U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content







AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window






Chapter XIX





In chlorine poisoning gassed dogs fall naturally into two chief classes: Dogs which die within three days (acute deaths) as a result of changes directly induced by the gas; and dogs which survive this acute period (survivals) and recover or later succumb to secondary factors, chiefly pneumonia. It is obvious that in the treatment of chlorine poisoning attention should be focused on the prevention or alleviation of the acute effects--the direct results of the action of the chlorine-and the success of the treatment should be judged by the total number of survivors, that is, those living through the acute period, rather than by actual recoveries, since infection and the other factors which prevent the recovery of most of the survivors are merely incidental to the gassing.

In the investigation of the physiological action of chlorine it was imperative frequently to draw samples of blood after gassing. Though the quantity of blood drawn at any one time was small, it was noted that many of the animals seemed to be favorably influenced by the experimental procedure.

Venesection, of course, has long been employed in conditions involving pulmonary edema and congestion. Recently it has been applied, with conflicting results, in the treatment of these conditions induced in man by poisonous gases in warfare. Inasmuch as the incidental observations with blood letting indicated a favorable influence upon gassed dogs it was deemed desirable to determine, under exact conditions, the effect of venesection upon the mortality of dogs receiving a lethal concentration of chlorine gas.


A series of twenty dogs, which had been gassed at a lethal concentration b by the standard method, were studied to determine the effect of venesection on chlorine poisoning. Blood to the amount of not less than 1 percent of the body weight was drawn from each animal soon after exposure. It was assumed, on account of the rapid development of serious symptoms, that it was impera- tive to treat the animals as soon as practicable after gassing. Obviously it is easier to inhibit the establishment of an abnormal state than to alleviate the condition when fully established. However, variations up to two hours in the time of treatment apparently play little role in its efficacy.

The blood was drawn, without anesthesia, by aspiration through a needle inserted through the skin into a jugular vein. This procedure was carried out easily and without discomfort to the animal. In most cases the blood was

a The data in this chapter are based on investigations made by the Medical Division, Chemical Warfare Service, at Yale University, New Haven, Conn., and published by Lieut. Col. Frank P. Underhill, C. W. S., in: The Lethal War Oases, Physiology and Experimental Treatment, Yale University Press, 1920.
b From 800 to 900 parts chlorine per million of air (2.53 to 2.85 mgm. per liter) for a period of one-half hour.


drawn all at one time, but in a few instances it was removed by several small bleedings. In either case the general result was the same, as was determined by separate analysis of the data and accordingly all the data are included together in Table 64. The standard toxicity figures are given for comparison. It is apparent from the table that 35 percent of the animals which had been bled survived the acute period as compared with 13 percent survivals for the untreated animals. The percentage of dogs that died acutely was decreased from 87 percent to 65 percent. These figures indicate clearly the therapeutic value of venesection after chlorine poisoning in dogs under the conditions of the experiment.

TABLE 64.- Effect of venesection on chlorine poisoning after standard gassing

The result is at variance with that reported by British investigators upon goats. But it should be emphasized in the present experiment upon dogs that the bleeding was performed early (usually from one-half to two hours after gassing) and therefore before the onset of severe symptoms. At times it was observed that experimental animals died acutely during treatment. In all cases the animal struggled violently and in several instances death occurred before any attempt was made at manipulative procedure, bleeding or otherwise. Under these conditions it would seem that the observed collapse was caused by failure of the heart, a resultant of violent struggling rather than venesection. Furthermore, it should be observed in this connection that the conclusions reached are based on the actual death or survival of the animals experimented on and that the gassing procedure was carried out under carefully controlled conditions which permitted accurate comparison with previously determined toxicity figures.


Having established the therapeutic value of venesection, attempts were made to determine measures to still further increase the percentage of survivals. Infusion of physiologic salt solution is distinctly indicated when large volumes of blood are withdrawn. Under these circumstances the salt solution tends toward the temporary maintenance of the normal blood volume and helps to maintain in equilibrium other important functions. Furthermore, the thickened condition of the blood subsequent to gassing suggested the infusion for its probable diluent effect. Therefore, a series of 15 gassed dogs were bled 1 percent of body weight and immediately thereafter were infused with warm isotonic


sodium chloride solution, equal in volume to the blood withdrawn. The infusion was accomplished by inserting a needle under aseptic conditions through the skin into the jugular vein. The solution was allowed to flow slowly from a burette.

The results (Table 65) show 40 percent of survivals as compared with 35 percent susvivals for venesection alone, and 13 percent for the untreated series. Conversely the acute deaths are reduced to 60 percent as compared with 65 percent for venesection alone and 87 percent for the untreated dogs.

TABLE 65.- Effect of venesection and infusion of salt solution

A closer analysis of the data, however, for each of the three days which comprise the acute period is suggestive. It will be noted that, while both bleeding alone and bleeding plus infusion markedly reduced the total acute deaths (deaths during first three days) as compared with those in the untreated series, and, while bleeding plus infusion showed a greater reduction in this regard--and therefore was more efficacious, judged by the standard of this investigation--than bleeding alone, nevertheless bleeding plus infusion showed a much greater number of deaths during the first 24 hours and no deaths during the remainder of the acute period. In other words, the infused sodium chloride apparently had a deleterious effect on the most severely poisoned dogs--i. e., the dogs which, if untreated, would have succumbed acutely--in that it brought about the death of all within the first 24 hours after exposure, while it had a decidedly beneficial effect on another group of animals, presumably less severely affected by the gas, and successfully tided them over the three-day period and put them into the group of survivals.

Therefore the results from this experiment establish beyond peradventure that venesection is a valuable therapeutic measure, and indicate that venesection plus infusion of sodium chloride is better. Furthermore, the analysis of data for the respective 24-hour periods suggests that measures must be sought to save at least the percentage of untreated animals which are forced, as it were, from the latter part of the acute period into the first 24 hours.

Accordingly, the effect of substituting the infusion of isotonic acid, alkaline and neutral solutions of other salts in place of sodium chloride was investigated. The chief results are given in Table 66, from which it is apparent that both Na2HPO4 and Na3PO4 gave at least as high a percentage of survivals as sodium chloride. But Na3PO4 as well as NaH2PO4, gave rise to marked respiratory disturbances and so were not considered further. Na2HPO4, however, offered a suggestive lead, since this alkaline solution gave a slightly higher percentage of survivals than sodium chloride.


TABLE 66.- Influence of various salt solutions in the treatment of chlorine poisoning

Investigation of the urine after chlorine poisoning shows that there is a characteristic increase of acidity and an augmented excretion of ammonia, acid phosphates, and organic acids. All of these indicate acidosis. In addition, this is substantiated by direct determination of the bicarbonate value of the blood. Since alkali administration is suggested in all conditions where an acidosis is present, a study was made of the influence of sodium bicarbonate (5 to 10 gm. in 50 to 200 c.c. of water) per os, in conjunction with venesection and infusion of sodium chloride.

The results on a series of 28 dogs are summarized in Table 67. The animals given the sodium bicarbonate after bleeding and infusion show a distinctly greater percentage of survivals, 54 percent as compared with 40 per cent for those merely bled and infused. Also, the latter series shows 60 percent acute deaths, while the bicarbonate treated series shows only 46 percent acute deaths. It is to be noted again, however, that there was still the tendency, though by no means so marked, for acute deaths to be forced, as it were, from the later to the early part of the three-day acute period--there being more acute deaths during the first day after gassing followed by bleeding plus infusion and bicarbonate than after bleeding alone--though the net result of the former method is decidedly more beneficial.

