Oxygen Bends

K. W. DONALD? From the Admiralty Experimeqtal Diving Unit, H.M.S. Vernon, Portswouth, England

T IS REMARKABLE that there is no serious The experiments so far discussedwere under mention of the possibility of bubble for- conditions little related to those encountered I mation in the body after exposure to in- in diving and submarine escape.However, in creased tensions of oxygen in any of the modern the report of the Royal Naval Deep Diving reviews or articles on decompression sickness. Committee in 1933 (4), a number of instances Paul Bert, and many workers since, assumed of severe paralyses in goats after simulated that, even when there is considerable super- submarine escapesfrom pressures correspond- saturation of the body with oxygen, the excess ing to 200 feet of seawater were reported. oxygen will be so rapidly used by the tissue These animals were breathing 70% oxygen for metabolism that there is no danger of bubble 20 minutes at an average depth of 135 feet, formation on return to air breathing at atmos- and for another 5 minutes at 200 feet; they pheric pressure. Yet perusal of the early lit- were then brought to atmospheric pressure in erature shows that not all investigators were IOO seconds. In view of the relatively small of the same opinion. Leonard Hill (I) de- percentage of nitrogen present, these paralyses scribed bubbles of oxygen in the cortex of mice were presumed to be due to oxygen bubbles. after exposure to 8 atm. of oxygen. These mice Recompressioncured only a very few of these convulsed for a long time after decompression. goats, and in no case was bends (flexion of A number that were not killed for examination leg due to pain in joint) seen. Lennox (5) re- recovered completely without therapeutic re- ported a possible instance of mild decompres- compression. Hill also demonstrated oxygen sion sickness in a human subject after breath- bubbles in the heart and many other organs ing 2-3 atm. of oxygen, consisting of bend-like of toads, rats and guinea pigs, following rapid pains in the shoulder which passed off in a decompression from between IO and 20 atm. short period. Stadie (6) ventured to state that of oxygen. Finally, Hill pointed out “that in decompression from 2 to 5 atm. of oxygen only one dog out of many experiments did would not cause the formation of bubbles. Bert record ‘s’agitant demi-convulsivement’ In the author’s personal experience, divers while under compression. In all other cases have surfaced immediately after an exposure the convulsions only came on after rapid cle- of 2 hours to oxygen at 2.5 atm. absolute, and compression. Rapid decompressionfrom air, after 30 minutes exposure to oxygen at 3.4 on the other hand, did not produce convulsions, atm. absolute, without symptoms. In the case but embarrassed respiration, paralysis or of self-contained oxygen-nitrogen mixture death.” He concluded that “the convulsions divers, who go to depths of 140 feet of sea which Bert details as occurring in dogs are water or more, the risk of decompression sick- clearly decompression results and due to the ness is assessed by calculating the depth at effervescence of oxygen gas in the central which an air diver would be breathing a similar nervous system.” Although it is now known tension of nitrogen (equivalent air depth). that sudden decompression from high tensions The rate of decompression is controlled by the of oxygen may precipitate the convulsions of ordinary ‘air tables’ according to the time at oxygen poisoning (2, 3), this is an infrequent the equivalent air depth. The risk of oxygen occurrence and Hill’s observations concerning poisoning is similarly determined by calculat- Bert’s experiments are still pertinent. ing the depth at which a diver on pure oxygen e-P would be breathing a similar tension of oxygen Received for publication November 24, 1954. (equivalent oxygen depth). As the breathing 1 Present address: Department of Medicine, Uni- versity of Birmingham, Queen Elizabeth Hospital, of oxygen at more than 2 atm. is known to be Birmingham, England. dangerous (3), divers are instructed not to go

639 640 K. W. DONALD Volume 7

