Submersion and increased pressure (depth) characterize the diving environment and may independently increase demand on the respiratory system. To quantify changes in respiratory mechanics, this study employed a unique protocol and techniques to measure, in a hyperbaric chamber, inspiratory and expiratory alveolar pressures (interrupter technique), inspiratory and expiratory resistance in the airways (RawI and RawE, esophageal balloon technique), nitric oxide elimination (thought to correlate with Raw), inspiratory and expiratory mechanical power of breathing, and the total energy cost of ventilation. Eight healthy adult men underwent experiments at 1, 2.7, and 4.6 atmospheres absolute (ATA) in dry and fully submersed conditions. Subjects rested, cycled on an ergometer at 100 W, and rested while voluntarily matching their ventilation to their own exercise hyperpnea (isocapnic simulated exercise ventilation). During isocapnic simulated exercise ventilation, increased O2 uptake (above rest values) resulted from increased expired ventilation. RawI decreased with submersion (mean 43% during rest and 20% during exercise) but increased from 1 to 4.6 ATA (19% during rest and 75% during exercise), as did RawE (53% decrease with submersion during rest and 10% during exercise; 9% increase from 1 to 4.6 ATA during rest and 66% during exercise). Nitric oxide elimination did not correlate with Raw. Depth increased inspiratory mechanical power of breathing during rest (40%) and exercise (20%). Expiratory mechanical power of breathing was largely unchanged. These results suggest that the diving environment affects ventilatory mechanics primarily by increasing Raw, secondary to increased gas density. This necessitates increased alveolar pressure and increases the work and energy cost of breathing as the diver descends. These findings can inform physician assessment of diver fitness and the pulmonary risks of hyperbaric O2 therapy.
- alveolar pressure
- airway resistance
- nitric oxide
the diving environment demands more of the ventilatory system than the terrestrial environment because static lung loading (SLL) is often imposed during submersion, and breathing gas density increases as a function of depth. These factors increase the work of breathing and may limit exercise performance in healthy divers or present unexpected risks in patients undergoing hyperbaric O2 therapy (HBOT).
Many studies have assessed the effects of the diving environment on the work of breathing and outlined changes imposed on ventilation by diving. SLL is imposed when the source of breathing gas is at a hydrostatic pressure different from that of the chest centroid (54). Negative SLL is typical and is encountered by divers in a head-up orientation (36), when the breathing gas is at less pressure than the chest. Negative SLL places greater demand on the inspiratory muscles. In addition, when the body is immersed, blood that normally pools in the extremities translocates to the thorax (17), decreasing total lung capacity, residual volume (17, 41, 53), and lung compliance (17). These changes increase the elastic work of breathing. Additional blood in the thorax may confer a small benefit, however, by increasing the ability to generate pressure in the airways. Since blood is essentially incompressible, less work may be lost to compression during maximal expiratory pressure (PEmax) maneuvers. The effects of submersion on respiratory muscle efficiency, however, remain unknown.
Increased depth affects respiratory mechanics in several ways. Work of breathing increases as a function of depth, primarily due to increased airway resistance (Raw) secondary to increased gas density (38). Historically, Raw was measured using the pressure differential between the mouth and an air-filled balloon placed within the lower third of the esophagus (46). Recently, clinical practice has shifted to measuring Raw by modifying the “P0.1 technique,” which was first developed by Otis and Proctor (45). Use of this technique involves brief interruption of airflow, and pressure in the system is taken to approximate alveolar pressure (Pa). This indicates the degree of respiratory muscle activation (26) and central nervous system respiratory drive (60). Raw can be calculated by dividing the pressure differential between the alveoli and the mouth by the flow rate immediately prior to occlusion. Raw values at sea level assessed by the P0.1 technique are comparable to those obtained by the traditional esophageal balloon technique (30, 40), and the P0.1 technique has been used to assess Raw during simulated dives at depths up to 45 atmospheres absolute (ATA) (12). To our knowledge, however, the P0.1 and esophageal balloon techniques have not been compared during wet simulated dives.
Another recent development in pulmonary mechanics is a more thorough understanding of the role of nitric oxide (NO), a free radical that induces smooth muscle relaxation. Raw may be reduced by NO production (V̇no) (2), and animal studies have shown that production or administration of NO decreases Raw in a dose-dependent manner (21, 27). However, the effects of diving on V̇no in human divers have not been reported, and the relationship between Raw and V̇no has not been examined. It is particularly important to study the effect of depth on V̇no, because there is evidence that V̇no is associated with maintenance of tissue normoxia (6), and it has been observed that V̇no in the cerebral vasculature decreases with depth as Po2 rises (20). If V̇no is similarly regulated in airway smooth muscle, the decrease in V̇no could result in bronchoconstriction and could be expected to exacerbate the increase in Raw with increasing depth. To maintain alveolar ventilation, it is likely that increases in Raw demand concurrent increases in Pa. Thus it is important to measure the effect of depth on V̇no to form a more complete understanding of how respiratory mechanics change in the diving environment.
Interestingly, depth may provide some benefit to respiratory muscle function. Because gas in the lungs at depth is compressed relative to gas at sea level, less work may be lost to compression of the gas during PEmax maneuvers, and higher pressures may be generated.
Many early studies sought to maximize the depth at which humans could operate, but more recent studies have attempted to improve diver performance at more moderate depths. To do so, it is important to clearly define the effects of the diving environment on respiratory mechanics.
The purpose of this study was to describe the effects of submersion and increased gas density (depth) on respiratory mechanics, with a particular emphasis on correlating newer methods (e.g., P0.1 measurement and V̇no calculation) with established measures of pulmonary mechanics during diving. These findings will create a more complete picture of the effects of submersion and depth on ventilation, which is important not only for improving diver performance, but also for making a more thorough medical assessment of fitness to dive and determination of whether HBOT is appropriate for patients with respiratory systems compromised by disease or injury.
