Research and Technology Work Unit Summary, Entitled: Scientifically Based Decompression Tables for Air and Mixed Gas Diving (Redacted)

Page 1

7ESEARCH AND TECHNOLOGY WORK UNIT SUMMARY SUM ' RY SUMMARY SCTY =CUR ITY 861001 ID. CHANGE IUII cx I 0. NO.ICOOES IPROGRAMELEMENT I PROJECT NUMBER IT ASK AREA NUMSER I WORK UNIT NUMBER 06 19 Stress physiology; 06 02 Eioengineering; 06 06 Environmental biolom 3. START OATE 114. ESTIMATED COMPLETION OATE115. FUNDING ORGANIZATION116,PERFORMANCE MET HOO DATE EFFECTIVE IE XP I RATION Manual were last revised over 30 years ago using a variation of a mathematical method applied in 1908. The calculation has grown in complexity but not in understanding or experience will be produced to replace presently unsatisfactory schedules. , 24. APPROACH. (U) Physiologically based models with a tractable number of parameters 'i (less than 6) will be constructed and fitted to a data base of human decompression replaced b y experimental data on rat (Rattus norvegicus) population response to decompression, measured values of gas solubility in dog (Canis familiaris) tissue, and physiological responses of guinea pigs (Cavia porcellus) to. decompression. Limited experimental decompression of military and civilian government personnel will be guided by with results suggesting that 6-hour exposures at 60 fsw are safe without decompression. Thirty-two man-dives for 5 days at 20 tsw on air followed by no time for decompression produced no cases of bends, a finding relevant to submarine rescue. gas-switching experiments with rats were completed. relative gas potencies and gas exchange time constants of N2 and He and suggested a transient detrimental effect or switching from air to He02 and a transient beneticial effect for the reverse switch. Experiments comparing air versus He02 recompression treatment at various depths following air dives with guinea pigs were completed. A positive linear relationship was observed for air between increasing recompression depth and better recovery; with He02 treatment, recovery appeared less dependent on recompression depth and less effective. rublications: 6 (U) The air decompression schedules of the U.S. Navy Diving Revised schedules based on plausible physiology and well documented Parameters and algorithms will be empirical initially, and then will be Results agreed with previous work on 1498 SNO~OZ-CPOILOJOI EO lTl ON OF MAR 68 I S OBSOLETE

Page 2

D. Scientifically based decompression tables for air and mixed gas Work Unit Number 1002 1. OBJECTIVES AND REQUIREMENTS: Develop safe and efficient human decompression procedures to avoid decompression sickness (DCS) in air, mixed gas, and saturation diving by: a. Development of a comprehensive, predictive model of the population incidence of decompression sickness in rats as a function of the inspired inert gas and oxygen composition and the pressure profile. b. Analysis of human dives and decompression outcome by models similar to (a.) and estimation with maximum likelihood. c. Measurement of inert gas exchange with radiotracers in men. d. Comparison of recompression treatment techniques utilizing different gas mixtures and pressure profiles in guinea pigs. e. Development of a table of canine organ inert gas solubilities which can be used in developing the model outlined in (a) and in human decompression table development. f. Human trials of proposed new procedures. This work responds to requirements identified in the CNO Deep Sea Floor Mission Requirement Study (C), December 1982, and the Long Range Technology Needs for Ocegn Engineering Missions (C), March 1984. Except for very modest exposures. all undewater or pressure exposures require a gradual reduction of pressure or the diver will suffer decompression sickness (DCS). Despite extensive empirical modification and testing of decompression tables, DCS remains the major disability among Navy divers in new or unusual operational settings. decompression schedules appear to be reasonably safen(1ess than 1% incidence of DCS) for the most common operational dives. experimental dives appear to be unsafe (more than 10% DCS), or safe but too long. Furthermore, current decompression schedules do not consider variations in underwater workloads before and during ascent, and do not optimize use of oxygen. Currently promulgated decompression tables were obtained by several different calculation procedures. or algorithms. Their successful use can be attributed to laborious human testing and modification, rather than to choice of the biologically appropriate algorithm. improve currently published tables, to extend the useful diving range, and to allow maximum efficiency by providing for diver-worn decompression computers. Infrequently used and A new algorithm is needed to