TABLE 67.- Treatment of chlorine poisoning


From the investigation of therapeutic measures as thus far outlined, it is clear that bleeding is beneficial; that bleeding plus infusion of isotonic sodium chloride solution is more beneficial; and that bleeding and infusion plus the administration of sodium bicarbonate per os is most beneficial, since, under the conditions of the experiment, 54 percent of animals survived after the latter treatment as compared with 13 percent of survivals for the untreated series.


Attempts were then made to improve the treatment further by variations in the concentration of the infused sodium chloride solution; variations in the volume of the infused chloride solution; the infusion of isotonic nonelectrolytes; the substitution of other infusion fluids for the sodium chloride; other methods.


Forty dogs were given the treatment outlined above except that the concentration of the sodium chloride solution was slightly hypotonic or hypertonic. The results (Table 68) show at a glance that the infusion of the slightly hypotonic and hypertonic solution of sodium chloride is markedly inferior to isotonic solutions, the survivals varying from 16 percent to 35 percent for the former as compared with 54 percent for the latter. In fact, non-isotonic infusion is less beneficial than venesection without infusion at all. Emphasis, therefore, should be placed upon the necessarily narrow limits of concentration allowable in the infusion fluid.

TABLE 68.- The influence of various strengths of sodium chloride solutions in the treatment of chlorine poisoning

Twenty-eight dogs were studied in order to determine whether the infusion of smaller or larger volumes of sodium chloride solution than the volume of the blood (1 percent of body weight) withdrawn would improve the treatment. In this series sodium bicarbonate was not administered. Comparing the figures obtained with those for one volume of sodium chloride infusion without bicarbonate administration shows unquestionably that infusion of more or less


solution than blood withdrawn is far less beneficial (Table 69). Experiment also demonstrated (Table 70) that repeating the saline infusion 7 to 10 hours after the first treatment does not increase the percentage of survivals above the figure obtained by single treatments

TABLE 69.- The influence of the volume of fluid injected in the treatment of chlorine poisoning

TABLE 70.- The influence of repeated infusion in the treatment of chlorine poisoning


The infusion of solutions of nonelectrolytes such as isotonic dextrose alone and 3 to 6 percent gum acacia in isotonic saline (1 to 3 infusions) was also found to increase the acute mortality as well as decrease the survivals as compared with the standard toxicity figures for untreated animals (Table 71).

TABLE 71.-The influence of dextrose and acacia solutions in the treatment of chlorine poisoning



A series of 13 dogs was infused with Ringer's solution with double the usual amount of calcium. Another series of 15 dogs was infused with isotonic Na2SO4. Both of these fluids were markedly less beneficial than isotonic sodium chloride (cf. Table 72).

TABLE 72.- The influence of calcium and other salts in the treatment of chlorine poisoning


Morphine (5 mgm. per kilo) was administered to 10 dogs in addition to the prescribed treatment and found to be distinctly harmful. In this series the survivals were only 10 percent as compared with 54 per cent for dogs receiving the regular treatment alone and 13 percent for untreated animals (Table 73). Another series of 18 animals was treated in an entirely different manner. To 8 dogs 5 to 10 grams of NaHCO3 was administered by stomach, sound bleeding and infusion being omitted. This treatment was without beneficial effect (Table 74), the total acute deaths and the survivals being the same as for untreated animals. The 10 other dogs of the series were merely bled 1 percent and then given a subcutaneous infusion of 1 percent calcium chloride in 0.5 percent sodium chloride solution, according to the method outlined in the "Translation of Captured German Document." This procedure was found to increase the percentage of total acute deaths and decrease the percentage of survivals as compared with that for animals which received no treatment whatsoever (Table 74).

TABLE 73.- Effect of morphine in treatment of chlorine poisoning


TABLE 74.-The influence of subcutaneous injections of calcium in the treatment of chlorine poisoning


Finally, a study was made of the effect of environmental conditions, chiefly temperature, on the efficacy of the prescribed treatment. Throughout the definitive experiments, every endeavor was made to keep the gassed animals under as favorable conditions as possible. However, one series of 10 dogs was allowed to become chilled after exposure to the gas, and another series of 9 dogs was kept at a relatively low temperature. Both series received the prescribed treatment. The data (Table 75) show 20 percent survivals for the "chilled" and zero survivals for the "cold" series as compared with 54 percent survivals for the regular series kept under favorable temperature conditions. The result is a conclusive demonstration of the important part played by external temperature in determining the survivals and recoveries after gassing.

TABLE 75.- Effect of environmental conditions (temperature) on the efficacy of the prescribed treatment


The investigation just described of therapeutic measures for chlorine poisoning involved the study of over 300 dogs gassed at the standard lethal concentration d as worked out in the laboratory. It is obvious that the acute symptoms of chlorine poisoning are most prominent in dogs gassed at this concentration.
d From 800 to 900 parts chlorine per million of air (2.53 to 2.85 mgm. per liter) for a period of one-half hour.


Therefore, it must be emphasized that the intensive changes induced by this lethal concentration put to the most rigorous test any method of treatment of the condition. A treatment which is reasonably successful at this lethal concentration may be expected to produce much greater results when administered in conditions less severe; that is, in animals exposed to lower concentrations of chlorine gas.

Again, environmental conditions, such as temperature, are crucial factors in determining the efficacy of therapeutic measures.


Over 20 different therapeutic procedures were tried. These may be classified, on the basis of the data already presented, as follows:


(a) Bleeding, infusion of isotonic sodium chloride solution in volume equal to the blood withdrawn, administration of NaHCO3 by stomach sound (Table 68). (b) Bleeding, infusion of isotonic Na2HPO4 solution (Table 66). (c) Bleeding alone (Table 64).


(a) Bleeding, infusion of hypertonic or hypotonic sodium chloride solution, NaHCO3 by mouth (Table 68). (b) Bleeding, infusion of isotonic sodium chloride solution repeated 7 to 10 hours after the original treatment, NaHCO3 by mouth (Table 70). (c) Bleeding, infusion of isotonic sodium chloride solution in one-half or double the volume of the blood withdrawn (Table 69). (d) Bleeding, infusion of isotonic NaH2PO4 or Na3PO4 solution (Table 66). e (e) Bleeding, infusion of isotonic Na2SO4 solution, NaHCO3 by mouth (Table 72). (f) Administration of NaHCO3 by stomach sound (Table 74).


(a) Bleeding, infusion of isotonic sodium chloride solution, NaHCO3 by mouth, morphine (0.5 mgm. per kilo) (Table 73). (b) Bleeding, infusion of isotonic dextrose solution, NaHCO3 by mouth (Table 71). (c) Bleeding, infusion of 3 or 6 percent gum acacia in isotonic saline solution, NaHCO3 by mouth (Table 71). (d) Bleeding, infusion of neutral PO4 mixture (Table 66). (e) Bleeding, subcutaneous infusion of 1 percent calcium chloride in 0.5 percent sodium chloride (Table 74). (f) Bleeding, infusion of Ringer's solution with double quantity of calcium, NaHCO3 by mouth (Table 72).