below the depth where this oxygen tension is When working with goats, in high tensions of oxygen exceeded. It is assumed that oxygen need not under pressure, scrupulous care must be taken that no ‘live’ electric wires are left inside the chamber lest the be considered in relation to decompression bold and never failing search of these animals for new sickness and it is even suggested that its dietetic sensations causes a ‘short’ with sparking. presence will help in the elimination of nitro- A IOO cubic foot compression chamber was employed. gen from the tissues. During the recent war The goat was submitted to a pressure of 2.52 atm. many experimental dives, while breathing absolute (50 ft. of seawater) by means of compressed air. Pure oxygen was then run in from high pressure oxygen-nitrogen mixtures, were carried out by cylindersuntil a total pressure of 5.54 atm. absolute was Donald and J. B. S. Haldane at equivalent reached (150 ft.). The goat was now breathing 360/o oxygen depths up to 2 atm. combined with nitrogen and 64% oxygen and was exposed to the same tension of nitrogen as a diver on air at 50 feet of equivalent air depths up to IOO feet of sea sea water (2.52 atm. absolute), and to the same tension water. Routine decompression, according to of oxygen as a diver on pure oxygen at 84 feet of sea the equivalent air depth, caused no untoward water (3.56 atm. absolute). The animal remained at symptoms apart from some itching and a very this pressure for 60 minutes. It was carefully observed occasional mild bend. However, as higher ten- throughout for any signs of oxygen poisoning. A sions of oxygen were occasionally hazarded sample of the chamber gas was taken at 30 minutes and analyzed by the Haldane volumetric method. while breathing mixtures on certain urgent Carbon dioxide tensions were never in excess of 5 mm operational occasions, the author felt that the Hg. At the end of the hour, the pressure was released hypothesis that oxygen could not contribute at the rate of 1.25 ft/sec., and the animal watched to bubble formation on decompression should closely for the development of decompression sickness. Separate control experiments were carried out on be more severely tested. Further the problem each animal at similar tensions of nitrogen (air, 50 was of importance in relation to experiments feet) and oxygen (80% oxygen, I IO feet), respectively, concerning the possibility of submarine escape for the same period. An interval of at least 48 hours while breathing air or artificial oxygen-nitrogen was allowed to elapse between experiments on the mixtures (7). The time spent at high pressure same animal. under these conditions is of sufficient brevity RESULTS to allow exposure to tensions of oxygen greatly The results of a typical experiment were as in excess of the normal limit of 2 atm. follows. The goat surfaced apparently normal. It was therefore decided to carry out imme- In about 5 minutes, the animal began to de- diate decompression after exposures to nitro- velop a bend in the right foreleg. Three min- gen-oxygen mixtures containing considerably utes later, a bend appeared in the right hindleg more than 2 atm. of oxygen. and the animal began to pant vigorously, with loud rales in the chest. Three minutes later, METHOD multiple bends had developed and the goat In view of the lack of knowledge concerning the was unable to stand. Motor power was lost in subject, and the possibility of untoward symptoms, the hindlegs. The rales in the chest and respira- it was decided to employ goats, and not men, for these experiments. Goats are excellent experimental animals for the study of decompression sickness and were first employed by Boycott, Damant and Haldane in 1908 (8). These animals have since been used in a number of investigations concerning decompression sickness Twenty minutes later the goat was led away (7, 9) and the results obtained have been safely applied to the human diver. The development of a ‘bend’ in the with no bends, paralysis, or signs of any kind. leg can be watched in all its stages. Firstly, the animal To anyone associated with work on compressed will rest its weight very lightly on the affected leg. air illness, th is series of events flavored of the Soon it lifts it with increasing frequency and vigor. miraculous. Sometimes the goat paws the ground with the affected limb. When the ‘bend’ is fully developed, the