We hypothesized that submersion would increase PEmax while increased pressure would 1) decrease V̇no, 2) increase Raw because of the increase in gas density and the decrease in V̇no, 3) increase Pa and, thereby, increase the respiratory work rate, 4) increase the O2 cost of breathing, and 5) increase PEmax.
- Atmospheres absolute
- Forced expiratory volume in 1 s
- Fraction of CO2 in inspired gas
- Forced vital capacity
- Hyperbaric O2 therapy
- Isocapnic simulated exercise ventilation
- Maximum voluntary ventilation in 15 s (l/min)
- Nitric oxide
- NO analyzer
- Alveolar pressure
- Expiratory Pa
- Inspiratory Pa
- Maximal expiratory pressure at the mouth
- Pleural pressure measured by esophageal balloon
- Pressure at the mouth
- Cardiac output
- Resistance in the airways
- Expiratory Raw measured at peak flow using the esophageal balloon technique
- Inspiratory Raw measured at peak flow using the esophageal balloon technique
- Raw measured by the P0.1 technique
- Expiratory Rint
- Inspiratory Rint
- Slow vital capacity
- CO2 production
- Vital capacity
- Minute ventilation
- NO elimination
- O2 consumption
- O2 cost of breathing
- Peak V̇o2
- Respiratory work rate (power)
- Expiratory Ẇ (power)
- Inspiratory Ẇ (power)
The study protocol was approved by the University's Health Sciences Institutional Review Board, and the volunteer subjects provided written informed consent before participating in the study.
Eight male certified SCUBA divers were recruited from the local diving population. At the start of the study, they were 28 ± 8 (SD) yr old with an average weight of 81 ± 12 kg and height of 179 ± 4 cm. A physical examination, including a chest X-ray and pulmonary function tests, was performed (Table 1). A medical history was obtained to exclude current or previous illness of consequence. Volunteers with abnormal pulmonary function tests were excluded from the study. All subjects were nonsmokers, but subject 10 had a history of smoking (1 pack/day for 9 yr) ending 2 yr prior to enrollment in the study and subsequently chewed tobacco.
Six experiments were performed in each subject: in dry and submersed conditions, at three simulated depths (1, 2.7, and 4.6 ATA). The order of the experiments was randomized, with the exception that each subject's first experiment was performed in the dry condition to allow familiarization with the protocol and techniques. Prior to the start of each experiment, each subject was outfitted with a three-lead ECG and a thin balloon-tipped catheter [100-mm balloon (Nolato, Torekov, Sweden) attached to 0.034-in. ID polyethylene tubing]. The balloon was passed through a nostril until it sat within the lower third of the esophagus (∼42 cm from the nostril to the tip of the balloon) and was inflated with 0.5 cm3 of air upon reaching the designated pressure. Proper placement was verified by observation of increasingly negative pressure during a deep, slow inspiration.
During experiments, subjects sat on a cycle ergometer behind a Lanphier-Morin barrier system (54) in the hyperbaric chamber (Fig. 1). During submersion, water was 30.9 ± 0.1°C, which is critical water temperature for resting subjects (16). Thus, subjects likely experienced near-maximal vasoconstriction of the skin and subcutaneous tissue without increases in metabolic rate while at rest but were thermoneutral during exercise. The water level was adjusted at the start of the experiment to impose a −15-cmH2O SLL.
Each experiment consisted of four 10-min periods of activity: rest, exercise, recovery, and isocapnic simulated exercise ventilation (ISEV). During exercise, the subject pedaled the cycle ergometer at 60 ± 5 rpm, against a work rate of 100 W [O2 consumption (V̇o2) = 1.80 ± 0.39 l/min, i.e., 59% of subjects' peak V̇o2 (V̇o2peak) determined at 1.0 ATA]. The ergometer controller was set to 25 W less when the subject pedaled in water to account for the added hydrodynamic drag of pedaling in water (54). During ISEV, the subject watched a computer screen, which allowed him to mimic a replay of the ventilation pattern recorded during his earlier exercise.
Determination of minute ventilation, V̇o2, and V̇no.
Subjects wore a nose clip and breathed through a mouthpiece connected to two large-bore (5.1-cm-diameter) hoses, which were part of a bag-in-box spirometer system. Breathing gas was humidified before reaching a vinyl reservoir bag. A 55-gallon drum containing a second (Douglas) bag functioned as the bag-in-box. A spirometer and potentiometer (Ohio Medical Products, Houston, TX) in the system continuously recorded volume changes [tidal volumes (Vt)]. A mass spectrometer (MGA 1100, Perkin-Elmer, Pomona, CA) and a NO analyzer (NOA; Sievers 280i, GE Water and Process Technologies Analytical Instruments, Boulder, CO) continuously sampled gas for composition at the subject's mouth. The mass spectrometer and the NOA were calibrated prior to each test. The NOA was calibrated using filtered room air (0 parts per billion NO) and a mixture of known gas concentration (45 parts per million NO). The mass spectrometer was calibrated to O2, CO2, and N2 concentrations using room air (21% O2 and 78% N2) and two commercially available gas compositions: 100% O2 and 9% CO2 in O2.
After each collection, exhalate was vented through a dry gas meter (750 Meter, Rockwell International Meter Division, Pittsburgh, PA) outside the chamber. A thermocouple (DP-80 Series, Omega Engineering, Stamford, CT) displayed the temperature inside the gas meter, and the mass spectrometer and NOA sampled a small volume of gas before it passed through the gas meter.