Therefore, the investigation leads unmistakably to the conclusion that, under the conditions of the experiments, the first therapeutic method of those listed as "distinctly beneficial" is the one in which emphasis should be placed. This prescribed method may be restated as follows:

(1) Venesection, involving the withdrawal of approximately 1 percent of the body weight. (2) Intravenous infusion of warm sterile isotonic solution equal in volume to the blood withdrawn. (3) Administration of NaHCO3 by a stomach tube--5 to 10 grams in 50 to 100 cubic centimeters of water. Treatment to be carried out from one-half to two hours after exposure to the gas.

e The figures indicate a large increase in survivals, as compared with the untreated animals, but marked respiratory disturbances rendered the use of these salts inadvisable.



It was not until a study of phosgene poisoning was made that an adequate conception was obtained of the fundamental principles involved in gas poisoning. The treatment for chlorine poisoning was evolved by more or less empirical methods, whereas in the case of phosgene the changes taking place in the blood were well recognized and the stages of phosgene poisoning quite clearly defined previous to any extensive attempts at therapeutic measures. This understanding of the problem led to a much more rational plan of attack with regard to treatment.

It is of interest that the treatment for phosgene formulated on this plan was in principle the same as outlined for chlorine poisoning except that the time factor varied. With chlorine poisoning early bleeding and early infusion of salt solution were advocated. With phosgene early bleeding and late infu- sion were imperative. With chlorine a significant acidosis was present; hence, alkali by mouth was indicated. Inasmuch as acidosis failed to be manifested in phosgene poisoning except as a terminal acidosis, the administration of sodium bicarbonate was omitted. Since the time of infusion is directly indicated by the abnormally high concentration of the blood, a rapid method of determining blood concentration is essential. Blood concentration and hemoglobin content follow similar courses. Determination of hemoglobin, therefore, gives an exact indication of blood concentration and hence a means of determining when salt infusion should be made in order to restore concentrated blood to more normal conditions. A further indication of the condition of the animal is found in the temperature changes, which should be followed at one-half hour periods.


Approximately one hour after gassing, blood is drawn from a vein to the extent of 1 percent of the body weight. Bleeding at any time up to four hours after gassing is beneficial, but the best results are obtained when the withdrawal of blood is practiced about one hour after gassing.


In the first stage, blood concentration may exhibit one of two features after bleeding: (a) The blood becomes markedly dilute and slowly returns to normal concentration; (b) there is no significant dilution of the blood. The latter is an exceptional condition. The time of treatment, therefore, will depend upon which of these two conditions obtains. When the blood becomes markedly dilute and then slowly returns to the normal, infusion of 0.97 percent sodium chloride solution equal in amount to the blood withdrawn should be practiced when the concentration reaches the normal level. This usaally takes from 8 to 10 hours. On the other hand, when even after bleeding the concentration of the blood is not definitely decreased, infusion of salt solution should be delayed until there is a clear indication that the blood is becoming concentrated. This usually occurs from 6 to 8 hours after gassing. In any case the infusion should not be delayed beyond the point where the blood has reached a concentration of more than 25 percent above normal.

In following blood concentration, estimations of hemoglobin are made at intervals of an hour.

After the infusion of salt solution the concentration of the blood is followed at one-hour intervals in order to determine whether subsequent salt infusions are


indicated. In general after the first infusion the blood mnay begin to concentrate again within one hour, andl when this concentration continues it may be desirable to infuse subsequently, but judgment must be exercised in order to strike a proper mean between insufficient and excessive infusion. Insufficient infusion leaves the blood concentrated. Excessive infusion augments edema. So long as the concentration of the blood remains constant, infusion is unnecessary, and when the concentration diminishes the animal is on the road to recovery.


Usually rest and warmth are all that are necessary in this stage; but if the blood should become greatly diluted again and remain so, a second bleeding may be necessary. This condition, however, rarely occurs.


The adoption of early bleeding as a therapeutic measure in phosgene poisoning was made with the idea that it would be beneficial in preventing excessive dilution of the blood and would relieve the heart of the strain placed upon it by a significantly increased blood volume. This view was adopted previous to actual determinations of blood volume. It will be remembered that Eyster and Meek (Chap. XI) claim blood volume is not increased. If this is true then early bleeding plays little role with regard to the original conception. However, it is quite possible to interpret the beneficial effects of bleeding on the assumption that thereby pulmonary edema is modified in that extraction of considerable quantities of blood makes it more difficult for the organism to obtain from its tissues fluid sufficient to furnish an intensive edema. In other words, one must assume that bleeding tends to delay or inhibit blood concentration. It will be shown that this actually occurs.

From toxicity experiments it may be accepted that phosgene in concentrations of 70 to 80 parts per million (0.30 to 0.35 mgm. per liter) may be regarded as the lethal concentration.


The experience had been gained in phosgene poisoning that acute symptoms do not appear so rapidly as they occur with chlorine. It became, therefore, a matter of particular importance to determine whether early bleeding or later bleeding was more beneficial in the treatment of phosgene poisoning. To test this point bleeding was practiced at a period of one hour in one series of animals and at a period of five hours in a second series, one series comprising 38 animals, the other 30. The results are shown in Table 76.

TABLE 76.- The influence of the element of time of bleeding in the treatment of phosgene poisoning


It is quite evident from these figures that bleeding is decidedly beneficial in phosgene poisoning and that there is very little difference either in acute deaths or ultimate recoveries of the animals bled one hour or five hours after gassing. One may conclude that it is immaterial whether bleeding is practiced at one or five hours after gassing.


It is possible that repeated small bleedings (20 to 30 c.c. at frequent intervals) would exert equal or greater beneficial effects upon dogs gassed with phosgene than a single venesection. In Table 77 the results of such an investigation are recorded for different concentrations of phosgene.

TABLE 77.- The influence of multiple small bleedings

It is quite apparent that multiple small bleedings tended to have a beneficial influence at a concentration 61 to 70 (0.26 to 0.30 mgm. per liter) phosgene. The total acute deaths were reduced from 42 percent to 18 percent, and the recoveries increased from 46 percent to 63 percent. Beyond this concentration, however, the favorable effect of this procedure was apparently lost.


Although convinced that early infusion was not directly indicated, tests of its influence were made first by infusion of isotonic salt solution alone one hour after gassing, and in a second series of animals venesection one hour after gassing, followed immediately by infusion of salt solution. In both experiments salt solution to approximately 1 percent of the body weight was introduced. (Table 78.) These data demonstrate that infusion of 0.97 percent sodium chloride solution to the extent of 1 percent of the body weight one hour after gassing is distinctly detrimental in phosgene poisoning. It is also quite evident that the results were not as good as those obtained with bleeding only after one hour. The conclusion may be drawn, therefore, that immediate infusion is detrimental under the experimental conditions.


TABLE 78.- Influence of early infusion (recorded by percentage)


When venesection was practiced one hour after exposure to phosgene and infusion of salt solution was made at a period when blood concentration was occurring rapidly, results were obtained as recorded in Table 79. Of these 35 dogs only 30 received infusion, inasmuch as 5 animals died previous to the time when it was possible to introduce the salt solution. It is quite evident, therefore, that the figures do not adequately express the influence of the procedure. If only those animals receiving infusion are included, then the total recoveries are significantly increased and the animals living beyond the three-day period are increased from 58 percent with bleeding alone to 80 percent with bleeding plus delayed infusion. Even without consideration of these 5 dogs, delayed infusion after early bleeding noticeably reduces death even though ultimate recoveries are not increased over bleeding alone.