leg is well In the two separate control experiments flexed, and kept clear of the ground. The animal walks which were carried out on each animal in simi- reluctantly with a severe limp. If myelitis occurs, lar tensions of air and oxygen respectively, for owing to bubbles in the spinal cord, the animal first the same period (60 min.), no animal developed shows slight unsteadiness, and is inclined to lean any signs of any kind to indicate bubble forma- against the side of the chamber. Next, the affected legs (usually rear legs) begin to straddle. In a few tion, although in the exposures to oxygen, minutes the animal is unable to stand. certain signs of oxygen poisoning were noted. May w5.5 OXYGEN BENDS 641

The results of the experiments are detailed be- hr. low. Pressures are stated throughout as feet of Lying down, in pain 111s Standing appears to have bend sea water (D). The absolute pressure in atmos- in both hindlegs 1122 pheres equals I + D/33. Greatly improved, still has slight bend in right hindleg 1128 Subj. I. 0 , 29.6 kg, dive to 150 ft. for I hr. Animal completely normal 1200 Analysis, 63.04% Oz. Equiv. 02 depth 82.3 ft. Control dive in oxygen, depth IIO ft. Equiv. air depth 52.3 ft. Analysis, 85.3y0 02. Equiv. 02 depth 90 ft. hr. At IIO ft. for I hr. No signs during exposure or on decompression At ISO ft. 1000 Restless, no convulsant signs 1040 Left 150 ft. 1100 Subj. 4. 3, 40 kg, dive to ISO ft. At surface 1101+ Analysis, 61.78% 02. Equiv. 02 depth 80 ft. Bend in right foreleg 1107 Equiv. air depth 55.4 ft. Bend in right hindleg and At 150 ft. for 25 min. convulsed severe respiratory embarrassment III0 In pain, moving continuously 1112 Subj. 5. Q , 30.8 kg, dive to 150 ft. for I hr. Sitting down, unable to stand 1112+ Analysis, 64.05% 02. Equiv. 02 depth 84 ft. Lying down, unable to stand, Equiv. air depth 50.1 ft. loud rales in chest 1114 hr. In severe pain with multiple bends 1117 At 150 ft. 1033 Lying peacefully, respirations normal 1126 Animal somnolent and ‘dopey’ towards end Animal completely normal 1150 Left 150 ft. 1133 Arrived surface 1135 Control dive in oxygen, depth IIO ft. Bends in both hindlegs 1140 Analysis, 82.275 02. Equiv. 02 depth 84.7 ft. In severe pain, restless, multiple bends II45 hr. Lying down, loud rales and panting I 148 At IIO ft. 1015 Lying down with bends in all legs 1152 Lip twitching 1100 Stood up 1157 Convulsive tremors I 108 Normal 1200 Sustained lip twitching 1114 Control dive in oxygen, depth IIO ft. Left 110 ft. 1115 Analysis, 83.3y0 02. Equiv. 02 depth 81.8 ft. Surfaced 1117 hr. No signs of decompression sickness after dive At IIO ft. 1002 Convulsant respiration 1042 Subj. 2. Q , 37.8 kg, dive to rp ft. for I hr. Eyelids ptosed, appears stupified 1047 Analysis, 63.137~ 02. Equiv. 02 depth 82.3 ft. Multiple tremors 1049 Equiv. air depth 5 I .4 ft. Appears very sleepy 1101 At 150 ft. for I hr., no signs during, Surfaced in 90 sec. Arrived surface I 1034 or after, the exposure. No signs on decompression Couttrol dive in oxygen, depth 110 ft. Analysis, 81.2% 02. Equiv. 02 depth 82.8 ft. subj. 6. Q, 36.3 kg, dive to ISO ft. for I hr. hr. Analysis, 63.04% 02. Equiv. 02 depth 82.3 ft. At IIO ft. 1408 Equiv. air depth 52.3 ft. Lip twitching 1433 hr. Left 110 ft. 1508 150 ft. 1005 At surface 1510 No symptoms at depth. Surfaced in 130 sec. No signs after decompression Arrived surface I 107 Bend in left hindleg 1110 Subj. 3. 3, 37.7 kg, dive to .rjo ft. for I hr. Bend in right foreleg 1117 Analysis, 62.087~ 02. Equiv. 02 depth 82 ft. Bends in both hindlegs and in severe pain 1125 Equiv. air depth 54.8 ft. Lying down, unable to stand, hr. restless, panting heavily 1130 At ISO ft. 1008 Lying quietly 1136 Moving uneasily 1052 Animal completely normal IIS5 Twitching of lips 1053 Control dive in oxygen, depth 1x0 ft. Sustained lip twitching 1058 Analysis, 82.27a/* 02. Equiv. 02 depth 84.7 ft. Restless, but no convulsant symptoms 1106 At depth I hr. Left 150 ft. 1108 Lip twitching after 35 min. Arrived at surface 1110 Sleepy and bemused last 20 min. Bend in left hindleg 1113 No signs after decompression 642 K. W. DONALD V&me 7