During experiments at the surface, the gas expelled from the Douglas bag was sampled through a gas meter (model 506164, Harvard Apparatus, Holliston, MA). A second mass spectrometer (MGA 1100, Perkin-Elmer) sampled the gas for composition just before it entered the gas meter, and a thermistor (Yellow Springs Instrument, Yellow Springs, OH) measured the temperature. There was no practical way to sample NO directly with this setup, so a 60-cm3 syringe was used to draw a gas sample for analysis by the NOA. Dry and submersed experiments were conducted in exactly the same manner, with the addition of a tight-fitting oronasal mask (as used in recreational SCUBA diving) with a pass-through port in the nose for the esophageal balloon tubing during submersion.
Measurement of pressure at the mouth and esophageal pressure.
A pressure transducer (model DP15-28, Validyne Engineering, Northridge, CA) connected directly to the mouthpiece continuously measured pressure at the mouth (Pmouth). The transducer sat within a third wide-bore hose, which allowed its reference port to remain open to ambient pressure (chamber pressure), even during submersion. A second pressure transducer (model DP15-30, Validyne Engineering) continuously measured esophageal pressure (Pes).
Measurement of Pa.
Pa was measured by a modification of the P0.1 technique. The subject's mouthpiece was connected to a custom-made breathing interrupter device, which was designed and built in-house. The interrupter consisted of a thin metal plate seated in a slotted plastic housing, which served as a guillotine valve that occluded airflow through the mouthpiece. The plate was driven by a pneumatic piston (model MRS-012-5DBFGUEE1JJ, Bimba Manufacturing, Monee, IL) connected by two pressurized air lines to a solenoid valve (model 125-4E1, Humphrey Products, Kalamazoo, MI). A computerized controller, designed and built in-house, directed the timing of guillotine valve operation by activating the solenoid valve. The piston housing held a magnet, which signaled when the piston was fully extended (occlusion complete). This signal was recorded by the data acquisition system for verification of occlusion during data analysis. Occlusion began ∼100 ms after the onset of inhalation or exhalation, occurred in ∼2 ms, and lasted ∼150 ms. These characteristics were chosen because they are similar to the closure characteristics of commercially available devices (48). Pressure excursions were measured by the Pmouth transducer and recorded as described below (see Data collection).
During exercise gas collection, a software program (written in-house) recorded the spirometer volume signal. At ∼30 s after the start of the collection, the subject performed a vital capacity (VC) maneuver, which provided volume reference values for use during ISEV.
The gas mixtures for ISEV were designed to maintain normocapnia and contained 4.20 ± 0.17% (mean ± SE) CO2 at 1 ATA, 1.67 ± 0.04% CO2 at 2.7 ATA, and 0.95 ± 0.01% CO2 at 4.6 ATA. The gas mixtures also included 21% O2, and the remaining fraction was N2. The computer software that was used to record the subject's breathing pattern during exercise simultaneously displayed, on a computer screen in front of the subject, the earlier recording and the current breath-to-breath volume changes of the spirometer. This allowed the subject to mimic the earlier recording and, thereby, simulate his own exercise breathing. The subject directly followed the pace on the computer screen while an investigator outside the chamber visually confirmed that he adhered to the recorded pattern. The subject performed a VC maneuver at the same time as the recorded tracing displayed VC, which allowed the investigator to adjust the baseline of the real-time spirometer display so that it and, therefore, the subject's end-expiratory lung volume were the same as during exercise.
Cardiac output (Q̇) was measured using a portable cardiac measurement device (Finometer model 1, Finapres Medical Systems, Amsterdam, The Netherlands). Because there was no way to render the device safe for hyperbaric exposure or the presence of water, the device was used only in the dry condition at 1 ATA. Furthermore, only the final four subjects were tested with the device.
A data acquisition system (MP150, BIOPAC Systems, Goleta, CA) recorded data for analysis. The signal from the spirometer, the Pmouth transducer, Pes transducer, ECG signal, fractions of O2 and CO2, concentration of NO, and interrupter signal were continuously output to the data acquisition system. Phase differences among the recorded variables (e.g., between Pmouth and Pes or between Pmouth and O2 concentration) were identified and corrected during data analysis.
Compression and decompression schedules.
US Navy compression and decompression schedules (42) were used conservatively in the dives. No decompression-related health problems were encountered.
V̇o2peak was determined at 1 ATA on the same cycle ergometer (pedaling rate 60 rpm) used for the experiments described above. The mouthpiece for determination of V̇o2peak was the same as that described above, except it did not include the interrupter device or pressure transducers. Subjects wore heart rate monitors (model 810i T61, Polar, Kempele, Finland). Before and after the test, blood drawn from a finger prick was analyzed for lactate (Accutrend Lactate, Roche Diagnostics, Mannheim, Germany).
Subjects initially pedaled at 50 W for 3 min and then at 100 and 150 W for 2 min each. The work rate was subsequently increased by 25 W every 2 min until the subject reached exhaustion. Frequency of breathing (f) and maximum heart rate were recorded at each work rate. Exhalate from the last minute of each work rate was collected and analyzed for volume, temperature, and gas composition using the equipment described above.
Subjects performed the test once in the dry condition and once while fully submersed in water. As described above, the ergometer was set to a work rate that was 25 W less than the corresponding dry work rate. Subjects wore diving face masks and nose clips while submersed.
Subjects were considered to have reached maximum V̇o2 if at least two of the following three criteria were met: 1) the increase in V̇o2 between the last gas collection and the gas collection preceding it was ≤150 ml O2; 2) the concentration of blood lactate following exercise was ≥4.0 mmol/l; and 3) the respiratory exchange ratio [CO2 production (V̇co2)/V̇o2] for the last collection was ≥1.15. Most subjects did not fulfill these criteria and, thus, were considered to have reached V̇o2peak under these conditions, as opposed to a true maximum.