TABLE 79.- Influence of early bleeding and delayed infusion



In order to put the method of treatment to an even greater test observations were made on dogs gassed at a concentration of 80 to 90 parts phosgene (0.35 to 0.39 mgm. per liter) per million of air. The results are presented in Table 80. In these experiments the animal was bled 1 percent of its body weight one hour after gassing, and then if it seemed desirable, bleeding to the extent of one-half of 1 per cent was later practiced. Water was given freely. It will be seen that at this concentration a single large bleeding did not yield results much better than the toxicity figures. Bleeding followed by immediate infusion was distinctly detrimental, as were multiple infusions. Bleeding plus delayed infusion, when emphasis is placed upon a definite time after gassing for infusion rather than an exact adherence to the degree of concentration of the blood, gave the results seen in column 5, Table 80. These are little better than bleeding alone. This observation is a clear indication that in order to yield the best results definite criteria of the animal's condition must be followed closely and the kind and time of treatment should be practiced in strict accord with these criteria. When changes in temperature and hemoglobin were employed in this connec- tion, and treatment was given to dogs gassed with phosgene at a concentration of 80 to 90 (0.35 to 0.39 mgm. per liter) parts per million of air, results such as those presented in column 6 were obtained. Obviously these observations lead to the conclusion that the method of treatment herein outlined is rendered decidedly efficacious when careful consideration is given to the time element with respect to infusion. In these results in certain instances more than one infusion was given at times, the number being determined entirely by the response of the change in hemoglobin. If after a single infusion the blood concentration again rapidly increased, a second or even a third infusion was given. Column 7 in Table 80 shows that even when the blood concentration and temperature are not followed so closely, excellent results may be obtained by following the method in general. One reason for the poorer results shown in column 7 of Table 80 is that in a number of animals included the wrong strength of salt solution was employed, every animal receiving it dying.

TABLE 80.- Influence of various types of treatment (recorded by percentage)


In this connection attention should be called to the necessity of infusion of the correct strength of sodium chloride solution. Excellent results were yielded when this salt solution varied in strength from 0.95 percent to 0.98 percent. Variations on either side of these limits were attended with detrimental effects. Thus in Table 80, column 8, it may be seen that when dogs were infused with sodium chloride solution of a strength of 1.02 percent all died, even though, in every other respect, the treatment given was in strict accord with the prescribed method.


The final step in the method of treatment was the recognition that it was unnecessary to infuse every animal exposed to phosgene and, further, that by early repeated bleeding the number of gassed animals needing infusion could be appreciably reduced. Moreover, the possibility was presented that, inasmuch as the essential feature of infusion is to dilute the blood even though only temporarily, the same result could be obtained by introducing water by routes other than by direct infusion, namely, by the mouth or intraperitoneally.

The following considerations are pertinent with respect to the modified procedure:

The change in hemoglobin of the blood and the temperature curve may be accepted as accurate criteria of the condition of the gassed dog.

These criteria may be depended upon as indicators of the time for treatment and the type of treatment.

Intensive treatment in the first state (that is, during the period of blood dilution) of phosgene poisoning, in the majority of cases, will prevent extreme concentration of the blood characteristic of the second stage. In other words, the second stage is very greatly modified by proper treatment of the first stage. During the first stage water should not be given.

Proper treatment of the first stage consists in venesection to the extent of 0.5 percent body weight as soon as practicable after gassing. The temperature and hemoglobin are then followed at one-half hour intervals. So long as the temperature remains normal and blood concentration does not diminish further treatment is not indicated. When, however, the temperature rises rapidly and a fall in blood concentration occurs (the two changes take place simultaneously) a second venesection of 0.5 percent body weight is practiced. This procedure may be repeated for the second time, that is, until blood to the extent of 1.5 percent of the body weight has been withdrawn.

The large majority of cases needed no further treatment and practically every animal survived.

If in spite of intensive treatment in the first stage, the blood became markedly concentrated and a marked fall in temperature took place, the condition of the animal was considered as very serious, and if left untreated it would surely die.

Under these conditions two types of treatment were carried through: (a) To those animals that would drink, large quantities of water were given, or water was given by a sound; (b) infusion or intraperitoneal injection of 0.95 percent sodium chloride solution was practiced.

Probably 50 percent of animals in a serious condition in the second stage might be saved by following either procedure.

Injudicious infusion of salt solution might be markedly detrimental, hastening the onset of death. Judicious infusion of salt solution resulted in the survival of a large number of animals.


Infusion of salt solution was necessary only with the most serious cases, and a considerable degree of judgment had to be employed.

Finally, the aim of treatment in phosgene poisoning should be to consider each case as an entity, and to give treatment only in accordance with changes in temperature and hemoglobin, the criteria of changes in the gassed animal's condition.

In Table 81 will be found the observations made with the modified treatment. It may be concluded that the results yielded are almost as good as the best obtained when every animal was infused, and that the modified method is more practical under field conditions than would be a method where intravenous injection is necessary. From a practical viewpoint then, this modification is of considerable importance. Equally important practically is the fact that not every animal needs even water introduction, the blood concentration being controlled in large measure by the early repeated bleeding. It is significant that by following the modified method delayed deaths are entirely eliminated.

TABLE 81.- Modified treatment


It was deemed desirable to determine what influence upon phosgene treatment would be exerted by changing the water intake of an animal previous to gassing. The results may be seen in Table 82. If water was withheld from a dog 48 hours previous to exposure to phosgene, response to treatment is very poor. In other words, water starvation previous to gassing exerted a distinctly detrimental effect. The administration of water just previous to exposure to gas gave results entirely comparable with those of the toxicity figures. In other words, too much water was also definitely detrimental. When water administration constituted the only treatment, little or no improvement over the toxicity figures was observed.

TABLE 82.- Influence of water intake upon treatment



In certain quarters objection was raised to the injection of salt solution on the ground that such infusion would tend to aggravate edema and hence perhaps exert a detrimental influence. From the experiments carried through by Capt. Samuel Goldschmidt and Capt. David Wright Wilson, of the Chemical Warfare Service, at Porton, England, with goats (see Chapter XX), it may be stated that this objection is without foundation. Goats gassed with phosgene, bled and infused with sodium chloride solution, gave no evidence that this procedure aggravated the existing edema. These results were obtained by determination of the lung to heart (L:H) ratio with and without infusion, and if edema were actually increased by the infusion the change would surely be indicated by alterations in this ratio. The results follow:


The average L: H ratios were slightly less in the animals that had been treated, both in the groups that died and in those which survived. Goats dying after being treated by venesection alone failed to show any marked difference in the L: H ratios. Therefore, the conclusion may be drawn that infusion does not increase the edema of the lungs in phosgene poisoning. Indeed, the evidence indicates that the edema is somewhat decreased.


The treatment of chloropicrin poisoning was carried through upon both dogs and goats. In the latter instance the observations were made at Porton, f England, by Captain Goldschmidt and Captain Wilson.


The method of gassing may be outlined briefly. Ten or 12 goats were placed in the gassing chamber and left for a period of 25 minutes in an atmosphere of chloropicrin, made by spraying in the liquid in quantities sufficient to develop a theoretical concentration of from 1 part of chloropicrin in 8,500 to 9,000 parts of air. The air in the chamber was kept continually in motion during the period of the gassing by means of an electric fan.

The animals were arbitrarily divided into two equal groups; one group was treated, while the other served as a control. An effort was made to have, so far as possible, an equal division of animals as regards weight and sex, in the two groups. This division was made without any reference to the animal, before the gassing was completed, and adhered to regardless of the condition of the respective animals after gassing.

