Pulmonary effects of repeated six-hour normoxic and hyperoxic dives

This study examines differential effects of immersion, elevated oxygen partial pressure, and exercise on pulmonary function after series of five daily six-hour dives at 130 kPa (1.3 ATA), with 18 hours between dives. Five cohorts of 10 to 14 divers participated. The exposure phases were resting while breathing O2 or air in the water (“wetO2”, “wetAir”) or O2 in the hyperbaric chamber (“dryO2”), and exercise in the water while breathing O2 or air (“wetO2X”, “wetAirX”). Respiratory symptoms were recorded during and after each dive, and pulmonary function (forced flow-volume) was measured twice at baseline before diving, after each dive both immediately and on the following morning, and three days post diving (“Day+3”). The incidences of symptoms and of flow volume changes from baseline greater than normal limits (“ΔFV”) were assessed, as were mean ΔFV. The parameters examined were forced vital capacity (FVC), forced expired volume in 1 second (FEV1), and forced expired flow from 25% to 75% volume expired (FEF25–75). The phases ranked from greatest to least fraction of diver-days with symptoms were wetO2X (56%) > dryO2 (42%) > wetO2 (13%) > [wetAir (2%) or wetAirX (1%)] (p<0.05). FEV1 and FEF25–75 were depressed in the morning following wetO2 and wetO2X and on Day+3 after and wetO2X, but increased immediately following each wetAirX dive. O2 exposures caused symptoms and ΔFV suggestive of pulmonary oxygen toxicity,exacerbated by exercise. Indices of small airway function showed late (17-hour) post-O2 exposure deficits, but, particularly with exercise, improvement was evident early after exposure with or without O2. FEF25–75 and FEV1 remained depressed on Day+3 after wetO2 and wetO2X.


Introduction
Pulmonary oxygen toxicity is a known hazard of long exposures to oxygen partial pressure (PO 2 ) greater than about 50 kPa (0.5 atm) [1]. In previous work, single long duration (four-to eight-hour) dives with PO 2 = 130 to 140 kPa (1.3 to 1.4 atm) provoked mild pulmonary oxygen toxicity in some individuals whether divers on the bottom were resting [2] or exercising intermittently on cycle ergometers for four hours [3]. When four-hour dives were repeated daily for five days, pulmonary symptoms and changes in pulmonary function were minimal for divers who rested on the bottom. However, when the divers exercised on the bottom reports of a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Ethics, consent, and permissions
All studies were approved by the Institutional Review Board at Navy Experimental Diving Unit. Each subject gave written informed consent before participating, and all procedures conformed to the Declaration of Helsinki.

Subjects
Healthy U.S. Navy divers dove for six hours per day on five consecutive days, with 18-hour surface intervals (SI). Before and after each day's diving, forced flow-volume loops (forced vital capacity maneuvers) were recorded using a spirometer-based pulmonary function machine (CPL, nSpire Medical, Longmount CO). The loops were examined for quality and consistency of values, but only FVC, FEV 1 , and FEF  were compared statistically and are presented. The times of measurement for each individual were consistent. Morning measurements were made from 5:30 AM to 8:30 AM and afternoon measurements from about 2 PM to about 5 PM. Forced vital capacity maneuvers were also conducted within the week before diving, on the morning following the fifth dive, and three days after the fifth dive. The average of parameters from three flow-volume loops reproducible according to American Thoracic Society standards [13] was used at each time point, and was compared to a baseline average of six reproducible values, three from the week before diving and three from the morning before the first dive. In-water ("wet") dives were conducted in a 4.6 m (15 foot) deep, fresh water pool, kept comfortably warm (31.7 ± 1.7˚C [89 ± 3˚F] rest, 30.6±1.7˚C [87±3˚F] exercise). At those temperatures the divers needed no thermal protection, and wore just shorts and T-shirts, as they did in the lab before submersion and after changing into dry clothes after diving. Divers with lung centroid at a depth of 3.4 to 4.0 m (11 to 13 feet) breathed surface-supplied, humidified gas, open circuit from full face masks with self-contained demand regulators (AGA Divator mask, Interspiro, Cliffwood Beach NJ).