Subj. 7. 9 , 22.8 kg, dive to 150 ft. for I hr. common experience that decompressionsick- Analysis, 59.5% OZ. Equiv. 02 depth 75.9 ft. ness of this severity, due to bubbles mainly Equiv. air depth 61 ft. hr. composedof nitrogen, doesnot passoff rapidly At 150 ft. 1040 in this dramatic fashion without recompres- No symptoms of oxygen poisoning sion, owing to the lack of solubility and resist- Left 150 ft. 1140 ance to metabolism of this gas. There is no Arrived surface 1142 reason to inculpate carbon dioxide in the Bends in left foreleg and right hindleg formation of thesebubbles and it is reasonably Multiple loud r%les in chest, dyspneic 1145 Bend in left foreleg disappeared 1150 certain that oxygen was responsible. Further Increasing weakness of legs, could not stand the use of this gas in body metabolism would In pain and distress 1156 account for the remarkable relief of these No relief. Therapeutic. recompression carried out I 2063 Complete recovery animals. Control dive in 02, depth, IIO ft. This is the first time severetransient decom- Analysis, 80& OZ. Equiv. 02 depth 82.8 ft. pression sickness,which was almost certainly At depth I hr. mainly due to oxygen, has been demonstrated No signs during exposure or after decompression in conditions not far removed from those of extreme mixture dives. They were obtained in Subj. 8. Q , 45.6 kg, dive to 150 ft. for I hr. Analysis, 62.95% 02. Equiv. 02 depth 82 ft. the most favorable conditions for bubble for- Equiv. air depth 52.8 ft. mation that could be devised, i.e., in a maxi- hr. mum air tension that did not of itself cause At 150 ft. 0957 bends, combined with a maximum oxygen No symptoms at depth Left 150 ft. 1057 tension that did not cause severe oxygen Surfaced IO59 poisoning in the majority of animals. Release Bend in right foreleg 1101 from the total pressure(5.5 atm. absolute) to 1112 Bend gone I atm., in 120 seconds,was too rapid for the Bend recurred in right foreleg 111s Goat sat down, lassitude, harmless elimination of the total excess of r%les in chest and dyspneic gases, although the body could eliminate Appeared to be in pain 1118 either gas without untoward signs after sepa- Goat normal 1130 rate exposures to similar nitrogen or oxygen Control dive in oxygen, depth IIO jt. tensions. Analysis, 77.7% OZ. Equiv. 02 depth 78 ft. I hr. at depth It is of interest to consider the probable No signs during or after exposure oxygen tensionsthat occurred in theseanimals. Control dives in air carried out on all animals for I hr. The ambient oxygen tension during the expo- at 50 ft. sure was of the order of 2500 mm Hg. Lam- No signs of decompression sickness after surfacing in any animal. bertsen et al. (IO) found arterial oxygen ten- sions of 2100 mm Hg in human subjects It will be noted that goat 7 was recompressed breathing oxygen at 3.5 atm. The oxygen after 25 minutes. The animal was paralyzed, pressure in the present experiments was 0.16 in pain and, as there appeared to be no im- atm. less and the arterial ~0, was probably of provement, therapeutic recompression was the order of 2000 mm Hg. Such arterial blood carried out for humanitarian reasons.Analysis would contain 6.0 vol. % of dissolved oxygen. of the chamber gas showed an unusually high Lambertsen and co-workers (IO) have shown percentage of nitrogen, and an equivalent air that, at these tensionsof oxygen, there is con- depth of over 60 feet during the exposure. siderable cerebral vasoconstriction and that the cerebral arterio-venous oxygen content DISCUSSION difference is therefore increased so that suffi- The signs developed by these animals were cient dissolved oxygen is used to causesome of undoubtedly those of severe decompression the oxygen combined with hemoglobin to be sickness,caused by the presenceof bubbles in utilized with a fall of jugular venous oxygen the body. As these signswere only transient it tension to about 70 mm Hg. Using a modifica- is reasonableto conclude that the bubbles were tion of Barcroft’s approximate method of calcu- largely reabsorbed in this short interval. It is lation, he suggested that the mean cerebral Jfv I955 OXYGEN BENDS 643 capillary oxygen tension was probably of the ration of the tissueswith oxygen, this will be order of 850 mm Hg during his experiments. immediately reduced by metabolism. However However, the whole body cannot be simi- this will only be at the approximate mean rate larly protected by widespread