The purpose of this chapter is to describe the anatomical lesions caused by the poisonous gases as they were actually observed and to present the detailed records of 107 autopsy protocols. The general description of the pathology is based chiefly upon these protocols, although advantage has been taken of experimental data to supplement the description of the skin lesions caused by mustard gas and to consider the possible effect of this agent upon the blood and blood-forming tissues.

Without questioning the fundamental importance of the work done upon animals in establishing the general mode of action of the different gases, it is advisable to point out again the difficulties in applying the knowledge thus gained to the human material.

To begin with, the experimental worker knew the gas that was being used, its concentration, at least approximately, the duration of exposure, and the manner of application, whether by contact or inhalation. The pathologist in the field had to rely upon uncertain evidence in regard to the nature of the gas; in many cases no information of any sort was obtainable. There was always the possibility, too, that the subject had been exposed to more than one kind of gas. As to the concentration and length of exposure, it was only in exceptional cases that even a crude and approximate estimate could be hazarded from the history of the case, as, for instance, where a shell burst at the door of a dugout, and it might be inferred that the inmate had been subjected to a very high concentration for a brief period.

There were other complications which made the interpretation of the cases very difficult. Frequently the soldier was severely injured by the very gas shell which poisoned him, or he may have been gassed as he was lying wounded by another projectile. In either event, the traumatic injuries complicated the picture. For example, it was not possible to distinguish between the terminal pneumonia which is always present in patients dying from infected wounds and the secondary pneumonia consequent upon the gassing.

The most difficult problem of all for the pathologist was to distinguish between the direct effect of the gas and the lesions caused by the secondary or supervening infection of the respiratory tract. Many of the gas casualties occurred during the months of October and November, when the pandemic of influenza was at its height amongst our troops. There will be pointed out in detail, later, the extraordinary resemblance between the respiratory lesions in this disease and those caused by certain of the irritant gases, notably mustard gas- a resemblance which extends to the finest histological detail and which is of considerable interest. Here will be indicated the possible interpretations which often arose in such a case. Were the lesions due to the gas alone? Had an influenzal infection followed upon the injury caused by the gas ? Was it certain at all that the patient had initial gas lesions of the respiratory tract, and could he not simply have acquired an influenzal pneumonia while he was in the hospital being treated for skin burns? The last possibility is indicated in some of our cases in which the respiratory complications developed long after the gassing, but is equally possible in some of the more acute cases in which death ensued.

The experimental work upon animals seemed to establish a clear-cut distinction between such gases as have a vesicant or irritant effect upon the skin and upper respiratory passages, and gases of the edema-producing or suffocant type, such as phosgene and diphosgene, whose action is manifest only upon the parenchyma of the lung and which are without obvious effect upon the upper respiratory epithelium. In the majority of the human cases sufficient data are at hand to permit at least of this general differentiation in regard to the type of gas concerned. In individual cases, however, the matter is not so simple. Extensive pulmonary edema, whether as a result of the gas itself or of the succeeding pneumonia (particularly in the influenzal cases), or as a terminal event associated with the failing circulation, was of common occurrence. In the absence of precise clinical data, and of characteristic changes in the skin, eyes, and upper respiratory tract, the diagnosis might remain uncertain. It was particularly difficult to decide in such cases whether the subject might not have been exposed to a suffocant gas in addition to mustard gas.



In the series of 107 completely studied cases only four may with probability be ascribed to poisoning with this type of gas (Cases 1, 2, 3, 4). Death occurred in all of these cases in 3 days or less after gassing. In addition 2 other cases are ascribed to phosgene in the records, but in these the evidence in support of the diagnosis seems inconclusive (Cases 78, 104). The first of these (Case 78) died 14 days, the other (Case 104) 72 days, after the alleged exposure to phosgene. The difficulties in interpreting these cases will be evident on reference to the detailed protocols and need not be discussed here.

This material is inadequate for a full consideration of the pathology and especially for a study of the late or residual lesions which, on experimental and clinical grounds, may reasonably be expected to follow the inhalation of gases of the suffocant type. The discussion, therefore, is limited to a brief description of a typical acute case. Through the courtesy of the French G.A.C. the writer of this chapter had the opportunity to witness several autopsies upon French soldiers and to study the large collection of histological preparations in the laboratory of Professor Mayer at the College de France, and the following account is based in part upon these experiences, in addition to the records of the series appended and the references in the literature. Reference is made, also, to the observations of Wilder on his series of 50 necropsies of phosgene cases.

All of these patients had passed through the initial suffocative stage of the intoxication and had died one hour or several hours later. Cyanosis was conspicuous in all. In some cases the skin was an ashy gray, in others it was distinctly violaceous. In all cases the blood in the veins was of a deep chocolate colour and very thick. The mouth and nose were usually filled and covered with an accumulation of frothy serous fluid. The lungs were greatly distended and very heavy; they completely filled the thoracic cavity and showed definite imprints of the ribs. The pleural cavities in bodies examined immediately after death contained from 100 cc. to 1,000 cc. of a thin serous or serosanguineous exudate. The surfaces of the lungs were mottled with areas of subpleural emphysema. Cut surfaces of the lungs were extremely wet and bloody from edema and passive hyperaemia. The trachea and larger bronchi were filled with foamy serous or serosanguineous fluid, and pressure on the lungs caused this to flow abundantly. The larynx, trachea, and bronchi were moderately hyperaemic but presented no edema, erosion, or ecchymosis.