Up to four divers participated during any dive week. They entered the water individually at intervals of about an hour. After three hours of immersion they surfaced to stand in shoulderdeep water and breathe room air during a 10-minute lunch break. The breathing gas underwater was either air ("air dives") or 100% oxygen ("O 2 dives"), for a PO 2 of approximately 30 or 130 kPa (0.3 or 1.3 atm), respectively. Divers either rested in chairs on the bottom ("resting dives") or exercised prone but about 30˚head-up on in-water cycle ergometers ("exercise dives"). During exercise dives, divers alternated 30 minutes of seated rest with 30 minutes of exercise, and ergometer brake power was adjusted to generate heart rates of 90 to 110 beats/ min. The load was an attempt to mimic the workload of a hypothetical average dive without current.
Five different sets of dive conditions were investigated, four in the water and one in a dry hyperbaric chamber. Different divers participated in the five phases. Dry hyperbaric exposures ("dry dives") were conducted in a large hyperbaric chamber that was compressed with air to 130 kPa. Seated divers breathed humidified oxygen from hoods (Amron 8891, Amron International, Vista, CA). Divers locked into the chamber individually at approximately one hour intervals. They removed the hoods after three hours exposure to breathe chamber air and eat and drink for no more than 10 minutes before resuming oxygen breathing.
Specific symptoms-inspiratory burning, cough, chest tightness or dyspnea-were considered to indicate pulmonary oxygen toxicity. Subjects were queried about symptoms, which were scaled as "none", "mild", "moderate", "moderately severe", or "severe", whenever pulmonary function was measured and hourly during the dives. A diver who reported any symptom from that list was considered to have possible pulmonary oxygen toxicity on that day (S1 Table, S2 Table). When the fraction of person-days with symptoms was calculated, a full diverday was allocated.
Changes in flow-volume parameters that fell outside the normal non-diving 95% confidence limits as previously measured at the Navy Experimental Diving Unit [2] also were considered to be evidence of pulmonary oxygen toxicity. Specifically, decreases from baseline in FVC greater than 7.7%, in FEV 1 greater than 8.4%, or in FEF 25-75 greater than 16.8% were considered to be deficits, collectively termed "ΔFV". Incidences of symptoms and of ΔFV were compared across dive conditions using Fisher's Exact Test.
To account for differences across dive phases, fractional changes from baseline in pulmonary function indices were compared rather than the raw values. The average fractional changes from baseline in flow-volume parameters FVC, FEV 1 , the ratio of FEV 1 /FVC, and FEF  were assessed across dive weeks.
For the in-water data, the initial statistical analysis was an analysis of variance (ANOVA) with two between-subjects factors, gas (oxygen or air delivered) and rest or exercise, with repeated measures on the within-subject factors dive day (Dives 1 to 5) and time of measurement (immediately after a dive ["PM"] or the following morning ["AM"]). Differences were considered to be significant if α = 0.05 after Greenhouse-Geisser (G-G) correction for nonsphericity as appropriate. If a significant interaction of gas was found with rest and exercise or with time of day, repeated measures ANOVAs were repeated on each of the conditions, that is, one ANOVA for rest and one for exercise, or one for AM and one for PM. Further interactions including gas were similarly investigated, then time of measurement, with the final drillingdown a Şidak-corrected pairwise comparison of dive days and comparison of dive days to baseline.
The resting oxygen dives in the hyperbaric chamber (DryO 2 ) were compared to the resting in-water oxygen dives (WetO 2 ) with a one factor (wet or dry) repeated measures ANOVA. Measurements within the condition were made using repeated measures ANOVA with no factors. Follow-up measurements were compared to baseline using t-tests.

Results
Numbers of divers and their characteristics are presented by type of dive in Table 1.
All participants were healthy and had met physical examination criteria for diving. Nevertheless, a number had baseline FEV 1 /FVC ratios lower than 75%: one diver for wet resting O 2 , two for wet resting air, one for dry resting O 2 , two for wet exercise O 2 , and five for wet exercise air. Divers were included in the study regardless of baseline FEV 1 /FVC ratio. All analysis was of comparison to baseline, and population normative values were not considered.