Bleeding was performed by aspiration through a needle inserted through the skin into the right or left jugular vein. This procedure was performed so far as practicable under aseptic precautions.
f We are greatly indebted to Lieut. Col. Crossley and Mr. Joseph Barcroft for placing these facilities at our disposal as well as for many helpful suggestions and criticisms.


One bleeding only was made one hour after the animals were removed from the chamber.

All of the animals which survived were kept under careful observation for a period of 5 to 7 days, and then sacrificed for post-mortem examination. As nearly 90 percent of the mortalities occurred within the first 24 hours, and most of the surviving animals improved rapidly thereafter, the length of the period of observation is considered ample. Hence it may be safely stated that life or death has been the criterion of the beneficial effect of the treatment employed. A large number of experiments were performed, but for present purposes only those groups in which from 60 to 100 percent of the goats in the control group died are selected. Tables 83 and 84 give the results of these experiments.

TABLE 83.- Effect of bleeding in goats gassed with chloropicrin

TABLE 84.- Effect of bleeding upon goats gassed with chloropicrin

As maybe seen in Table 83, the mortality of the treated animals is decreased from 90 percent to 57 percent by a single bleeding in a severely gassed goat. This large reduction in mortality in the 95 animals used in these experiments demonstrates conclusively that bleeding is an efficacious therapeutic procedure when applied early to goats gassed with high concentrations of chloropicrin. These data have been subjected to statistical analysis and found to be conclusive.

Table 34 shows the percentage deaths (expressed as per cent of animals surviving at the beginning of each period) of experimental and control animals at four periods after gassing. The first critical period occurred within 8 hours after gassing. An inspection of Table 84 shows that whereas 40 per cent of the control animals had died at this time but 17 per cent of the treated animals had succumbed at this period.

The next and more critical period lay between 8 and 24 hours. Here also a greater percentage of the untreated animals died. Hence a single treatment had furnished protection for the treated animals through the severest portion of the acute stage of the gassing. After 24 hours the number of treated animals.


which died was very small; on the other hand, the untreated continued to give an appreciable percentage of deaths. It may, therefore, be concluded that the treatment gave a permanent advantage, and did not merely delay the death of the animal.

Unfortunately the observations upon goats were not sufficiently extended to show the influence of infusion upon this animal.


The first type of treatment which was tried with dogs subjected to chloropicrin poisoning was multiple bleeding. In Series I (Table 85) of this treatment 21 dogs were treated which had been gassed at a concentration of 0.81 to 0.95 mgm. per liter, and in Series V 41 dogs which had been gassed at a concentration of 0.96 to 1.10 mgm. per liter. The animals, as soon as received from the gassing chamber, generally about a half hour after the close of gassing, were bled one-half per cent of body weight. The animals were followed every hour for hemoglobin. If the hemoglobin showed a concentration the animal was bled another one-half per cent of body weight, and further concentration was followed by further bleeding.

TABLE 85.- Chloropicrin treatment

As shown in Table 85, the results obtained from this type of treatment at both concentrations were practically identical with the regular toxicity figures, and it appears clear, considering the large number of animals used, that this type of treatment is of no benefit.


The fundamental idea of this treatment is that if the animals are bled as the blood concentrates, additional fluid will be supplied from the tissues, and as a result the blood will become less concentrated. So far as one can tell this does not happen. The records, in general, give no evidence that the multiple bleedings bring down the concentration of the blood, and in fact in some cases at least it continues to increase more rapidly than without bleeding. It appears probable that as soon as the edema begins to develop in the lungs, with a consequent loss of fluid from the blood, that the tissues are immediately called upon to supply additional fluid. As long as fluid is supplied from the tissues in this manner, the concentration of the blood will be kept to a minimal figure, but as soon as the fluid reserve of the tissues is exhausted the concentra- tion of the blood will increase in exact ratio to the development of edema in the lungs. Inasmuch then as the fluid reserve of the tissues has been used by this time it appears logical to suppose that additional bleeding would only further aggravate the concentration and this is apparently what happens.


SERIES I.-Multiple bleeding. Animals bled one-half percent of body weight within one-half hour after close of gassing. Further bleeding of same amount upon concentration of blood.
SERIES II.-No bleeding. Animals given 2 percent of body weight of H2O by stomach tube within one-half hour after close of gassing. Further fluid supplied according to hemoglobin and clinical symptoms.
SERIES III.-Bled 1 percent of body weight within one-half hour after gassing and given 2 percent H2O by stomach tube immediately afterbleeding. Additional fluid supplied by intravenous or intraperitoneal infusion of isotonic saline solution indicated by hemoglobin or clinical symptoms.
SERIES IV.-Same as Series III except animals were bled one-half percent of body weight instead of 1 percent.
SERIES V.-Same as Series I.
SERIES VI.-Bled 1 percent and given intravenous infusion of 1 percent isotonic saline within two hours after close of gassing.
SERIES VII.-Same as Series IV.

In Series II the animals were not bled, but were given 2 percent of body weight of water by stomach tube within an hour after gassing. The treatment of this series was just the opposite of that in Series I. To counteract the concentration of the blood, fluid was supplied at once which could be absorbed into the blood stream, and the blood thus kept at a lower concentration. The recoveries by this treatment were only very slightly above the toxicity figures.

In Series IV and VII, the final treatment adopted, the animals were bled one-half percent of body weight within a half hour after the close of gassing and 2 percent water was given to the animal at once by stomach tube. The animals were closely followed, and in case a need of further fluid was indicated by hemoglobin or temperature it was supplied. Fifteen animals which had been gassed at a concentration of from 0.81 to 0.95 mgm. per liter and (Series VII) nine animals which had been gassed at a concentration between 0.96 to 1.10 mgm. per liter were treated by this method. Out of the 15 animals at the lower concentration, there were no deaths the first day and only three deaths


altogether, giving a percentage of recoveries of 80 percent. At the higher concentration there were two deaths the first day and one the second day, leaving 6 animals which recovered, or practically 67 percent. The percentage survivals in this treatment series was very much greater than in the regular toxicity groups.

This same treatment was tried, with the exception that the animals were bled 1 percent of body weight instead of one-half percent. As shown in Series III nine animals were thus treated, of which five recovered, giving a percentage recovery of about 56 percent, as compared with 80 percent when the one-half percent bleeding was used. Although the number of animals used in this series was small the results clearly indicated that the one-half percent bleeding was the better.

In Series VI nine dogs which had been gassed at a concentration of 0.96 to 1.10 mgm. per liter were bled 1 per cent of body weight andgiven an intrav- enous infusion of 1 per cent of body weight of normal saline immediately after- wards, the whole process taking place within two hours after gassing and at the time when the blood began to concentrate. The percentage of survivals for this was 44 per cent, which is an increase of 12 per cent over regular toxicity recoveries, and a decrease of 23 per cent when compared with the final adopted treatment as given in Series VII.

The final treatment, as given above, in Series IV and VII, appears to be a logical one when the condition of a gassed animal is considered. The bleeding of one-half percent of body weight is not sufficient to weaken the animal in any way, but a priori would appear to be efficacious in two ways. In the first place, providing any toxic substances had been taken up by the blood, as a result of gassing, the bleeding serves to relieve the blood of a certain amount of these substances. In the second place, the bleeding initiates influx of new fluid from the tissues into the blood. As the edema in the lungs develops and the blood concentrates, the influx continues until the fluid reserve of the tissues, augmented by the 2 percent of body weight of water given at the time of bleeding has been depleted, or, in the less serious cases, until the edema of the lungs has ceased to develop.

