Pulmonary function one and four years after a deep saturation dive

The pulmonary function of 24 Norwegian divers who had participated in a deep saturation dive to pressures of 3.1-4.6 MPa was reevaluated one and four years later. Twenty-eight divers performing ordinary saturation diving to pressures of 0.8-1.6 MPa and followed over a three-year period served as referents. A significant reduction in forced expiratory volume in 1 s (FEV1.0) of 210 (SD 84) ml (P < 0.001) occurred the first year after the dive. Thereafter the annual reduction in FEV1.0 was 28 (SD 62) ml.year-1; this value did not differ from the 35 (SD 80) ml.year-1 of the referents. The forced midexpiratory flow rate and forced expiratory flow rates at low lung volumes were also significantly reduced one year after the deep dive, and the closing volume was increased. No significant changes occurred in forced vital capacity. The results agree with those of cross-sectional studies on divers' lung function and indicate the development of airflow limitation in relation to diving exposure.

Saturation divin g is the method used to dive to such depths. The di ver s are co mpress ed to the pressur e corres pondi ng to the work ing depth in a hyperbaric chamber complex, where they live at incre ased pressure for periods of up to four weeks . They are tran sferr ed to the in-water worksite in a diving bell. There can be as many as 16 divers liv ing in the hyperbaric chamber complex at the same time. With thi s di ving method all tis sues are saturated with the inert gas of the atmosphere, and therefore the avail abl e bottom time is not limited as in surface-oriented diving or boun ce divin g. Th e rate of decompression is, however, limited by the rate of inert ga s washout from the tis sues. If decompression sickness is to be avoided, the mean rate of decompression mu st be 10-1 2.5 kPa . h' or 0.24-0.30 MPa . d-' . The inert gas of the atmosphe re is usuall y helium to reduce the den sit y of the breath ing gas and ther eby the work of breathing. The parti al pressure of oxygen is raised to 35-40 kPa during the isop ression phase of the dive, to ensure oxy genation, and to 50-60 kPa during the decompress ion pha se, to facilitate inert gas elimination.
, Norwegian Underwater Technology Centre A/S, Ytre Laksevag, Norway. Cro ss-section al studies of div ers' lung fun ct ion indicate the development of airflow limitation in relation to divin g exposure (1-4). Their lung function is, ho we ver , also influenced by the effects of adaptation to the hyperbaric envi ronme nt with increase d gas den sit y, which wo uld result in an increase in forced vital capaci ty (FVC) but not in forced expir atory volume in I s (FEVl.o)' givin g a low FEVl.o: FVC ratio (5)(6)(7). Th e selection of subjects as diver s could also result in a group of very fit subje cts with these lung function characteristics, as there is a negati ve relationship between FVC and the FEVl.o: FVC ratio (1, 8).
To study the co urse of lung fun cti on changes in pr ofess ional satura tion divers, a longitudinal followup of pulm ona ry funct ion was made of a group of profes sional Norwegian diver s who had parti cipated in a deep experiment al saturation dive to pressures of 3.1-4.6 MPa. Thi s group of deep div ers was compared with a gro up of divers performing ordinary sat uration diving in the North Sea to pre ssures of 0.8-1.6 MP a.

Subjects and methods
Forty-three profes sional saturation divers participated in eight different deep saturation dives in Norway in 1983-1 986. Th e maximal pressures we re 3.1-4.6 MPa, equ ivalent to sea water depth s of 300-450 m, and there were four to nine divers participating in each dive. The cha racter istics of the dives and the acute effects of the dives on pulmon ary func tion have been descr ibed elsewhere (9, 10). Of the 43 divers, 24 were Norwegians. The Norwegian divers were all reexam ined one year after the deep dive, and 22 of Scand J Work Envir on Health 1993, vol 19, no 2 them were available for another examination 4.0 (SD 1.0) years after the deep dive. Their base-line anthropometric data, smoking habits, and diving experience are shown in table 1. Ex-smoker s were classified as nonsmokers if they had stopped smoking more than one year earlier; otherwis e they were considered smokers. Three divers had participated in a deep dive to pressures of 3.1 and 5.1 MPa one and four years before entry into this study.