Minute ventilation (V̇e) and V̇o2 were calculated using standard equations (13) and are presented in liters at standard temperature dry at ambient (chamber) pressure. V̇no was calculated using an equation adapted from that for V̇co2 (13). Pa was calculated from the Pmouth interruption data as described by Pao et al. (48), because this method of calculation generates reproducible values and is similar to calculations performed by commercially available devices (48). Briefly, Pa was determined by application of a linear regression to the region of increasing pressure in the Pmouth signal following the initial upstroke (expiration) or downstroke (inspiration). A second line was drawn perpendicular to the x-axis (time) at the time of closure [assumed to be complete at ∼25% of the overall rise (50)], and the slope from the linear regression was used to extrapolate back to the time of closure. The value of the regression at the time of closure was taken as Pa. Interrupter pressure curves that displayed reduced pressures prior to reopening of the interrupter were discarded, because leakage at the mouth could not be ruled out. Raw (P0.1 method) was calculated using the standard equation [Rint = (Pa − Pmouth)/flow]. Flow was the time derivative of volume at the onset of occlusion, and Pmouth was the pressure immediately prior to occlusion. Raw (esophageal balloon method) was calculated using the standard equation (13) [Raw = (Pmouth − Pes)/flow] at the time of peak flow during each breath.
Calculation of the work of breathing was based on long-established methods where the work of breathing is taken as the area within the pleural pressure-Vt loop (46). Here, the work of breathing per minute (power) was calculated from the product of flow and the pressure differential between Pes and Pmouth. This product was plotted vs. time, and the area was taken to provide work per breath. This was multiplied by f to provide work per minute (Ẇ). One limitation of this technique is that a portion of the work of breathing falls outside the loop, which leads to a slight underestimation of the work of breathing and prohibits separation of work into elastic and resistive components.
O2 cost of breathing and respiratory muscle efficiency.
The absolute O2 cost of breathing (V̇o2B) was calculated in two steps. 1) The increase in V̇o2 from rest to ISEV was calculated: it was equal to the difference in V̇o2 from rest to ISEV divided by the difference in V̇e from rest to ISEV. 2) This value was multiplied by V̇e during ISEV, which resulted in the total absolute V̇o2B for the ventilation pattern employed during exercise and ISEV. Respiratory muscle efficiency was the work of breathing per minute divided by the absolute V̇o2B.
After the V̇o2 and work of breathing experiment but prior to decompression, PEmax was determined at several different lung volumes (20, 40, 60, 80, and 100% of VC). Dry (i.e., nonsubmersed) VC was determined from the pulmonary function screening test. VC for submersed experiments was determined by having the subject perform several VC maneuvers while on the mouthpiece-spirometer system during rest in the previously described experiment. The largest peak-to-minimum value recorded from the spirometer was taken as submersed VC. To perform PEmax maneuvers in the wet, the subject was brought to the barrier interface and seated with the water level 1 in. above the suprasternal notch, which generated the same SLL as during the rest of the experiment. Pressures were measured by a pressure transducer (model DP15-28, Validyne Engineering) and recorded using the BIOPAC data acquisition system.
Analysis and statistics.
Data were analyzed using SigmaPlot 11.0 and SigmaStat 3.5 and are presented as means ± SE. A two- or three-way ANOVA was used to determine whether experimental variables, such as depth, submersion, and effort (i.e., rest, exercise, and ISEV), interacted or were independent of one another in their effects on the physiological variables. Differences between groups were assessed using paired t-tests. If normality and equal variance tests failed, a signed-rank test was used instead. Statistical significance was set at P ≤ 0.05 for all tests.
The ISEV phase of the protocol used in this study was designed to determine the energy cost of ventilation by quantifying the difference in V̇o2 from rest to normocapnic hyperpnea at rest. To our knowledge, this is a unique technique for measuring the O2 cost of exercise hyperpnea. The protocol required six experiments, which in most cases were conducted over a 6-wk period. This ostensibly could have produced learning and training effects. Repeated exercise bouts during the experiments did not, however, affect V̇o2peak in the dry or submersed condition (36 ± 2 and 32 ± 2 ml·kg−1·min−1, respectively). Repeated bouts of ISEV within the same experiment separated by 10 min did not affect V̇o2 during ISEV, which was 308 ± 54 ml/min during the first ISEV and 297 ± 38 ml/min during the second ISEV. Similarly, Q̇ was not significantly affected by ISEV (7.3 ± 0.2 l/min vs. 7.2 ± 0.4 l/min at rest), although it increased significantly during exercise (17.0 ± 1.2 l/min), as was expected. From these results, it was concluded that V̇e and V̇o2 during ISEV were not influenced by learning or training effects and that the increase in V̇o2 was not due to increased cardiac metabolism but, rather, to increased V̇e alone.
Ventilation and duty cycle.
During rest, V̇e was unaffected by submersion (compared with the dry condition) at all depths tested. In the dry condition, resting V̇e increased with increasing depth (20% increase from 1 to 4.6 ATA; Table 2).
During exercise, the effects of submersion and depth were markedly different from the effects observed at rest: V̇e increased with submersion and decreased with depth. Submersion induced a 14% increase in V̇e from dry to submersed conditions at 1 ATA (P = 0.033) and a 6% increase from dry to submersed conditions at 4.6 ATA (P = 0.018). As a result of increasing depth, V̇e decreased 11% from 1 to 4.6 ATA (P = 0.015) in the dry condition and 18% from 1 to 4.6 ATA during submersion (P = 0.006).
During ISEV, subjects were successfully paced at exercise ventilation, as indicated by the absence of significant differences in V̇e between exercise and ISEV. Therefore, by design, the effects of submersion and depth on V̇e during ISEV followed the same trends as during exercise: increases with submersion and decreases with depth.