In the animals which were seriously affected this basic treatment of bleeding and supplying of fluid by mouth or intraperitoneally was not sufficient to save the animals. Evidently in such cases the edema of the lungs continues to develop so rapidly that even the augmented fluid reserve becomes exhausted. In serious cases of this type it was found that an intravenous injection of about 1 percent of body weight of normal saline would, at once in many cases, reduce the concentration of the blood to such an extent that the animal would be carried through the period of rapid edema development. Such an infusion can be repeated if necessary.

It is quite obvious that from the data the most efficacious method of treatment in chloropicrin poisoning is to bleed one-half percent of the body weight as soon as possible after gassing and then at once administer water by way of the mouth. Additional fluid may be supplied later by intravenous infusion of sodium chloride solution in accord with hemoglobin readings and temperature changes. With this treatment there are no "delayed deaths."



Clinical experience in gas poisoning leads to the conclusion that lack of oxygen plays a significant role in this condition. This inference has been confirmed by experimental evidence and various possibilities may be cited as causes for the establishment of anoxemia. Thus it is possible that the inflammatory changes in the lungs may hinder the absorption of oxygen. Vas- cular obstruction in the lungs, owing to swelling of the alveolar tissues, and the development of edema may also result in poor oxygen absorption; or again, there may be an irregular distribution of air in the lungs, owing to shallow breathing or partial obstruction of some of the bronchi. Any one or all of these conditions may operate to decrease oxygen absorption. All of these possibilities, however, carry with them the assumption that there is an inadequate oxygenation of the blood, meaning thereby an insufficient supply of oxygen in the arterial blood. According to experimental results there is little or no evidence of an inadequate supply of oxygen in the arterial blood during the first part, if not the whole, of the first period of phosgene poisoning. When, however, blood concentration becomes marked insufficient oxygen in the arterial blood is quite apparent. In a previous section of this chapter emphasis has been laid upon changes in blood concentration as the responsible factor leading to the condition of anoxemia. According to this view the viscosity of the concentrated blood leads to impaired circulation through the tissue capillaries thus accounting for the abnormally low oxygen content of the venous blood. Oxygen lack in the arterial blood is not present to the same degree.

From these observations it becomes apparent that the need for oxygen in the first period of phosgene poisoning is not great. In the second period, however, oxygen lack becomes quite pronounced, especially in the venous blood. The question naturally arises: Will administration of oxygen eliminate anoxemia? Again, if anoxemia is alleviated will this allow an individual to survive the effects of phosgene poisoning? In other words, is inadequate oxygenation of the blood responsible for death? From clinical experience there seems to be conflicting evidence as to the value of oxygen in the treatment of phosgene poisoning. On the whole, however, it would appear that the consensus of opinion indicates that oxygen administration is decidedly beneficial in the circumstances under discussion.


In order to ascertain whether oxygen administration is of therapeutic value dogs gassed with the lethal concentration of phosgene (70 to 80 parts per million of air; 0.30 to 0.35 mgm. per liter) were placed in a respiratory chamber through which oxgyen was circulated. The concentration of oxygen selected was 50 percent, small variations from this figure occurring during the course of the experiments. Analysis of the air in the chamber was made from time to time. The dogs were placed in the chamber 2 to 5 hours after gassing and kept there continuously up to a period of 72 hours. This period was selected for the reason that it was chosen as the limit of the duration of the acute stage of poisoning. Arrangements were made for the absorption of carbon dioxide and regulation of moisture and temperature within the chamber. Water was supplied to the experimental animal.


The investigation was carried out upon 30 dogs to which oxygen was administered. A control series of 27 animals gassed at the same concentration but without any treatment, except that they were kept quiet and warm in the same room with the experimental dogs, was also observed. (Table 86.)

TABLE 86.- The influence of oxygen administration upon the mortality of dogs gassed with phosgene

The results indicate, under the experimental condition, that: (1) In both series by far the larger percentage of deaths occurred within the first 24 hours. (2) All the animals that died (exclusive of "delayed deaths") did so within 48 hours. (3) Oxygen increased but slightly the percentage of survivals, from 26 percent to 37 per cent. (4) Oxygen had even less effect on prolonging the life of the dogs beyond the first 24 hours. (5) Oxygen did not increase the number of dogs that survived the "acute period" to die later of secondary causes. (6) Oxygen did not markedly increase the percentage of dogs that ultimately recovered.

If, instead of employing the exact controls, comparison is made between the so-called "toxicity figures" and the results obtained with oxygen administration, the beneficial influence of oxygen upon dogs gassed with lethal concentrations of phosgene is entirely lacking. The results, especially with respect to the survivals beyond the acute period and the ultimate recoveries, were almost identical (Table 87).

TABLE 87.- Comparison of toxicity figures for phosgene and those obtained by oxygen administration


Although these figures are very striking, too much emphasis should not be laid upon them, inasmuch as the number of animals employed in the two series were too dissimilar. This comparison, however, emphasizes the fact, revealed by the comparison with the actual control series, that continuous oxygen ad- ministration after phosgene poisoning did not appreciably decrease the percentage of deaths nor can it be said to have materially prolonged life. There is the same difference present that may be obtained when two series of animals are gassed by different observers under unlike environmental conditions. Though this conclusion is inevitable from the data it must be conceded that oxygen administration seemed to relieve the animal. It rested more quietly, respiration was less difficult and obvious cyanosis disappeared or was absent. From the failure of oxygen to prolong life in spite of obvious improvement in clinical symptoms it would appear that such a therapeutic measure does not alleviate the fundamental difficulty in phosgene poisoning.


Thirty dogs were gassed with a lethal concentration of phosgene (71 to 80 parts per million, or 0.31 to 0.35 mgm. per liter of air) and one hour later 1 percent of the body weight of blood was withdrawn from the jugular vein or from the femoral artery, in both cases without anesthesia. The animals were then placed in the oxygen chamber and kept there up to 72 hours. The results of these observations may be seen in Table 88, where a comparison is made with the effects of bleeding alone. It is quite obvious from these results that bleeding plus oxygen administration as a therapeutic measure has no appreciable advantage over simple venesection. Although carried through by different men in separate laboratories at different times of the year, the results in the two series of observations have a striking similarity, both of kind and degree.

TABLE 88.- The influence of oxygen upon gassed dogs treated by venesection


Of a series of 14 dogs gassed at a concentration of 50 to 60 parts of phosgene per million parts of air (0.21 to 0.26 mgm. per liter) for one-half hour and treated with oxygen, 57 percent survived beyond the three-day period. The toxicity figure is 54 percent. Oxygen can be said, therefore, to have no beneficial


effect on dogs gassed at a concentration lower than "lethal." This is corroborated when the recoveries in the two cases are compared-43 per cent with treatment as against 44 percent without treatment (Table 89).

TABLE 89.- Influence of oxygen upon dogs gassed with sublethal concentrations of phosgene and treated by venesection

The next logical step would have been to discover the influence of oxygen administration upon the final prescribed treatment for phosgene poisoning, namely, venesection followed by infusion of sodium chloride solution.

Just before it was imperative to discontinue the work a series of experiments was started in this direction from which results of a positive nature were anticipated. In this series the dogs were bled 1 percent of their body weight as soon after gassing as possible, followed at a proper time by an infusion of warm, sterile, isotonic (0.95 percent) NaCl, and the dogs then placed in the oxygen apparatus. When the outcome of the work dealing with the study of the blood-oxygen after gassing, the dog breathing air as compared with breathing oxygen, before and after bleeding, and before and after infusion is read (see p. 711), it will be realized why the carrying out of this series seemed to be so promising of results. The particular point involved was the ability of the gassed dog to get hold, as it were, of the additional amount of oxygen provided, as well as of that in atmospheric air, as evidenced by the arterial and venous oxygen content when infusion was carried out. Three dogs only were thus treated. The first died just under 72 hours after gassing, the other two recovered.