The deep divers were compared with 28 professional saturation divers who did not participate in deep dives (reference divers), but who perform ed ordina ry saturation diving in the North Sea to pressures of 0.8-1.6 MPa. They were divers who had attended the Norwegian Underwater Technology Centre for their annual medical certification and who had at least two examinations one year or more apart during the period 1983-1 990. The mean observation time was 3.0 (SD 0.9) years. Their base-line anthropometric data, smoking habit s, and diving experience are shown in table 1. This group of divers was not significantly differ ent from the group of deep divers with respect to previous diving experience and base-line pulmonary function. Both groups had a reduced forced midexpiratory flow rate (FEF 25 _ 75 ) and reduced maximal flow rates at low lung volumes (FEF 15 _ 85' FEF so ' FEF 1S ) ' consistent with the characteristics of divers' lung function in general (1--4).
There were no significant differences in age, height, weight, smoking habits (table I), or in initial pulmonary function (table 2) between the deep divers and the referent s at the time of entry into the study. The control divers were all, except two, active in operation al diving in the North Sea throughout the observation period, and their median time in saturation dives was 38 (range 12-64) d each year at maximal pressures of 0.8-1.6 MPa. The divers who were not active had advanced to diving supervisors during the observation period. Three of the deep divers had stopped diving after the deep dive for reasons not related to their pulmon ary function , and three had advanced to diving supervisors during the follow-up period. The others were still active in operational diving in the North Sea, and their median time in saturation dives after the deep dive was 27 (range 1~0) d each year at maximal pressures of 0.8-1 .6 MPa. Regular diving in the North Sea was resumed three to six months after the deep dive. The diving exposures, measured as cumulative hyperoxic exposure, cumulative hyperbaric exposure, days in saturation, and days in decompression, were all interrelated and correlated significantly with the maximal pressure to which they were exposed (correlation coefficient 0.7-0.9, P<O.OI).
On every occasion the subjects were given a structured interview registering pulmonary symptoms and diving exposure. Diving exposure was registered as the number of saturation dives between the observations with the registration of the pressure and duration of the dives partitioned into days in compression, isopression, and decompression. In the operational diving in the North Sea, as well as in the deep saturation dives, modified saturation decompression tables of the United States Navy were used with rates of decompression of 0.24--0.30 MPa . d' , and partial pressures of oxygen of 40 kPa during the isopression and 50 kPa during the decompression phases. The cumulative hyperoxic and cumulative hyperbaric exposures were calculated for each diver during the observation period . There were no differences between the deep divers and the referents in their reporting of respiratory infections or respiratory symptoms. Abnormal breathlessness on exertion or nonproductive coughing in the week following an ordinary saturation dive was reported by seven of the deep divers and ten of the reference divers, but no persistent pulmonary symptoms were reported in either group. The dive rs ' smoking habits did not change, and there were no cases of decompression sickness during the observation period.
Dynamic lung volumes and flows were measured by at least three satisfactory forced expiratory maneuvers from the total lung capacity in the sitting position. The FVC, FEVl. O ' and peak expiratory flow rate (PEF) were taken as the highest readings obtained. FEF z5 • 75 ' FEF 75 _ 85 , FEF 50 , and FEF 75 were taken as the highest readings from flow-volume loops not differing by more than 5% from the highest FVC (11) . Transfer factor for carbon monoxide (TIeD) was measured by the single-breath holding method (11). Effective alveolar volume (VA) was measured simultaneously by helium dilution, and the transfer of carbon monoxide per unit of alveolar volume (K eo) was calculated. Static lung volumes were measured with the multibreath nitrogen washout technique from the functional residual capacity (FRC). In combination with the measurements of expiratory reserve volume (ERV) and inspiratory vital capacity (IVC), the total lung capacity (TLC) and residual volume (RV) were calculated. The closing volume (CV) and the slope of phase III of the single-breath oxygen test (L\-N z ) were also measured (12). All of the measurement s were taken on a 1000 IV Computerized Pulmonary Function Laboratory (Gould Inc, Dayton, Ohio, United States ). Volume and test gas calibrations were done before each test with the instrument's automatic dynamic calibrat ion procedure , and the volume calibration was verified with a calibrated syringe both before and after each test (Gould calibration syringe 3.00 I, Gould Inc). The same instrument and calibration syringe were always used. The results were corrected to conditions of body temperature and pressure saturated with water vapor (BTPS ).