Occurrence of pulmonary signs and symptoms
The fractions of divers who reported symptoms or who had measurable changes in flow-volume (ΔFV) parameters are reported in Table 2 for the five phases. Divers frequently reported symptoms without measurable flow-volume deficits ("signs"), and occasionally vice versa; symptoms, signs, and both (i.e., any pulmonary oxygen toxicity) are tabulated individually. The incidences include those reported during a dive, those evident immediately post-dive and those first measured or reported on the following morning. From Table 2 one cannot differentiate between multiple individuals each affected in relation to a small number of dives and a small number of people affected across many dives. Fig  1 shows the number of individuals affected one or more times during the dive weeks. The values are the number of divers with symptoms, ΔFV, or both any time during the dive week expressed as a fraction of total number of divers for the condition. They are an estimate of the probability that any one diver will have symptoms or signs of pulmonary oxygen toxicity at any time during a dive week.
Exposures with elevated PO 2 , wet or dry, were associated with higher probabilities of symptoms and of combined symptoms and signs than were air dives. By Fisher's Exact Test, the fraction of divers with either signs or symptoms even once during the dive week with dry resting O 2 dives did not differ from wet resting O 2 dives. Similarly, the number affected during wet resting O 2 or air dive weeks did not differ from those during wet O 2 exercise or air exercise dive weeks, respectively. The numbers during wet exercise air dive weeks were lower than those during wet exercise O 2 dive weeks (p<0.01), and those during wet resting air dive weeks were marginally lower than those during wet resting O 2 dive weeks (p<0.06). Another index, the fraction of diver-days with evidence of pulmonary oxygen toxicity ( Fig  2), may give a better indication of total pulmonary insult. The number of diver-days with evidence of pulmonary oxygen toxicity was divided by the total number of diver days (e.g., 50 for ten dives for a dive week). Here also, elevated PO 2 was associated with greater incidence of pulmonary oxygen toxicity; wet O 2 dives showed a higher fraction of diver-days with symptoms than did wet air dives (p<0.04 at rest, p<0.01 with exercise). However, for this statistic, submergence appears protective; dry O 2 dives had a significantly larger (p<0.01) incidence of symptoms than did wet O 2 dives. Further, submerged exercise increased the effects of O 2 ; wet resting O 2 dives produced fewer symptoms than did wet exercise O 2 dives (p<0.01), but wet resting and wet exercising air dives did not differ from each other.
The numbers of person-days with ΔFV showed fewer differences across conditions than did the numbers with symptoms because ΔFV was a rarer event. The number of person days with ΔFV was not different between wet resting air and dry resting oxygen, nor between resting and exercise air. Additionally, the number of person days with ΔFV did not differ between wet resting and wet exercise oxygen. However, exercise exacerbated the effects of wet O 2 to the extent that exercise dives with O 2 produced significantly more (p<0.01) person-days with ΔFV than did exercise dives with air. Fig 2 reflects the aggregate for all divers across the dive weeks but does not describe the variability among divers. As shown in Fig 3, the variability is quite large both for number of days with signs and symptoms, and for severity as assessed by number of separate signs and symptoms ("number of problems") per day. The full individual tabulation appears in the Supporting Information (S1 Table, resting dives and S2 Table, dives with exercise). Probability of evident pulmonary oxygen toxicity on one or more days of a dive week. The incidence was calculated as the number of divers with evidence of toxicity even once during the week divided by the total number of divers. Solid grey: incidence of symptoms; hash marks: incidence of ΔFV; and grey with hash marks: symptoms and/or ΔFV, that is, any evidence of pulmonary oxygen toxicity. Error bars represent the binomial 95% confidence intervals. The statistical differences that are noted apply to the combined evidence of pulmonary oxygen toxicity. Dive conditions labeled "a" or "b" do not differ from other similarly-labeled conditions by Fisher's Exact Test. " Ã " and the arrow indicate a significant pairwise difference (p<0.01). Two divers who did not complete the dive week because of symptoms (wet exercise O 2 dives) were included in both numerator and denominator for the days when they participated. Fraction of person-days when divers showed evidence of pulmonary oxygen toxicity. Solid grey: incidence of symptoms; hash marks: incidence of ΔFV, and grey with hash marks: symptoms and/or ΔFV. Error bars represent the binomial 95% confidence intervals. Statistics in roman face represent symptoms comparisons indicated by the ends of the solid lines, and those in italics, the ΔFV comparisons indicated by the dashed lines (Fisher's Exact Test). Dives labeled "a" do not differ from each other for either indication. Two divers who did not complete the dive week because of symptoms (wet exercise O 2 dives) were included in both numerator and denominator for the days when they participated.