Page 3

a. Predictive model of decompression sickness in rats. Despite many years of operational experience diving with nitrogen, helium, and oxygen, many questions remain regarding the role of these gases in DCS (1,2,7). Specific questions concern use of mixtures of these gases and gas switching procedures. minimized because of reduced DCS potency of particular mixtures or reduced total inert gas tension brought about by gas sequencing. In addition, the validity of the equivalent air depth premise and the importance of O2 in decompression outcome remain unclear. Decompression stress could be Small rodents have been used extensively to examine decompression problems. Large numbers of these animals can be used relatively cheaply and easily compared with animals such as dogs. advantage when dealing with a phenomenon such as DCS in which the outcome is a binomial event (i.e. the outcome is recorded as one of two events, bends or no bends). This is because the results of repeated dives are subject to the constraints of binomial sampling which result in large confidence limits, or errors, associated with the observed incidence values of DCS. To avoid this severe limitation, maximum likelihood (a curve-fitting method that is ideally suited to analyzing binomial data) (4) can be used to express the probability of bends as a continuous function. to be important in affecting decompression outcome such as gas composition and likelihood, rather than a large number of incidence values as in binomial analysis, a greater degree of confidence can be obtained in the estimated be used to replace initial empirical parameters and algorithms for human This has a great This function includes variables thought By estimating a small number of parameters using maximum These models and parameters experimentally derived from rat data will b. Human decompression analysis. Inefficient use of previous human data has also been addressed in recent work. A class of mathematical models is used which predict the probability of a case of DCS from the known exposure profile. are being used in other areas of experimental biology such as carcinogenesis and environmental toxicology. a large number (hundreds or thousands) of actual known outcomes by a adjusted until the likelihood is maximized, or the agreement is as good as possible. Statistical techniques are available to compare how well different models agree with the data, to check whether data sources are comparable, and to estimate uncertainty in any predictions. probabilistic models with 1700 known human air dives using maximum likelihood, excellent agreement was found between predicted and observed DCS incidence (17). Using models from that exercise, new sets of standard air decompression tables were prepared with projected incidences of 1X and 5% (111, the safety of current U.S., British, and Canadian air tables were compared (16), and the work was extended to air and N2-02 saturation decompression (5). applications, the dives are frequently to multiple depths and use the Mk-15 UBA which maintains a constant PO of 0.7 ATA. The predicted probability of DCS is compared to ; The constants or parameters of the model are After *"calibrating" some simple Most of those model applications were to single dives using a For special warfare Other types of diving use the

Page 4

techniques of in-water decompression and surface decompression, which have aot yet been examined with these models. Those models used gas exchange kinetics suggested by the xenon Presumably better agreement would be found if kinetics of nitrogendog study. in man were available. Reasonable model extension into varying O2 during the dive and decompression is possible due to confidence gained by a large series of human test dives just completed (10). Exposure data for 0.7 ATA O2 dives, repetitive dives, decompression using high oxygen, and surface decompression are available in recent experiments both at NEDU and DCIEM in Canada. Furthermore, the models all assumed that inspired c. Human gas exchange. Gas exchange is a pregnant topic because all views of DCS view the original insult as an excess of inert gas somewhere in the body. Since the 1800's, the most comtnon view of gas exchange holds that the body is composed of several parallel single exponential compartments. A number of theoretical approaches to model plausible anatomy and mass transfer all predict more complex kinetics. In a study of many thousands areas of dogs breathing the inert trace gas xenon, no areas were found to exchange as single exponential compartments (14). Direct measurement of kinetics in the most relevant species (man) and using the most relevant gas (nitrogen) are not easy to perform. Since 1980, a collaborative project between NMRI and the Lawrence livennore National Laboratory (LLNL) has seen the development of techniques of N tracer studies in human divers. Striking results have been observed in the human knee which are not predicted by any current gas exchange theory ( 1 5 ) . Technical problems with contaminated radioactive gas has caused suspension of those experiments. Assuming that problem is resolved, the work can continue as a NSMRL-NMRI-LLNL collaboration. d. Recompression techniques in guinea pigs. The few studies documenting cardiorespiratory responses to DCS in anesthetized animals suggest that a series of physiological changes often occurs during development of the disease. events has been observed to occur in unanesthetized guinea pigs that develop DCS and has been used to develop a cardiorespiratory model to examine the response of a sick animal to different recompression techniques. This model will allow the investigation of several important decompression treatment questions. First, are there counterdiffusion (9) problems when recompression with air or He-02 is performed following air or He-0 Second, can O2 mixtures be varied to optimize recovery from the beds. These recompression results will enable formulation of a more unified human decompression model by incorporating algorithms and parameters based on experimental recompression. Likewise, a predictable pattern of The guinea pig model will also lend itself well to addressing an additional question. What is the composition of intravascular bubbles following decompression and how is It affected by the gas composition of the development of DCS and its treatment. This knowledge would be valuable in understanding both