Changes in V̇e resulted from changes in Vt and f (Table 2). At rest, there were no statistically significant differences in Vt between the dry and submersed conditions; however, Vt increased as a function of depth during submersion (26% from 1 to 4.6 ATA, P = 0.007). Similarly, during rest, f was unaffected by submersion. Depth had no effect on f, with the exception of a 16% decrease from 1 to 4.6 ATA during submersion (P = 0.008).
During exercise, trends in Vt were similar to those observed during rest: no change with submersion but an increase with depth [18% from 1 to 4.6 ATA in the dry condition (P = 0.024) and 13% in the submersed condition (P = 0.008)]. On the other hand, the effects of submersion and depth on f during exercise were very different from the effects observed during rest. During exercise, f increased with submersion (20% at 1 ATA, P = 0.020) and decreased with depth [27% from 1 to 4.6 ATA during submersion (P = 0.004) and 20% from 1 to 4.6 ATA in the dry condition (P = 0.029)].
Since subjects were paced during ISEV, the effects of submersion and depth on Vt and f, as could be expected, followed the same trends as during exercise. Thus, Vt was unchanged by submersion but tended to increase with increasing depth; similarly, f increased with submersion and decreased with depth.
During rest, submersion and depth had no effect on duty cycle, which is the duration of inspiration as a fraction of the total breath. During exercise, depth had no effect, but duty cycle increased 7% due to submersion at 1 ATA (P = 0.007). During ISEV, there were no statistically significant effects of depth or submersion on duty cycle.
Raw and V̇no.
RawI and RawE decreased from rest to exercise under all conditions tested. RawE did not differ between exercise and ISEV, nor did RawI, except at 2.7 ATA in the dry condition, where the value of RawI was slightly higher during ISEV than exercise (8.97 vs. 8.01 cmH2O·l−1·s, P = 0.031). For RawI and RawE, the effects of depth were independent, but the effects of submersion (dry and submersed) and effort (rest, exercise, and ISEV) were interrelated.
During rest, RawI decreased with submersion and increased with depth (Fig. 2A). The decrease in RawI with submersion was significant at 2.7 ATA (46% decrease, P = 0.009) and 4.6 ATA (34% decrease, P = 0.008). The increase in RawI with depth was not surprising, given the concurrent increase in gas density, and it reached significance at 4.6 ATA during submersion, where there was a 42% increase compared with 1 ATA (P = 0.013). Similarly, RawE decreased with submersion and increased with depth (Fig. 2B). Submersion decreased RawE 59% compared with the dry condition at 1 ATA (P = 0.016) and 68% compared with the dry condition at 2.7 ATA (P = 0.013). As a result of increasing depth, RawE increased 62% from 1 to 4.6 ATA during submersion (P = 0.023).
During exercise, RawI and RawE increased as a function of depth (Fig. 2, C and D), but only RawI decreased with submersion. The effect of submersion on RawI during exercise only reached significance at 1 ATA, where there was a 30% decrease (P = 0.008). Depth, on the other hand, produced a larger and more consistent increase in RawI during exercise than at rest: a 62% increase from 1 to 4.6 ATA in the dry condition (P = 0.008) and a 94% increase from 1 to 4.6 ATA during submersion (P = 0.007). RawE was unaffected by submersion but, like RawI, increased with depth to a greater degree than at rest: during submersion, there was a 139% increase in RawE from 1 to 4.6 ATA (P = 0.017).
Rint values were significantly higher than Raw values and were not as responsive to changes in submersion, depth, and effort as Raw. Neither RintI nor RintE was different from rest to exercise, and the effects of depth (1, 2.7, and 4.6 ATA), submersion (dry and submersed), and effort (rest, exercise, and ISEV) were independent of one another in their effects on RintI and RintE.
During rest, neither RintI nor RintE was affected by depth or submersion (Table 2). During exercise, submersion still had no effect on RintI, but depth increased it. There was a 449% increase in RintI from 1 to 4.6 ATA in the dry condition (P = 0.023) and a 153% increase from 1 to 4.6 ATA during submersion (P = 0.008). It is worth noting that these increases are substantially larger than the increases in RawI for the same conditions, perhaps due to greater inertial effects and less uniform flow at the start of the breath. RintE tended to increase with submersion during exercise, which was significant at 2.7 ATA (48% increase from the dry condition, P = 0.047). RintE also increased with depth, which was significant from 2.7 to 4.6 ATA in the dry condition, where there was an 89% increase (P = 0.049), a similar magnitude to the increase in RawE.
V̇no was studied for its reported relationship with Raw and was expected to correlate negatively with Raw as a function of depth (i.e., as depth increased, V̇no would decrease due to the higher Po2 and Raw would increase due to increased gas density and reduced V̇no). In fact, there was no correlation between Raw or Rint and V̇no as a function of depth or within individual test conditions (e.g., surface dry). At rest, V̇no was unaffected by depth and submersion (mean 60.2 ± 9.3 μl/min). Submersion had no effect on V̇no during exercise. The effect of depth on V̇no during exercise was also minimal, except at 2.7 ATA during submersion, when V̇no reached a nadir (99% less than at 1 and 4.6 ATA, P = 0.023 and 0.016, respectively). Additionally, V̇no was not different from rest to exercise, except at 4.6 ATA during submersion, when it was 582% greater than the resting value (P = 0.008).
The magnitudes of PaI and PaE increased from rest to exercise for all conditions, as expected. The effects of depth (1, 2.7, and 4.6 ATA), submersion (dry and submersed), and effort (rest, exercise, and ISEV) were independent of one another in their effects on PaI and PaE.