The measurements were taken in the morning at least 2 h after breakfast without tea or coffe e and with no smoking for 2 h before the measurements. At the time of the examination at least four weeks had elapsed since the last saturation dive . The dynamic lung volumes and flow s and TieDof all of the divers were measured on every occasion. Static lung volumes, CV and L\-N z were measured for 18 deep divers only, before and one year after the deep dive. For FVC, FEVl.O' FEF z5 _ 75 ' and Tl co' the predicted values of Gulsvik (13) and Guls vik et al (14) were used ; these value s are based on cros s-sectional surveys of asymptomatic subjects in a general population. For FEF 75 _ 85 , FEF 50 , and FEF 75 the predi cted values were taken from our previous study of divers ' lung function, which included a matched reference group (4), as updated predicted values for the general Scandinavian population are not available. All of the subjects gave their informed consent, and the protocol for the medical and physiological monitoring of the deep dives had been approved by the Regional Ethical Review Committee.
All of the results are given as means and standard deviations or medians and ranges . The individual rate of change of the lung function variables was calculated as the annual change between observations. The two-tailed t-test was used for comparing the deep divers with the referents and for comparing the determined values with the predicted values. A Pvvalue of less than 0.05 was considered to be significant, but, when multiple comparisons were made, the Bonferroni correction method was used for adjusting the significance level.

The mean annual reduction in FEVl, o in all divers
(N = 52) over the whole observation period was 82 (SO 61) mi· year", which was significantly higher than the predicted annual loss of 34 ml -yearI (P<O .OOI ) (13). There were no significant changes in FVC over the observation period. The control divers had a mean annual reduction in FEVl,o of 35 (SO 80) ml . year" which was not different from the predicted value. The deep divers, on the other hand , showed a reduction of 210 (SO 84) ml (P < O.OOI) in FEVl.o in the first year after the deep dive. Thereafter the annual reduction in FEVl,o was 28 (SO 62) ml . yearI , which was not significantly different from the corresponding reduction of the reference divers or from the predicted value (figure 1). The forced expiratory flow rates at given fractions of the FVC expired showed changes similar to those of the FEVl.O ' with a significant reduction in the first year after the deep dive (table 3). In the referents the rate of change in FEF 75 was larger than predicted (P < 0.02), whereas in the deep divers one year after the deep dive and thereafter the rates of change in the maximal expiratory flow rates did not differ from those of the reference divers or the predicted values. The PEF did not show any significant changes.
The deep divers had a reduction of 0.34 (SO 0.71) mmol . min-I . kPa-! in TieDin the first year after the deep dive, and thereafter the annual reduction in Tim was 0.19 (SO 0.87) mmol · mini . kPa-' . year" , which was not significantly different from the 0.10 (SO 0.78) mmol . min-I. kPa-1 • year! of the reference divers or the predicted annual loss of 0.06 mmol . min" . kPa-1 • year-'. There was no significant change in VA in the referents or the deep divers, the changes in K co therefore being similar to the changes in Tl co (table 3). There was a small increase in CV (P< 0.01) one year after the deep dive. TLC, FRC, RV, and L\-N z did not show any significant changes (table 4).

Discussion
The results showed that in the first year after the deep dive, there were significa nt reductions in FEVl. o and forced expiratory flow rates that did not recover over the following three years of observation. There were no signifi cant differences between the deep divers and reference divers at the time of entr y into the study, and there were no obvious respiratory infections or other exception al activities dur ing the observation period that could explain the differences between the two group s of divers. Except for the deep dive, there were no significant differenc es in the diving exposure of the two groups during the observation period s.