The administration of oxygen to chlorine dogs placed in the circulating apparatus was substantially the same as used for phosgene, except that since the work on chlorine was done first the apparatus was not as complete as it was later.

Three dogs were given oxygen, with no previous treatment, after gassing. Of these none survived. The first died on the second day; the other two on the first day.

Eight dogs were given oxygen after treatment as developed at the time. Within one-half hour after gassing, the dogs were bled 1 percent of their body weight by aspiration from the jugular vein. The same amount of warm, sterile, isotonic (0.95 percent) sodium chloride was then slowly infused, after which 50 cubic centimeters of 10 per cent NaHCO3 were given by a stomach tube.


The dog was then placed in the oxygen box, and the concentration of oxygen in the circulating air gradually raised to approximately 50 percent in the manner described in the report of the work on phosgene. The outcome of this experiment is recorded in Table 90.

TABLE 90.- The influence of oxygen upon dogs gassed with chlorine treated by the standard method

At this concentration of chlorine the toxicity series shows a percentage recovery of 9 percent. The number of dogs involved in the present series was entirely too small to warrant a definite conclusion. The fact that the percentage of recoveries was increased by 16 percent may be of significance. On the other hand, when comparison is made with treatment figures obtained, it is true, in a large number of dogs, it would appear that oxygen has little influence in increasing the number of recoveries; that is, with oxygen the recoveries were 25 percent, without oxygen 29 percent.


After it had been shown that oxygen administration alone or in combination with venesection was of doubtful value in reducing the mortality of dogs gassed with phosgene it was deemed important to determine the exact changes that occurred in the respiratory functions of the blood after exposure of animals to phosgene. Therefore the oxygen of the blood was determined in gassed dogs breathing air and breathing oxygen. As an ideal procedure for this purpose blood should be drawn from the right and left ventricles for the sample of venous and arterial blood, respectively, inasmuch as the blood contained in these receptacles undoubtedly more nearly represents the average composition than blood secured at a greater distance from the heart.

Although the figures may not be identical, the general results obtained by analysis of heart blood and that of vein and artery blood are of the same kind. This has been demonstrated. In the present investigation the venous blood was obtained by aspiration through a needle inserted through the skin into a jugular or femoral vein, and the blood drawn into a test tube under albolene. The arterial blood was obtained in the same way through a needle inserted through the skin into a femoral artery, the position of the artery being determined by palpation.



It was first ascertained whether the oxygen content of the arterial and venous blood of the normal dog can be raised by the administration of oxygen. As a result of these experiments (Table 91) it is quite evident that this was possible in those cases in which the arterial percentage saturation was low originally. It was not, however, an invariable rule.

TABLE 91.-The arterial and venous oxygen content and percentage saturation breathing air as compared with breathing oxygen (approximately 50 percent). The capacity of the blood (oxygen-combining power) is also given

It should be noted that, in the cases where the percentage saturation of the venous blood showed a decrease while the dog was breathing oxygen as compared with the value obtained when the animal was breathing air, the contents did not vary by much more than 1.5 percent. This was the error of the method used for analyzing oxygen (Henderson-Smith method).


The study of the blood oxygen was undertaken primarily to discover whether the determination of its content in the arterial and venous blood would not indicate the reason why oxygen alone seemed to have no therapeutic value, since, as has been noted above, the percentage of survivals in those series in which oxygen was given was practically identical with that obtained in the " toxicity series," for the same concentration of gas.

A study of Table 92 will show that in spite of the administration of oxygen and the ability of the animal to raise the oxygen content of the blood, in general, the percentage saturation became successively lower. In other words, although additional oxygen could in many cases be forced into the blood in relatively great amounts either soon after or several hours after gassing, nevertheless, owing to the gradual decrease of the oxygen content of the blood and the consequent lowering of the percentage saturation, the blood was further and further away from saturation.


TABLE 92.-The influence of oxygen administration upon the percentage saturation and oxygen content of the blood of dogs gassed with phosgene (71 to 80 parts per million) (0.31 to 0.35 mgm. per liter)

It was often observed that it was impossible to raise the venous content, although the arterial content could be markedly increased. This would seem to indicate that while there may have been an abundance of oxygen in the arterial blood for ordinary purposes, there was such an increased demand for it on the part of the tissues that the increased amount not only did not get to the venous blood, but the oxygen there might show a continued marked decrease. Dog No. 89 was an example of such a case, in which, although the oxygen in the arterial blood was increased, the percentage saturation being raised from 90 percent to 95 percent--the capacity showing relatively little change--it was impossible to prevent a very rapid fall in the oxygen content of the venous blood. As long as the dog was breathing oxygen the arterial percentage


saturation remained high, but fell off rapidly when atmospheric air was substituted for the 50 percent oxygen. Dog No. 113 exemplified a similar condition.

Speaking of the "capacity" is perhaps applying a misnomer for the condition which we had in these gassed dogs. That is, the term did not here represent the true combining power of the blood, in the body, with oxygen, in the conditions where the blood had thickened considerably. A cubic centimeter of thickened blood contained more hemoglobin, in ratio to the plasma, than did a cubic centimeter of blood of normal or average viscosity. In the body the capacity of the blood as a whole must be really lower, since it has lost in volume, due to the loss of water. This probably explains, in part, the drop in oxygen content. Furthermore, the increased viscosity slowed the blood stream, and led to poor oxygenation of the tissues.

Dog No. 105, an animal that recovered, illustrates a recovery which was plainly not due to the oxygen. Seventeen hours after gassing the percentage saturation of the arterial blood was practically the same as the normal figure (93 percent as compared with 91 percent). The arterial content rose parallel with the capacity, the former increasing from 22.4 to 25.1, the latter from 24.4 to 27.1.

The venous content, on the other hand, had dropped from 61 percent to 23 percent. Oddly enough, when the blood was examined after the dog had been breathing oxygen, the arterial content was found to be lower, while the venous content was higher. This was obviously a mistake in the one or the other analysis. But it is interesting to note that the next arterial sample, taken at 2.15, had a lower oxygen content still, although the percentage saturation was the same as the oxygen breathing sample, due to a gradual fall in the capacity; and that the sample of venous blood when the dog was breathing air had about the same content as a previous sample taken when the dog was breathing air at 9.30. Be that as it may, the next administration of oxygen, beginning at 2.20, raised the arterial content to 100 percent saturation, although the capacity was decreasing all the time, and kept it raised while the oxygen of the venous blood seemed to have been prevented merely from decreasing. There was a sharp falling off of the oxygen in the arterial blood, the saturation dropping to 55 percent, after the dog was returned to air, and a drop in the capacity, which had the effect of raising the saturation of the venous blood, although the venous content was actually lower. The later figures show the condition 41 to 65 hours after gassing, respectively. The capacity decreased still further, below normal, and at 41 hours after gassing the venous saturation dropped as low as 12 percent, although the arterial was practically normal (93 percent). Sixty-five hours after gassing the venous percentage saturation had risen to 32 percent, or about the value for it 17 hours after gassing, but only about one-half that of normal.