During rest, PaI was more negative as a result of submersion and depth (Fig. 3A). The effect of submersion reached significance at 2.7 ATA: PaI was 25% more negative than in the dry condition (P = 0.020). PaI was also more negative with increasing depth: 166% more negative at 4.6 than 1 ATA in the dry condition (P = 0.008) and 115% more negative at 4.6 than 1 ATA during submersion (P < 0.001). In contrast to PaI, resting PaE was largely unaffected by depth and submersion (Fig. 3B). Depth had an effect only in the dry condition, when PaE increased 99% from 1 to 4.6 ATA (P = 0.001).
During exercise, the effects of depth and submersion on PaI were similar to those during rest (Fig. 3C). PaI was more negative with increasing depth and tended to be more negative with submersion, although the effect of submersion did not reach significance. PaI was 23% more negative at 2.7 than 1 ATA in the dry condition (P = 0.026) and 72% more negative at 4.6 than 1 ATA during submersion (P < 0.001). In contrast, during exercise, PaE increased with submersion but was unaffected by depth (Fig. 3D). Submersion caused an 81% increase in PaE compared with the dry condition at 1 ATA (P = 0.004) and a 90% increase from the dry condition at 2.7 ATA (P = 0.008).
Power of breathing.
ẆI and ẆE increased from rest to exercise under all conditions, as expected, and the effects of submersion (dry and submersed), depth (1, 2.7, and 4.6 ATA), and effort (rest, exercise, and ISEV) were independent of one another in their effects on ẆI and ẆE.
During rest, ẆI decreased with submersion and increased with depth (Fig. 4A). The decrease from submersion was statistically significant at 2.7 ATA, where there was a 26% decrease compared with the dry condition (P = 0.031). The increase in ẆI with depth was 53% from 1 to 4.6 ATA in the dry condition (P < 0.001) and 35% from 1 to 2.7 ATA during submersion (P = 0.043). In contrast to ẆI, ẆE was unaffected by submersion but increased with depth during rest (Fig. 4B). There was a 62% increase in ẆE from 1 to 4.6 ATA in the dry condition (P = 0.007) and a 72% increase from 1 to 4.6 ATA during submersion (P = 0.021).
During exercise, ẆI tended to increase with submersion and depth (Fig. 4C). The effect of submersion reached significance at 4.6 ATA, where there was a 19% increase in ẆI from the dry condition (P = 0.020). Increasing depth resulted in an increase of 20% from 1 to 4.6 ATA in the dry condition (P = 0.022). In contrast, ẆE was unaffected by submersion and depth (mean 639 ± 305 mW; Fig. 4D).
O2 cost and efficiency of breathing.
It was hypothesized that V̇o2B would increase as a function of depth due to the increased work of breathing and would increase with submersion due to the mechanical disadvantages imposed by blood redistribution to the lungs and the presence of the SLL; however, there were no significant effects of submersion on V̇o2B. Depth tended to increase V̇o2B during submersion, but this effect did not reach statistical significance (Fig. 5A).
Respiratory efficiency decreased with submersion (49% decrease from dry condition to submersion at 2.7 ATA, P = 0.023) but was unaffected by depth (Fig. 5B).
As expected, PEmax increased with increasing lung volume and with increasing depth (Fig. 6). Submersion had no effect on PEmax at any depth, so wet and dry data have been pooled. PEmax increased from 20% of VC to total lung capacity at all depths [14% increase at 1 ATA (P = 0.030), 19% increase at 2.7 ATA (P = 0.003), and 17% increase at 4.6 ATA (P = 0.002)].
PEmax also increased as a function of depth, although the increase only reached significance at the higher lung volumes. At a lung volume of 80% of VC, there was a 17% increase from 1 to 2.7 ATA (P = 0.035). At total lung capacity, there was an increase of 15% from 1 to 4.6 ATA (P = 0.033).
V̇e and its components.
During rest, Vt was the primary determinant ofV̇e: as Vt increased with depth, so did V̇e. Such an increase has been documented previously (43, 57) and may be explained by several processes. The most recent hypothesis is hyperoxic hyperventilation, which may occur following blunting of the peripheral chemoreceptor response by the high Po2 of the hyperbaric environment. After 5–10 min at depth, reactive oxygen species accumulate in brain tissue and stimulate central chemoreceptors in the caudal solitary complex (19), which should stimulate hyperventilation, although this effect is likely dampened by feedback within the system and has only been observed in vivo following carotid body denervation (15, 24). It should also be noted that this mechanism would lead to a decrease in arterial Pco2, and the majority of studies have not produced this result (23). Another mechanism that may contribute to the hyperventilation observed during rest is the Haldane effect, whereby incomplete O2 unloading from hemoglobin interferes with CO2 transport and leads to CO2 retention, which in turn raises Pco2 and increases respiratory drive (59). As discussed by Lambertsen et al. (34), this may be particularly deleterious to the brain and could further stimulate central chemoreceptors. It should be noted, however, that modeling based on quantitative studies indicates that the resulting increase in arterial Pco2 would be <1 mmHg and is therefore likely insignificant (23). Finally, it has been reported that as gas density increases, respiratory dead space and the dead space-to-Vt ratio increase, most likely as a result of decreased gas mixing in the lungs (41). Thus, V̇e and Vt would have to increase to maintain alveolar ventilation at depth. This is the most likely explanation for the increase in Vt and V̇e with depth during rest.
In contrast to rest, V̇e increased with submersion but decreased as a function of depth during exercise. The increase in V̇e during submersion was most likely due to a slightly higher metabolic rate (as evidenced by increased V̇o2 and V̇co2), despite an attempt to set the work rate equivalent to the dry condition. Although increases in V̇o2 and V̇co2 during submersion generally were not statistically significant, they may have been sufficient to elevate V̇e. An alternative explanation is the effect of submersion in resting critical water temperature, which is cool enough to cause an effect (18); however, during exercise, this water temperature is thermoneutral and may not have had an effect (10).