https://doi.org/10.1371/journal.pone.0202892.g002  1. WetAir, WetAirX: almost no one was affected, although one diver in the resting air series showed more than one change at nearly all assessments; 2. WetO 2 , DryO 2 : half of divers showed evidence of pulmonary oxygen toxicity on one quarter to one half of days and one or two divers had evidence of toxicity every day, though three divers wet and one dry had no evidence of pulmonary oxygen toxicity; 3. WetO 2 X: half of divers manifested pulmonary oxygen toxicity on nearly every assessment day, with an average of 1.5 problems per day, though one diver had no signs or symptoms on any day, and one averaged three problems per dive.

Mean values
Baseline. Baseline values are listed in Table 3. Changes in mean values of pulmonary function indices were minimal in relation to normal variation, but some statistically significant changes did occur. Fractional changes from baseline for all divers are shown in Figs 4-8.
Submersion effects. None of FVC, FEV 1 , or FEF  was significantly different between WetO 2 and DryO 2 (both resting dives). However, resting wet and dry O 2 exposures were not considered together because of marginal interactions of the wet-or-dry condition and dive day for some of the variables.
Within the dry dive FEV 1 and FEF  varied with time of day (p<0.006, p<0.003, respectively), with AM values lower than PM values. FEF 25-75 also changed with dive day, with only AM values after Dives 1 and 2 less than baseline. O 2 effects in the water. FVC. There were no significant differences in FVC ascribable to breathing gas.
FEV 1 . FEV 1 tested across all in-water dives showed no effect of breathing gas per se (no main effect), and no significant interactions of breathing gas with other variables.
FEF  . For FEF  there was a significant effect of breathing gas (main effect p<0.008), as well as a significant interaction of breathing gas with time of day (p<0.04), and a marginal interaction with rest or exercise (p<0.06). Breathing gas had no significant effect on FEF  PM, while FEF 25-75 AM was significantly more depressed with O 2 than with air (p<0.002) with no interactions with rest or exercise or dive day.
Exercise effects in the water. FVC. FVC showed only a marginal difference (p<0.07) between rest and exercise, with values slightly larger after resting dives than after exercise dives.
FEV 1 . FEV 1 showed a strong effect of exercise (main effect p<0.005, but the exercise effect also interacted significantly with time of day (p<0.05) and with dive day (p<0.04). FEV 1 PM  was greater after exercise dives than after resting dives (p<0.002), though for neither rest nor exercise dives was any individual PM value different from baseline. However, FEV 1 AM was only marginally different (p<0.1) between rest and exercise, with more depression from Pulmonary effects of repeated six-hour dives  baseline after resting dives than after exercise dives. For FEV 1 AM, rest or exercise interacted significantly (G-G p<0.003) with dive day; the exercise conditions differed significantly (exercise FEV 1 AM was lower) only on the mornings after Dives 3 and 4. After WetAirX, FEV 1 PM was greater than baseline after Dives 1 and 5. For all other phases and dives, it did not differ from baseline.    . There was a strong effect of exercise FEF  (main effect <0.006), but also an interaction of exercise with time of day (p<0.03). FEF  PM was larger after exercise dives than after resting dives (p<0.0003), not different from baseline after the resting dives but significantly elevated in the afternoon after Dives 1, 4, and 5. FEF  AM showed only an interaction between rest and exercise and dive day; after resting dives, dive day did not significantly affect FEF 25-75 AM, but FEF 25-75 was depressed more after Dive 4 at rest than Dive 4 with exercise. After WetAirX, FEF  PM was greater than baseline, and after all other phases, it did not differ from baseline for any dive.
Follow-up. At the follow-up measurements three days after the fifth dive, FVC and FEV 1 were below baseline for WetO 2 , FEV 1 and FEF  values were below baseline for WetO 2 X, FVC was depressed for WetAirX, and a few other values remained marginally different from baseline (Table 4).