vasoconstric- of 3.4 ml/min/kg, varying greatly in different tion and, as no significant change of oxygen tissues,and thus about IO minutes will be nec- uptake has been shown under theseconditions, essary to reach normal levels of tissue oxygen the only way in which the arterial-mixed tension. It would appear that the commonly venous blood oxygen content difference could accepted view that significant supersaturation be increasedwould be by a fall of cardiac out- with oxygen for any period will be prevented put. There is no evidence that such a fall by metabolism is not well founded. occurs. If these considerations are valid then the Behnke et al. (II) measuredthe arterial and absence of signs of bubble formation in the mixed venous blood p02 of anesthetized dogs control ‘high oxygen’ exposures with 80% breathing 3.8-3.9 atm. of oxygen and obtained oxygen is worthy of further investigation. mean figures (6 expers.) of arterial pOz of 2840 Nims’ contention (5) that “it is only an acci- mm Hg and mixed venous blood pOz of 630 dent of nature that nitrogen is the chief factor mm Hg (range 90-1440). Those animals with in decompression sickness” may be an over small arterio-venous oxygen content differences, simplification. and presumably a raised cardiac output, The occurrence of violent bubble formation showed, as would be expected, the highest ve- and decompressionsickness when the animals nous oxygen tensions. In all these experiments were exposed to the sameoxygen supersatura- but one, the mixed venous blood oxygen ten- tion but with added nitrogen supersaturation, sions were well above the levels in which which, by itself, caused no signs, could be ex- significant dissociation of oxy-hemoglobin plained in the following manner. It would occurs. Assuming that oxygen demands and appear very likely that, in the control air transfer are the samealong each unit length of experiments, small nitrogen bubbles causing each capillary, then the fall in oxygen tension no appreciable signs (‘silent bubbles’) oc- in most of these animals capillaries would be curred. If such bubbles also occurred in the ‘linear in space’ and the mean capillary oxygen body after the combinedexposure then, in view tensionwould be the mathematical mean of the of the very considerable supersaturation with arterial and mixed venous blood pOz, that is, oxygen and its greater solubility and apparent 1735 mm Hg* speedof diffusion, this gas would enter these Under the conditions of this experiment it is ‘silent bubbles’ in large quantities. In the case not possible to be certain of the arterio-venous of the animals which recovered so rapidly, it is oxygen content differences, the mixed venous difficult to seehow bubbles which were almost blood pOs, or the tissue tensions. However it a third of nitrogen could disappear completely would appear most likely that the mean capil- in this time. This would suggest that the ab- lary tension lay between IOOO and qoo, ex- sorption of oxygen to normal tensionsrendered cept in those animals with a very high arterio- these bubbles ‘silent’ again as in the control air venous blood oxygen content difference. The experiment. The goat which was exposed to the mean tissue tension would be but little below highest nitrogen tensions in the combined ex- this after I hour’s exposure. For the sake of posuredid not obtain relief without therapeutic simplicity, let it be assumed that all tissues recompressionand it is probable that it had have an oxygen solubility coefficient of the genuine ‘nitrogen’ bends as well as the short sameorder as blood. Then in I kg of tissue the lived exacerbation of its decompressionsickness amount of oxygen dissolved would be of the by oxygen supersaturation. After the control order of 36 ml. If the tissuecontained any fatty ‘high oxygen’ exposures with 20% nitrogen at substance, in which oxygen is approximately 4.33 atm. absolute, when ‘silent bubbles’ would five times more soluble, this figure will be a hardly be expected to occur with such a trivial considerable underestimate (I 2). When air nitrogen supersaturation, there were no signs breathing at atmospheric pressureis resumed, of bubble formation or decompressionsickness. although there will be considerable supersatu- These findings and