It is interesting to note that there was a concurrent increase in duty cycle with submersion, which may reflect the greater inspiratory effort required to overcome the SLL. The decrease in V̇e as a function of depth occurred despite increasing Vt at depth and may reflect a shift to a ventilation pattern that increased elastic work. This would help offset the increase in resistive work experienced at greater gas densities and may also help maintain alveolar ventilation in spite of increasing dead space, as discussed above. The alterations in V̇e may therefore reflect a rebalancing of respiratory drive toward a more efficient breathing pattern, a notion that is supported by the observation that human subjects at sea level spontaneously breathe at a frequency and Vt that closely approximate the most efficient pattern (39, 47).
Raw, Pa, and the work of breathing.
The overall trends in V̇no were surprising, given the established role of NO as a bronchodilator (5). During rest and exercise, V̇no was unaffected by submersion. The nadir in V̇no that was observed at 2.7 ATA during exercise is puzzling but likely results from the interaction of more than one variable. It is possible that V̇no decreased as a result of hyperoxia, as previously proposed (20), but this effect may have competed with increases in V̇no typically seen during exercise (11, 29). The balance of these two effects produced a nonuniform trend as depth increased but exercise intensity remained constant. This is also the most likely explanation for the lack of increase in V̇no from rest to exercise that has typically been observed (11, 29, 37). Our subjects exhibited a high degree of variability, which is most likely due to the fact that V̇no varies between trained and untrained subjects (11, 37) and as a function of height and age (44). Given that our cohort was small and encompassed individuals with a broad age range and varying heights and levels of physical fitness, it is not surprising that our results exhibit a wide standard error. The absence of correlation between V̇no and Raw, even within a single condition, suggests that control of airway caliber and, therefore, resistance was multifactorial and not primarily determined by V̇no, at least under the conditions tested. Several studies have shown, under normobaric conditions, that V̇no fluctuates according to Po2 in the surrounding tissue and have implicated NO in maintenance of tissue normoxia (6, 22). The observation that V̇no decreases with depth (i.e., increased Po2) is consistent with this understanding and suggests that the alterations in V̇no observed in this study were dictated more heavily by demands to maintain tissue normoxia than by other factors, such as Raw.
We chose to compare the traditional, esophageal balloon-based method of calculating Raw with the newer P0.1 method (Rint), as the P0.1 method may more closely reflect Pa, and its use has become more prevalent. If the two methods were found to produce similar values and were equally sensitive to alterations in Raw, the P0.1 method would be desirable for future studies because of the ease of use of the devices and improved subject comfort from elimination of the esophageal balloon. However, the commercially available devices make this measurement 100 ms after the onset of the breath, which may not reflect the same pressures and resistances observed during the portion of the breath with uniform flow (as in the esophageal balloon method). While both measurements of airway resistance (Raw and Rint) detected increased resistance as a function of depth, the values given by the P0.1 technique differed substantially from those of the esophageal balloon method. The P0.1 technique used in this study measured the initial portion of the breath and, thus, did not detect any effect of depth or submersion during rest. Furthermore, the values obtained using this method were substantially higher than those obtained using the esophageal balloon method. One possible explanation for this discrepancy is that Pa, used in the Rint calculations, describes the pressure differential between the lung and ambient pressure and, thus, includes the hydrostatic pressure exerted on the chest during submersion, while pleural pressure (approximated by the esophageal balloon and used for Raw calculations) only describes the pressure differential generated between the lung and the chest wall and, thus, does not account for hydrostatic pressure on the chest. It is not surprising, then, that Raw and Rint measurements differed during submersion. The discord between the two methods in the dry condition is puzzling, although it probably results from the difference in the timing of the measurement (midbreath vs. early breath). Rint measurements were made at the very start of the breath, when little air was flowing (which would increase the inertial component) and airway diameter was likely to be smaller (which would directly increase flow resistance). These effects apparently contributed to a significantly increased Rint compared with Raw and were so substantial that they overshadowed many of the smaller changes that were detected using the esophageal balloon method. For these reasons, we do not recommend the P0.1 method for assessment of Raw in the diving environment unless the measurement can be made during peak flow.
Overall, our measurements of Raw were in agreement with some published values (38) but were higher than those that are typically reported (56, 62), even in studies of asthma (25, 27, 32, 55). We propose that the elevated values we observed derived from the subjects' ventilation and rates, which were somewhat higher than normal during rest and would have proportionally increased resistance according to Poiseuille's law. Elevated ventilation may have occurred if subjects were anxious at the beginning of each experiment or as subjects adjusted to breathing through the mouthpiece (4).
As expected, Pa and Raw increased as a function of depth, most likely due to increased gas density, which has been shown to increase Raw by the square root of the increase in density (49). Pa increased from rest to exercise, probably due to increased demand and airflow, while Raw decreased during exercise, most likely due to increased airway caliber. These findings are in accordance with the works of others (31), and it is generally well accepted that exercise-induced bronchodilation stems from a combination of neural control, primarily via the vagus nerve (61), and increases in circulating catecholamines, such as epinephrine (52, 58).