Discussion
The comparisons of dry oxygen to wet oxygen dives and wet oxygen to wet air dives indicate that the constellation of signs and symptoms strongly suggested pulmonary oxygen toxicity rather than an immersion phenomenon of some sort: 1) strikingly few divers had any pulmonary signs or symptoms after wet air dives, even with exercise (Fig 3); 2) incidences of signs and symptoms were high only for oxygen dives ( Table 2); 3) the probability of symptoms or signs at least once during the weeks was not different among the different oxygen dives, including the dry dives, but was greater for oxygen than for air dives (Fig 1); and 3) dry oxygen dives generated more symptoms on all but the first dive day ( Table 2) and more person-days with symptoms (Fig 2) than did wet oxygen dives. Thus, the hypothesis that large pressure swings in the presence of pulmonary congestion is responsible for altered pulmonary function indices and for symptoms after diving was not substantiated. Exercise exacerbated the oxygen effect without itself causing any change in signs or symptoms after air dives.
The incidence of symptoms or of ΔFV in any individual is very different from the average change in pulmonary function across a group. This study, like others reported previously [2,3], shows that average changes are minimal. Nonetheless, some individuals experiencing pulmonary oxygen toxicity have results far from the group mean (Figs 4-8). Also like the previous reports, this study shows the lack of correlation between pulmonary symptoms and pulmonary signs.
Divers generally participated in only one phase of the study. This opens the possibility that the apparent effects of the dive conditions were, in fact, individual responses; some divers may be more aware of respiratory symptoms than others. Here, the fact that divers did not participate in all phases of the study reduces the ability to separate effects of conditions from individual succeptibility. However, as differences are usually greater, not smaller, in a repeated design, we have confidence that the differences seen betweem conditions were real. Divers knew what

Occurrences of signs and symptoms of pulmonary oxygen toxicity
The incidences of reported symptoms following the wet resting, six-hour oxygen exposures of these dive weeks were apparently lower than those seen previously following similar dives, while the incidences of ΔFV were greater. Previous single, six-hour resting dives with PO 2 of 130 kPa showed incidences for symptoms of 33% (95% confidence interval 20% -46%) after single dives (n = 44); 34% (18%-54%) after two dives with an 18-hour SI (n = 29); and 44%, 50%, and 31% (n = 16) after three, four, and five resting six-hour dives with 18-hr SI, closer to those from the current dry O 2 exposures. The concurrent incidences of ΔFV were 6% (1%-16%) after one dive; 3% (0.1%-18%) after two dives; and 25%, 6%, and 6% after three, four, and five dives, respectively [2]. Given the relatively small numbers of participants and the low incidences in the current dives, binomial confidence intervals are large, but the apparent difference between current and previous wet resting oxygen dive series (and between the incidence after three dives and the rest of the series) remains a puzzle. The differences between wet and dry oxygen dives in the current series was surprising, since an earlier study showed no difference in the incidence of signs or symptoms of pulmonary oxygen toxicity between single wet and single dry six-hour dives with a PO 2 of approximately 160 kPa, where 36% of 34 divers in both groups reported symptoms [3]. However, that study involved only one day of diving, and the current results also show no difference between the first dive of the dry and wet resting oxygen dives series. For days 2 through 5, the breathing apparatus is unlikely to have caused the difference; the slight threshold negative pressure needed to start the gas flow and the hydrostatic load of the seated posture underwater might be expected to have increased, not decreased, pulmonary injury relative to a hood with a very slight positive pressure surrounding the head. (Both devices had a slight positive pressure threshold for exhalation.) For both situations, the oxygen supply was humidified by bubbling it through water before it was delivered to the divers, although the bubblers were different (home built for the wet dives, commercial bubblers made for use with the hoods for the dry dives), and perhaps the gas delivered in the water was also humidified by water leakage into the masks. However, the difference is not that the dry dives had a higher incidence of pulmonary oxygen toxicity than that expected, but more that the wet dives had a lower incidence than had been reported earlier.