Page 5

e. Inert gas solubilities. In actual computation of decepression requirements, gas quantities such as diffusion and solubility coefficients are required. However, little data currently exist on the quantity and kinetics of inert gases at the tissue level under hyperbaric conditions. are made regarding gas behavior in tissue, and these are used in developing When a gas quantity is required in the computation of decompression requirements, Henry's law (the amount of gas dissolved in any fluid is equal to the gas phase partial pressure times a constant) is assumed. Error in this assumption could lead to incorrect predictions regarding decompression, especially if a mixture of more than one inert gas is being considered. Since Henry constants are available in only a few cases, most often reference is made to solubility in water and olive oil, any tissue is then presumed to be a linear combination of these solvents. usefulness of water-oil partition data is open to serious question. years ago, a review on solubility of inert gases in biological tissues re-emphasized the need for experimental data on solubility and diffusion above Because human decompression models are notoriously unreliable, even the scientifically based approach used here is not sufficient to ensure safety. Some direct human experimentation is required, both to fill gaps in available data, and to limit the potential of seriously unsafe procedures going to the Fleet. Because DCS is almost universally thought to result from gas bubbles in the diver's body, most algorithms mathematically track gas transport and bubble behavior. proposed, but none have been experimentally verified or combined into a mathematically consistent and biologically plausible model. In the overall 1, NMRDC approach, elements of this complex physiological process will be isolated into component chemical and physiological studies, and then reassembled and systematically tested. In the present work unit, basic measurements of gas solubility ("How much inert gas can a tissue hold?"), animal population response ("If the amount of decompression is doubled, is the frequency of symptoms doubled?"), and gas diffusi'vity ("How fast can gas move inside a tissue?") will be examined. Mathematical models that integrate this basic information and evaluate the mathematical parameters with measurements on human divers will continue to be developed. models will then be used to predict the safety of specific decompression profiles that will be tested with divers. This will lead to operational schedules and even diver-worn decompression computers that are optimally safe specific problems, which due to number required, the nature of decompression stress, or the need for invasive techniques, could not othewise be resolved. Inert gas-oxygen combinations that may optimize decompression in man will be studied first with rats. Combinations found promising will be considered for Many mechanisms of bubble behavior have been The candidate mathematical Animal studies (see a and d) will permit examination of

Page 6

testing on humans. in guinea pigs to examine differences in DCS treatment techniques. Cardiovascular and respiratory responses will be monitored a. Predictive model of decompression sickness in rats. Rats will be used to develop a comprehensive, predictive model of the population incidence of decompression sickness as a function of the inspired inert gas and oxygen composition and the pressure profile. Severe cases of decompression sickness are clearly manifested in these animals and provide a simple way to quantitate outcome to decompression following specific dives. Manipulation will be performed of the following variables: N2, Ar), and pressure profile. decompression will be employed to assist in parameter estimation and to determine the value of these techniques in operational diving. partial pressure of inert gases (He, Gas switching at depth and during Decompression outcome data from such experiments will be used to construct mathematical models using maximum likelihood techniques. Data from a set of sufficiently broad experimental designs which include adequate numbers of observations will allow one unified model for predicting decompression outcome in rats to be formulated based on the pressure history of the animal. Statistical models used successfully in predicting the safety of standard air dives will be used to analyze dives with other ratios of oxygen and nitrogen. Repetitive and multilevel dives and dives followed by surface decompression will also be examined. leverage of each recorded case of DCS, the natural ability of the models to predict the times of symptom occurrence will be invoked. has never been used in decompression analysis, there is a substantial probability of model failure, which will lead to further insight into human decompression and revision of the models. of computationally simple models will be explored in case of a fast algorithm for real time diver-carried calculation is required. To extend the statistical Since that approach At the same time, a different set As a final analytical task, Canadian scoring from extensive double-blind recordings of a Doppler ultrasonic device (3) in the recent NMRI experiment will be analyzed. become controversial and no other sizeable clinical trial …