On the basis of the literature, we expected that submersion would not alter Raw (7) or would increase it secondary to decreased functional residual capacity (9, 35), which immersion and negative SLL have been shown to produce (54). Thus we were surprised to observe decreases in RawI and RawE with submersion during rest. The decrease in RawE occurred, despite the fact that related variables (f, Vt, V̇e, PaE, and ẆE) were almost entirely unchanged as a result of submersion. Thus we can only assume that airway caliber must have been larger than in the dry condition. This is also unexpected, given that we observed a decrease in VC, consistent with the observations of others (3, 33), and such a decrease in lung volume is usually associated with increases in Raw (8). We can only surmise that immersion of the whole body, as opposed to head-out submersion, may have played a role. For instance, in attempting to prevent air from escaping around the mouth or through the nostrils, which experienced positive static loading, it is possible that subjects recruited accessory muscles in the neck and shoulders, which may have increased airway caliber. RawI also decreased with submersion during exercise. This occurred despite increased V̇e and f (greater flow). Again, the simplest conclusion is that the airways were more open during submersion than in the dry condition. The more-negative PaI observed during submersion ostensibly reflects the effect of SLL, and not submersion per se, which would increase chest wall elastic recoil at the start of inspiration (54); however, this cannot be stated with certainty, since functional residual capacity was not measured.
As might be expected, changes in Ẇ paralleled changes in Raw and Pa. In general, Ẇ increased as a function of depth. We surmise that this was due to the increased Pa required to move denser air against the increased Raw, although the effect of altered Pa cannot be stated for certain, since pressure was measured only at the beginning of inspiration and expiration, and not in the middle of the breath when flow was greater. Submersion decreased ẆI at rest but increased it during exercise. The decrease in ẆI during submersion at rest was due to less work per breath (data not shown) and correlated with the decrease in RawI. During exercise, f was elevated during submersed experiments, but the work per breath was unchanged between dry and submersed conditions (data not shown). The result was a higher ẆI during submersed exercise than in the dry condition. Finally, ẆE was largely unaffected by submersion. This probably results from the negative SLL, which assisted exhalation. Interestingly, depth increased ẆE during rest, but not exercise. The increase during rest was expected, given the increase in gas density, but the lack of increase during exercise is surprising. It is possible that expiratory assistance from the negative SLL overshadowed any hindrance as a function of depth.
V̇o2B and respiratory efficiency.
Accurate measurement of efficiency and V̇o2B hinged on three factors: 1) subjects' ability during ISEV to accurately duplicate the recorded breathing pattern, 2) the necessity that increases in V̇o2 from rest to ISEV were due solely to increased ventilatory effort, and 3) the assumption that respiratory and accessory muscle recruitment was similar between exercise and ISEV. On the basis of the data obtained, subjects reliably reproduced the exercise ventilation pattern, as there were almost no differences between exercise and ISEV ventilation (with regard to V̇e, Vt, f, etc.). Similarly, Q̇ was not different from rest to ISEV, which suggests that the increased V̇o2 was due solely to respiratory effort. While we cannot rule out the possibility that subjects recruited muscle groups differently during ISEV than during exercise, subjects were monitored during the experiment, and use of their arms for postural support during ISEV was discouraged. We thus believe that the calculated V̇o2B is representative of O2 use by the respiratory muscles during exercise. Finally, V̇o2B for our subjects was similar to that described by others (1), and any difference probably is due to the fact that our study subjects were less physically fit than those in the study by Aaron et al. (1).
For submersed conditions, V̇o2B tended to increase with depth due to an increase in the work of breathing and a concurrent tendency for respiratory muscle efficiency to decrease. Efficiency decreased slightly with submersion, so, overall, V̇o2B paralleled changes in Ẇ.
PEmax increased with lung volume and increasing depth. The increases in lung volume are consistent with previously published observations (14) and probably reflect more optimal respiratory muscle length at higher lung volumes (51). Additionally, preservation of a favorable muscle length during expiratory and inspiratory efforts is sustained by the reduced compressibility of gas in the lungs during the maneuver. These findings are not surprising for several reasons. It is difficult to generate pressure against a small volume of air in the lungs, because the elastic recoil of the chest prevents the ribcage from contracting to a small enough volume to exert significant pressure. At higher lung volumes, the elastic recoil of the chest wall is diminished, and past its relaxation volume, the elastic recoil of the chest wall inward supplements contraction of the expiratory muscles. Since subjects' VCs were measured in dry (i.e., normal perfusion) and submersed (when more blood is present in the subjects' lungs) conditions and these values were used to calculate the amount of air subjects inhaled, the volumes during submersion were normalized compared with the dry volumes. For this reason, it is not surprising that there were no differences in PEmax between dry and submersed conditions. It is possible that the presence of additional blood could have served to elevate PEmax by serving as an incompressible substrate (much as the dense air at increased depth), but either this was not significant or venous return decreased in response to the high pressures in the thorax and prevented the presence of a sufficient amount of blood to have an effect.
In summary, submersion did not have an effect on PEmax. Increased pressure did not reliably decrease V̇no or increase the work of breathing in the dry condition, although it did increase Raw, Pa, respiratory work during submersion, and PEmax. These changes should be considered when assessing the impact of submersion and depth on divers who are undertaking rigorous physical activity, when assessing fitness to dive, and when providing HBOT to patients with compromised ventilation.
This research was sponsored by Office of Naval Research Grant N00014-08-1-0255. H. E. Held's doctoral position is supported by the Office of Naval Research.
No conflicts of interest, financial or otherwise, are declared by the authors.
H.E.H. and D.R.P. are responsible for conception and design of the research; H.E.H. performed the experiments; H.E.H. analyzed the data; H.E.H. and D.R.P. interpreted the results of the experiments; H.E.H. prepared the figures; H.E.H. drafted the manuscript; H.E.H. and D.R.P. edited and revised the manuscript; H.E.H. and D.R.P. approved the final version of the manuscript.
We acknowledge the involvement of Claes Lundgren. We thank the staff of the Center for Research and Education in Special Environments, Andrew Barth, Lukas Eckhardt, Michael Fletcher, Ron Okupski, Curtis Senf, Amber Simpson, Eric Stimson, and Matthew Vargo, for assistance in data collection. We also thank Nancy Niedermayer for organizational work.
- Copyright © 2013 the American Physiological Society