Submersion itself seems an unlikely source of protection against pulmonary oxygen toxicity. Acutely, the centralization of blood volume leads to higher pulmonary blood volume and flow and higher pulmonary capillary pressure than that in a dry environment, all factors which would promote fluid transfer into interstitial spaces through any leaky capillary walls; alveolarcapillary leaks have been detected after 17 hours of dry normobaric hyperoxia [14]. It seems reasonable to assume that a small translocated volume can generate local interstitial edema sufficient to irritate nerves without affecting overall pulmonary function parameters. However, the hypovolemia that follows from immersion [15,16,17] may have moderated the increased filtration pressures in the lungs in the wet divers despite the sustained increase in central venous pressure [18]. Hypovolemia was likely cumulative across the dive week [15], and perhaps more severe than that during the previous series of six-hour wet resting oxygen dives [2] when divers were less restricted. The sympathetic responses to hypovolemia [15,19,20] may have reduced some of the immediate post-dive symptoms; for example, terbutaline has been reported to alleviate the impaired mucociliarly clearance seen after six-hour normobaric hyperoxic exposures [21].

Parameters of forced flow-Volume maneuvers: changes overall
None of the measured parameters changed from baseline immediately after the first dive of any of the series. However, FEV 1 and FEF  values about 17 hours after surfacing from the second and subsequent in-water dives were lower than those immediately post surfacing. Values immediately after surfacing from exercise dives were higher than those immediately after surfacing from resting dives. Finally, values after immersed dives with oxygen were lower than those after dives with air; in-water diving with oxygen generally depressed FEV 1 and FEF  , and the effect persisted through the follow-up day three days after the fifth dive.
Depressed values at follow-up indicate that the values not different from or above baseline immediately post dive do not represent recovery, but rather a phenomenon of immersion or immersion with exercise. The pattern is suggestive of an inflammatory effect that begins from the morning after the first oxygen dives until several days after the last oxygen dives, but with intervening immersion (and sometimes exercise) effects that somehow increased these indices of pulmonary function. The pattern of increase without inflammation is seen for WetAirX.
Both FEV 1 and FEF  provide information about small airway resistance and lung elastic recoil. FEV 1 is usually preferred in clinical practice because it is the more reproducible of the two parameters. The primary difference is that FEV 1 spans a wider and, across subjects, a more variable range of lung volumes than does FEF  . FEV 1 always includes information from the first 25% of the volume expired, flow dominated by resistance of the large airways. When FEV 1 /FVC is less than 75%, FEV 1 does not include the entire middle half of expiration. Thus, when FEV 1 /FVC is low, FEV 1 may appear normal while FEF  shows increased small airway resistance [22]. Of these dives only the wet air exercise cohort included a large number of divers with FEV 1 /FVC ratio less than 75%, and in that cohort, FEV 1 did not show as many changes related to diving as did FEF  .
Diving-related changes in FEV 1 and FEF 25-75 must be functional, not structural changes; they have rapid onset and are self-reversing. Possible sources of altered small airway resistance are changes in smooth muscle tone, degree of bronchiolar mucosal edema, airway fluid cuffing, or mucus obstruction. Lung elastic recoil could be altered reversibly by capillary engorgement or surfactant degradation.
The deficits in FEV 1 and FEF 25-75 on the mornings after diving but not at the initial postdive assessment strongly suggest a post-dive inflammatory cause. Oxygen-induced injury to the airway epithelium may have caused later mucosal swelling and even smooth muscle activation. The inflammation seems to have required immersion during the oxygen exposure. This is consistent with the argument advanced above, that high PO 2 causes capillaries to become leaky, and that the leaks become worse when a hemodynamic pressure head forces fluid through them to cause mild interstitial and perivascular edema and fluid cuffing around bronchioles and bronchial junctions. Mucus accumulation in the airways seems unlikely because cough was not a common symptom.
Restoration or increase in FEV 1 and FEF  immediately after both air and oxygen exposures in the water, enhanced by exercise, suggests a cause that is independent of any oxygeninduced injury. Water immersion causes lung vascular engorgement, but that has been shown to decrease lung elastic recoil at low lung volume [8], and is thus not an explanation for the increase in FEF  shortly after dives. Further, all measurements were averages of three valid tests, and the first vital capacity maneuver can be expected to expel much of the excess blood from the pulmonary vasculature. Breakdown of pulmonary surfactant would increase lung recoil pressure, and, at least in mice, surfactant turn-over takes between 3 and 6 hours [23], which makes the time course possible, but changes in surfactant are more likely after O 2 than air dives. Another possible cause for the early recovery of or increase in FEV 1 and FEF  is bronchodilation secondary to generalized increased sympathetic nervous system activity after the dives: Muscle sympathetic nervous activity increases after three days of dry immersion [19] and in response to hypovolemia [20], and six-hour water immersion induces significant hypovolemia [15]. Heart rate variability indices of sympathetic activity decrease after single sixhour resting air dives [24].