The annual redu ction in FEVl. o in the deep divers one year after the deep dive and thereafter did not differ from that of saturation divers working in the North Sea or from the predicted annual reducti on. The predicted annual reduction in FEVl. o was, however, based on the results of a cross-sectional study.
The majority of such studies report an annual reduction of 25-35 ml year" in FEVl. o (11) on the basis of linear regression models of cross-sectional data of healthy popul ations 20-70 years of age. Because of the age cohort effect (15), and because the decline in FEVl.omay not start until the age of 25-30 years, the predicted values may have been overestimated in the age range of our dive rs. There was no significant change in FVC over the observati on period. It has been suggested in earlier studies that adaptat ion to the hyperbaric environment with its increased gas density and thereby increased work of breathing may result in an increase in FVC (5)(6)(7). This phenom enon could explain the relatively high FVC of the divers and the disproportionat e elev ation of the FEV u)" For the relatively well-exper ienced divers in this study , no increase in FVC was demonstrated, neither immediately after the deep experimental dive  lung fun ction variables to more specific exposure factors than maximal pre ssure, which wa s used as the determinant differentiating between the two g ro ups of divers.  (9 ) nor at the end of the follow-up period. The crosssectional study in which these divers were included did not show any correlation between FVC and diving e xpos ure, and there were no differences in the static lung volumes between the divers and a matched reference group (4). The results are therefore consis tent with the development of airflow limitation in divers, and the characteristics of their lung function cannot be explained by adaptation or se lectio n. The larger-than-predicted rate of change in FEF 75 among the reference divers and the increase in CV in the deep di vers are al so in agreement with the development of airflow limitation.
In contrast to our previous cross-sectional study of div ers' lun g fun ction (4) and the studies of pulmonary fun ction immediately after deep dives (9, 10), there wa s no significant reduction in Tl co in this investigation. The mean reduction in Tl co immediately after the deep saturation dives wa s 10.6% or 1.24 (SD 0.87) mmol . min-I . kP a-1 (P < O.OOI), and 0.39 (SD 0 .78 ) mmol· min-I . kPa-' (P < 0 .05) one month later. There were, on the other hand, no changes in the FEV 1.0 or maximal expiratory flow rates immediately after or one month after the deep di ves. The changes in dynamic lung volumes and flows mu st therefore have taken place later than one month after the deep div e and may therefore not be rel ated to the deep dive at all. There was, however, a significant increase in TLC and FRC immed iate ly after the deep di ve s, and the increase could ha ve compensated for a real reduction in the maximal expiratory flow rates since these measurements are related to the fraction of FVC expired and not to ab solute lung volume.
Several factors may be in vol ved in th e development of airflo w limitation in divers. In saturation diving , the partial pressure of ox ygen is usually e levated to 40-60 kPa to en sure oxyg enation and facilitate inert gas elimination during decompression. Oxygen ha s well-established acute to xic effects on the lun gs ( 16) , and in an animal study by Ril ey e t al (17) it wa s shown that four to six weeks afte r exposure to hyperoxia loss of pulmonary elastic tissue and reduced pulmonary recoil pre ssure occur. Th ese findings are consistent with the development of an emphysematou s lesion. Venous ga s mi croemboli generated during decompression are common, e ve n in the absence of clinical de compression sickness (18), and can induce pulmonary inflammatory lesion s and gas exchange abno rmalities (19,20). This effect ca n be pot entiated by hyp eroxia since oxygen radi cals are in vol ved in the inflammatory reaction s induced by gas microembolism ( 19) . Th e exposure to fact ors such as cumulative hyperoxic exposure, cumulative hyperbaric exposure , time in saturatio n, and tim e in de compression we re all inte rrelated and co rre lated with the maximal pre ssure to which the deep divers were e xpos ed during the ob servation period . We we re therefore not abl e to ascribe the c hanges in the