considerations suggest a 644 H. W. DONALD Votume 7 possible new method of demonstrating the in bubble formation may have contributed to presence of ‘silent air bubbles’ that are thought this enigma. Further, oxygen supersaturation to occur after certain degrees of decompression may be a useful technique of demonstrating in air. Knowledge of their presenceis most im- the presence of the elusive ‘silent nitrogen portant in the evolution of safe decompression bubble’. tables as the occurrence of such bubbles at any stage of decompression,although causing no SUMMARY immediate signsor symptoms, will render de- Goats have been exposed to 64% oxygen compression to the next stage dangerous as and 36% nitrogen for I hour at 5.54 atm. they will increase in size and act as reservoirs absolute. On returning to air breathing at into which further gas can diffuse from the atmospheric pressure a number of these ani- supersaturated tissues. The reinforcement of mals develo.ped grave decompression sickness various degrees of symptomless nitrogen which, however, remitted rapidly without re- supersaturation with oxygen supersaturation compression.The sameanimals did not develop

------~ mav helD to solve the difficult Iproblem of the any signs of decompression sickness after ‘silent bubble’. separate control exposures to equivalent ten- The fact that theseanimals developed grave sions of oxygen (80% oxygen, 20 % nitrogen) decompression sickness while there was almost or nitrogen (in air). The danger of oxygen certainly a surfeit of oxygen dissolved in the supersaturation contributing to the occurrence tissueswould suggest that loss of function is of decompressionsickness in the presenceof a caused by other factors than immediate tissue relatively safe degree of nitrogen supersatura- anoxia due to interference with local circula- tion is discussed. tion. In conclusion it must be emphasized that The author acknowledges with pleasure the help of decompression sickness caused by oxygen Surgeon Commander W. M. Davidson, O.B.E., Royal alone has not been demonstrated. Nevertheless Navy, in these experiments which were carried out in 1945. Chief Petty Officer G. Derrick, Royal Navy, it would appear that, providing a certain degree also gave invaluable assistance. of supersaturation with nitrogen is present, the initial risk of dangerous bubble formation can REFERENCES be greatly influenced by increased tensions of I. HILL, L. Caisson Sickness. London : Arnold, 19 I 2. oxygen in the body. Although the resulting dis- 2. BEAN, J. W. Physiol. Rev. 25: I, 1945. turbances were only short lived in most of 3* DONALD, K. W. Brit. M. J. (i) 667 and 712, 1947. these animals, one continued to be severely 40 Admiralty Deep Diving and Ordinary Diving Committee. R.N. Diving Report, 1933. embarrassed. It is probable that, if the nitro- 50 FULTON, J. F. (L. F. NIMS) Decompression Sickness. gen tension during the exposure had been a Philadelphia & London : Saunders, I 95 I. little greater, not only would more animals 6. STADIE, W. C., B. C. RIGGS, AND N. HAUGEARD. have required therapeutic recompression, but Am. J. M. SC. 207: 84, 1944. DONALD, K. W., W. M. DAVIDSON, AND W. 0. some may have been fatally injured. Although 7* SHELFORD. J. Hyg. 46: 176, 1948. oxygen tensionsof the order used in these ex- 8. BOYCOTT, A. E., G. C. C. DAMANT, AND J. S. periments will never be encountered in stand- HALDANE. J. Hyg. 8: 342, 1908. ard air diving, it would be advisable to review 90 DONALD, K. W. AND W. M. DAVIDSON. Admiralty with great care the possibility of oxygen con- Experimental Diving Unit Report No. 17, Surface Decompression, I 945. tributing to bubble formation in submarine 10. LAMBERTSEN,~. J.,R.H. KOUGH, D.Y. COOPER, escape and after oxygen-nitrogen, oxygen- G. L. EMMEL, H. H. LOESCHCKE, AND C. F. helium and even air diving at considerable SCHMIDT. J. Appl. Physiol. 5: 471, 1953. depths. The precise risk of decompression II. BEHNKE, A. R., L. A. SHAW, C. W. SHILLING, R. M. THOMSON AND A. C. MESSER. Am. J. sickness, particularly after deeper dives, has Physiol. 107: 13, 1934. so far defied the most elaborate mathematical 12. LAWRENCE, J. H., W. F. LOOMIS, C. A. TOBIAS, and Dhvsical analyses and the role of oxygen AND F. H. TURPIN. J. Physiol. 105: 197, 1946.