Diurnal differences may also have influenced the post-dive vs. pre-dive values. FEV 1 is lowest during the normal sleep cycle when many of our morning measurements occurred, and highest in the early afternoon when some divers were exiting the water [25]. However, the measurement schedule was very similar for air and oxygen dives, wet or dry, while the patterns we observed were not. The time of day effect was not significant for DryO 2 . Further, the first diver of a series performed pulmonary function testing easily five hours before the last diver of a set of dives, an effect that should blur the diurnal effect [26].

Rest vs. Exercise
The pulmonary effects differed considerably between exposures at rest and those with exercise. For air dives, apparent improvement in indices of small airway function immediately postdive was more pronounced after exercise dives, when values after WetAirX were greater than baseline. For O 2 dives, the overall incidence and severity (number of separate problems) of pulmonary oxygen toxicity was much greater after exercise dives than after resting dives while average changes in pulmonary function were not evident until the end of the dive week. This may be related to the further elevation by in-water exercise of pulmonary vascular pressures, already elevated by immersion at rest relative to the non-immersed state [12].
Exercising divers had a slightly lower hydrostatic respiratory load than did those resting, because they were semi-prone at a 30˚head-up angle during their exercise periods rather than vertical in the water column as they were at rest. However, the exercise-induced increase in pulmonary blood volume is much greater than that for a small change in static load [11].
When divers breathed air underwater, the incidence of pulmonary signs and symptoms after diving was not increased by exercise; presumably, if the capillaries have not been damaged by hyperoxia, they don't leak. However, when divers breathed oxygen, the incidence of pulmonary oxygen toxicity increased with exercise, possibly because increased pressures "found" more leaks. After the second or later dives the increase was more pronounced, perhaps because the inflammatory response to the first dive's oxygen insult made the tissue more fragile. This is consistent with previous studies with four hour dives with rest or exercise, where the increase in pulmonary toxicity with exercise was not clear until the second day of diving [3].

Conclusion
Repeated shallow six-hour dives with air as the breathing gas generate very few respiratory symptoms or reductions in pulmonary function even when the divers perform moderate aerobic work half the time; the pulmonary vasculature appears to be able to handle the immersion effects without injury. However, exposures to a PO 2 of approximately 130 kPa cause symptoms of pulmonary oxygen toxicity and individual changes in pulmonary function parameters whether the exposure is in a dry chamber or underwater. If the underwater oxygen exposure includes moderate aerobic exercise for half the time, both incidence and individual severity of pulmonary oxygen toxicity are exacerbated. FEV 1 and FEF  are more affected than is FVC, and both show a pattern of late (about 17-hour) post-exposure deficit followed by improvement early after the next exposure. Evidence of increased small airway resistance remains three days after the end of in-water O 2 dive weeks. For individuals who must exert themselves while breathing PO 2 of 130 kPa, a schedule of 6 hours per day for more than two days is illadvised.
To further work in this area, studies including more functional parameters should be completed. Moreover, pulmonary function should be examined in the context of more global "whole-body" oxygen toxicity that may impact diver safety and physical performance.
Supporting information S1 Table. Incidence of pulmonary oxygen toxicity (PO 2 tox) for individual subjects at rest. A. Wet resting O 2 (WetO 2 ). B. Wet resting air (WetAir) C. Dry resting O 2 (DryO 2 ). Pulmonary oxygen toxicity in the morning after a dive is ascribed to the dive preceding it. Any reports from follow-up day post dive +1 were included with Dive 5. There are 6 possible days considered: 5 dive days, and the follow-up day, post dive +3, or day 8. (PDF) S2 Table. Incidence of pulmonary oxygen toxicity for individual subjects who exercised under water. A. Wet O 2 with exercise (WetO 2 X). B. Wet air with exercise (WetAirX). Pulmonary oxygen toxicity in the morning after a dive is ascribed to the dive preceding it. Any reports from follow-up day post dive +1 were included with Dive 5. There are 6 possible days considered: 5 dive days, and the follow-up day, post dive +3, or day 8. (PDF) Writing -original draft: Barbara E. Shykoff.