspired O 2 increases to 100% under normobaric conditions (Shibata et al. 2005), but the cardiac baroreflex slope has also been shown to decrease during hyperbaric hyperoxia (Gole et al. 2011). Nevertheless, the hyperoxic bradycardia is regulated by an increase and decrease in cardiac parasympathetic and sympathetic activity, respectively, as evidenced by changes in HRV (Lund et al. 1999(Lund et al. , 2000(Lund et al. , 2003Shibata et al. 2005;Gole et al. 2011;Florian et al. 2013).
Although baseline and HUT testing in the current study was completed approximately 1 h postexposure, many of the aforementioned hyperoxic responses were still present. Compared to preexposure HUT testing, post-O 2 HC SBP, DBP, and SVR remained elevated and _ Q was depressed throughout HUT testing. The hyperoxic constrictive effects also carried over to the post-O 2 WI HUT testing where, in contrast to the reduction in SVR and larger D tilt DBP max after Air WI, SVR increased and D tilt DBP max did not change. _ Q was also reduced across post-O 2 WI HUT testing.
Several studies have shown that cardiovascular adjustments to hyperoxia can persist after return to normoxia. For example, _ Q and HR did not return to baseline levels after 55 min of recovery following intermittent exposure to 100% O 2 at 2.2 ATA (Pelaia et al. 1992). The reductions in _ Q, SV, BRS, and LF BPV , and an increase in SVR were still evident 10 min after a 45-min exposure to 100% O 2 at 1 ATA, but returned to baseline levels by 30 min postexposure. The cardiac parasympathetic shift during the same hyperoxic exposure was more evanescent than the vascular responses, with a rapid postexposure shift to sympathetic dominance that lasted at least 30 min after return to normoxia (Gole et al. 2011). In the current study, no HRV or complexity measures differed from predive when they were measured 1 h postexposure after O 2 HC, suggesting full or nearly full recovery from dry hyperoxia in that time. In contrast, 1 h after either immersed condition, LFn HRV and SNS were increased and HFn HRV decreased relative to baseline, and SBP-SD was also higher, reflecting a persistent sympathetic dominance post-WI independent of hyperoxia. The reduction in GainLF after WI likely reflects a deterioration of BRS. Together, these changes point toward increased cardiac sympathetic activation up to 1.5 h after WI independent of breathing gas, but hyperoxia may still override some of the changes from WI such as reduced SVR and BRS.
Interestingly, during HUT after O 2 HC despite an increase in SVR, FBF was augmented and FVR diminished. This response contrasts with the reduced FBF and increased FVR after Air WI, and no change in either variable after O 2 WI. The observation that after O 2 HC BP and SVR were elevated while _ Q was slightly reduced and FBF was increased suggests a redistribution of blood flow and vasoconstriction in vascular beds other than muscle such as the splanchnic region. In animals, during shortterm hyperoxia, blood flow to the mesenteric circulation decreases, as, to a much lesser extent, does blood flow to muscle vasculature in the limbs (Bergofsky and Bertun 1966;Torbati et al. 1979). However, to our knowledge there are no data from which to extrapolate vascular recovery responses to the current posthyperoxic recovery data, nor are there any related experiments that have provided hyperbaric oxygen for as long a time. It is possible that the increase in FBF after O 2 HC is a compensatory response similar to the posthyperoxia HRV shift (Gole et al. 2011) or rebound to the hyperoxic constriction.
What is clear is that nonmuscle (or nonforearm muscle) constriction persisted long enough after cessation of hyperoxia to increase post-O 2 HC BP during HUT and to maintain BP during HUT after O 2 WI.

Experimental considerations
Several factors should be considered when interpreting the results of this study. Supine and HUT measurements were taken approximately 1-1.5 h postexposure, respectively, so the timeframe for manifestation of residual/compensatory physiological factors before that time cannot be determined. Similarly, since all exposures in this study were for 6 h, the duration or dose of oxygen required to provide protective effects cannot be determined from the current study. For comparison to previous studies, it is important to note that there may be differences between head-in WI and head-out WI (Pendergast et al. 2015). However, the seated head-in WI with a demand regulator at the mouth used in our experiments provides the same pulmonary hydrostatic loading as does seated head-out WI. Although most studies conduct HUT tests to failure, due to diving operation constraints, HUT was limited to a maximum tilt time of 15 min. Therefore, in addition to analyzing HUT time, other measures of orthostatic tolerance and hemodynamic stability such as the orthostatic index and changes in BP and HR during HUT were analyzed. An additional consideration is the use of echocardiography to measure Q. This indirect method provides reliable relative changes but may underestimate absolute values (Kiowski et al. 1981). Because HRV is an indirect measure of cardiac sympathetic and parasympathetic activity, it may not always accurately indicate sympathetic activation of the heart or be generalizable to other regions. However, LFn HRV and LF BPV responses during HUT have been shown to parallel changes in peroneal MSNA (Furlan et al. 2000). Lastly, depending on the PO 2 and duration of exposure, hyperoxia may cause central nervous system (CNS) oxygen toxicity and pulmonary oxygen toxicity. However, military and technical divers commonly use the PO 2 employed in this study because it is low enough that CNS oxygen toxicity is not a concern. Indeed, none of the subjects in the current study experienced symptoms of CNS oxygen toxicity, and we measured a generally mild pulmonary oxygen toxicity effect which is reported elsewhere.

Conclusion
This study demonstrates the divergent effects that WI, breathing 100% oxygen in a HC, and the combination of WI and hyperoxia have on hemodynamic responses and HRV during supine rest and 70°HUT. Breathing hyperbaric oxygen during a 6-h WI augments postimmersion cardiovascular compensatory responses to orthostatic stress. The primary protective effect appears to emanate from increased constriction in nonmuscle vasculature. Future mechanistic studies are required to determine both the timeline for hyperoxia residual/compensatory effects and which vascular beds are responsible for maintaining and even increasing SVR.