Dog No. 113 showed in summary fashion the inefficiency of oxygen as a therapeutic agent alone, in a crucial case, that is, one in which the effects of the poison were advancing rapidly. This case showed as marked and as rapid a rise in the "capacity," or of the thickening of the blood, as any that have


come under our observation. The arterial content rose with the capacity, but not fast enough to keep up the percentage saturation, which dropped from 97 percent to 91 percent four hours after gassing. The venous blood lost oxygen. Oxygen administration was begun at the fifth hour after gassing. The percentage saturation of the arterial blood six hours after gassing was 100 percent; the venous continued to fall rapidly. Forty-five minutes later, although still breathing oxygen, the percentage saturation had fallen to 74 percent, while there was no oxygen that could be detected in the venous blood. The content of the arterial blood at this time was the same as its value four hours after gassing while the dog was breathing air. But the capacity was then 30.5 as compared with 24.5 four hours after gassing. Another sample taken five minutes later, the dog having been returned to air, showed a still further decrease in oxygen content of the arterial blood, and a few minutes later the dog died.

This was particularly important as illustrating the futility of dealing directly with one of the secondary effects of gas poisoning, viz, the reduced amount of oxygen in the tissues. The lack of oxygen must be in one of two places, either in the arterial blood, because it can not be taken up from the lungs; or, if the arterial blood can take up oxygen in additional amount, the lack of oxygen is in the tissues themselves. The first possibility is ruled out, inasmuch as oxygen can be absorbed by the gassed lungs. In spite of the fact that the oxygen of arterial blood may be within normal limits under oxygen administration there still may be evident oxygen want in the tissues. Obviously oxygen administration alone does not eliminate the crucial condition in gas poisoning, namely, lack of oxygen in the tissues. Although an additional amount of oxygen in the inspired air may be of some benefit to the tissues, the primary object in treatment should be to remove or modify the cause for the lack of oxygen.

Although oxygen could not be shown to have a curative effect on dogs gassed at the "lethal" concentration (viz, 70 to 80 parts of phosgene per million of air; 0.30 to 0.35 mgm. per liter),the question was raised as to whether it would be beneficial to animals gassed at a lower concentration, and perhaps thus more in accord with conditions of men in the trenches. Accordingly the study of the oxygen in the blood of dogs gassed at a concentration of 50 to 60 parts phosgene (0.21 to 0.26 mgm. per liter) was made. Table 93 is a summary of the results. From these data the same general conclusions may be drawn regarding dogs gassed with a concentration of phosgene lower than lethal as pertained to dogs gassed at 70 to 80 (0.30 to 0.35 mgm. per liter). In some cases it was impossible to raise the venous content (dog No. 126), although this was exceptional. In others, after an initial raising of the percentage saturation of both arterial and venous blood, there was a more or less rapid fall even while the animal was breathing oxygen, which became more rapid after a return to breathing air (e. g., dogs Nos. 123 and 124).


TABLE 93.- The influence of oxygen administration upon percentage saturation and oxygen content of the blood of dogs gassed w'ith phosgene (.50 to 60 parts per million) (0.21 to 0.26 mgm. per liter)



It is apparent from the study of the blood oxygen as outlined above that oxygen alone can not be of more than temporary value in relieving the effects of gas poisoning. The next logical step in the development of the possible therapeutic use of oxygen was the study of the blood following bleeding and infusion plus oxygen administration. A large number of experiments were carried through. From the results of the experiments on the dogs which were placed in the oxygen chamber after venesection it is clearly evident that the respiratory function of the blood was not improved by venesection alone. Although no study of the influence of venesection on the arterial and venous content of the blood was made immediately after venesection, data are available which show that the beneficial effects of breathing oxygen instead of air are considerable, although practically the same as the effects obtained when venesection is not performed. From these data it is quite evident that bleeding does not materially alter conditions relative to the respiratory functions of the blood. However, at the time of venesection the oxygenation of the blood and tissues is not, in general, markedly deficient. The respiratory function of the blood is distinctly better when an animal is breathing oxygen than when ordinary air is breathed. The order of change is not, however, appreciably different from that which obtains when an animal breathes oxygen without venesection.

Bleeding followed by infusion may be said to have a decidedly beneficial effect upon the oxygenation of the blood. This is indicated by the fact that even when an animal is breathing air the venous blood carries more oxygen, which may be interpreted to mean that the tissues are being better supplied with oxygen. Oxygen administration at this time results, in general, in restoring the arterial blood to limits near the normal and in raising appreciably the venous saturation. It would appear that bleeding plus infusion so changes the physical character of the blood as to render possible a more complete oxygenation of tissues. The interpretation placed upon these results is that infusion decreases the viscosity of the blood, at least temporarily, so that capillary circulation is improved and, as a consequence, more adequate tissue respiration is possible. The dissociation of the oxyhemoglobin is probably also increased in accordance with the result of Barcroft on the effect of electrolytes on the dissociation curve of hemoglobin. It seems, therefore, that the optimum conditions for the use of oxygen are in connection with bleeding and infusion of salt solution. No single procedure is adequate; a combination of the three approaches the optimum treatment.


The extent of oxygenation of the blood will depend directly upon the rate of respiration and the volume of air breathed per minute and upon the circulation rate through the lungs. It becomes of importance, therefore, to determine these factors if a complete knowledge is desired concerning changes induced by gassing. Owing to insurmountable obstacles the circulation rate could not be determined in this investigation. The volume of air breathed and the respiration rate, however, have been followed in a measure and the observations


indicated that shortly after gassing the respiration was markedly increased but that the volume of air breathed per minute was slightly less than normal. In other words, there was present rapid shallow breathing and undoubtedly less air, hence oxygen was in contact with blood in the lungs during a given period on the assumption that the circulation rate was constant.

Bleeding had no noticeable influence upon either the rate of respiration or the volume of air breathed per minute. Infusion, however, caused a rapid increase in both. Oxygen administered did not exert a marked influence in either particular.

The pulse rate was distinctly decreased soon after gassing but gradually increased to a maximum several hours later. If oxygen was administered during the period of the increasing heart beat, only a slight decrease in the rate was to be noted.


From data collected in the study of the respiratory exchange it may be said that both oxygen consumption and carbon dioxide production were distinctly lowered by exposure to phosgene. Breathing oxygen under these circumstances increased the oxygen consumption. In general, oxygen administration increased the carbon-dioxide production to a slight degree.

Bleeding caused a slight increase in oxygen consumption, although still below normal.

Infusion raised the oxygen consumption and, in general, the carbon-dioxide production. Oxygen administration after infusion brought the oxygen consumption back to the normal level and might indeed carry it above. This should be considered in connection with the percentage saturation of arterial and venous blood. As has been pointed out above, the venous blood carried more oxygen after infusion than before. The administration of oxygen after infusion resulted practically in complete saturation of the arterial blood as well. The oxygen consumption was equal to or greater than normal, while the arterial blood was almost completely saturated, and the venous percentage saturation indicated that the tissues were getting an ample supply of oxygen.

The conclusion is warranted, therefore, that the method of treatment involving venesection, infusion, and oxygen administration is indicated for the reestablishment of normal conditions in the respiratory functions of the blood in an animal gassed with phosgene. It should be emphasized that oxygen administered alone is entirely inadequate to combat the effects of phosgene poisoning inasmuch as this procedure does not eliminate the primary cause, namely, the concentration of blood. When treatment succeeds in restoring blood concentration to a more nearly normal level oxygen administration is of decided benefit.