In those cases in which death was delayed for two or three days the edema was less homogeneous, hyperaemia was more extensive, particularly at the bases of the lungs, and zones of emphysema occurred at the apices and at the anterior margins. There was also some atelectasis, and if secondary infection had occurred nodules of pulmonic consolidation were in evidence.

Histologically the edematous material seemed to consist of liquid and the debris of epithelial cells. Leucocytes and erythrocytes occurred in late cases only. The cells of the alveolar walls and those of the terminal bronchi appeared swollen and indistinctly outlined, their nuclei staining poorly. Many alveolar capillaries appeared thrombosed and other capillaries were engorged. Still others appeared to be collapsed.


There was a dusky, livid hue to the lips, ears, finger tips, and dependent portions of the body. Thin, blood-stained fluid flowed from the mouth and nostrils, or a mushroom of foam (champignon d'écume) projected from the lips. The opened thorax showed the lungs bulky, meeting in the median line, with very little tendency to collapse when the air was admitted to the chest. They completely filled the pleural cavities, and often showed the imprint of the ribs. The pleural cavities almost always contained several hundred cubic centimetres of blood-tinged serous fluid. The visceral pleura was under great tension, moist, and dotted with petechial haemorrhages or larger reddish splotches. The colour was strikingly variegated, pale pink areas of emphysema alternating with dark red firmer patches or even lobes of atelectasis and haemorrhagic edema. The interlobular septa were wide and translucent from edema. The lung weight was greatly increased, often to 900 or 1,000 grams or even more. It offered a tense, doughy, cushion-like resistance, with a strong tendency to pit on pressure. The lobules were large and clearly defined by the edematous interlobular septa. The trachea and bronchial tree were completely or partially filled with pink fluid and foam and the smaller tubes appeared distended. The mucosa was pink or darker red, velvety, and smooth, and there was no necrosis, ulceration, or membrane formation. The sectioned surface of the parenchyma poured forth fluid and foam in great abundance. The alternation of paler, glistening areas, which under the loop were seen to be composed of distended vesicles, with the dark red collapsed regions, inundated with bloody edema, was very characteristic. Some of the cases showed disruptive emphysema beneath the visceral pleura, the air escaping to the subcutaneous tissue about the neck and shoulders.

The blood was thick and very dark in colour, but its coagulability was apparently unchanged. Thromboses have been described in the vessels of the lung and elsewhere but were not present in any of the cases of the series. The right chambers of the heart were usually greatly dilated and full of soft very dark cruor. Petechial haemorrhages were seen in the epicardium, in the sheaths of the great vessels, and in the gastrointestinal mucosa. The remaining organs, apart from very intense venous congestion, showed nothing abnormal.

No opportunity was presented for study of the central nervous system.


In the early cases the edematous fluid in the alveoli appeared in the sections as a uniform pink staining material or as a granular or shreddy coagulum. There was little or no stainable fibrin. The capillaries were congested and protruded in a tortuous manner into the lumen of the air spaces.

The preparations studied showed no clear evidence of thrombus formation, due to agglutination of red blood cells, to ante-mortem fibrin deposition or to the agglomeration of platelets in the pulmonary capillaries. Wilder, however, refers to the occurrence of thrombi, and the experimental findings of Dunn upon the impermeability of the pulmonary capillaries to injection of saline and carmine gelatine, and the observations of Meek and Eyster upon the agglutination of red corpuscles in vitro when phosgene is bubbled through a suspension, make it seem possible that the obstruction to the pulmonary circulation was not due merely to stasis hut to an actual clumping of the cells under the direct influence of gas. In addition to the intense edema and congestion there was diapedesis of red blood cells, and, in some areas, more extensive alveolar haemorrhage. The alveolar epithelium was exfoliated and in many of the air spaces the capillaries appeared to lie exposed. There was usually an abnormal number of leucocytes in the blood vessels, a few of which emigrated into the alveoli and septal spaces.

Their nuclei were caryorrhectic. The epithelium of the small bronchi might be lifted up by the underlying edema and partly exfoliated. The peribronchial, perivascular, and subpleural lymphatics were distended, usually with a granular or partly fibrinous coagulum. In view of the statement and the currently accepted belief, based upon animal experiment, that the epithelium of the upper respiratory passages is uninjured by phosgene and diphosgene, it was rather surprising to find that in three cases of the series the tracheal epithelium showed a definite lesion. In two of these cases the superficial cells were hyalinized, the ciliated border was destroyed, and there were signs of nuclear degeneration. In the third case there were in addition definite erosions, with local inflammatory reaction. Whether these lesions were due to the phosgene or diphosgene alone or, as seems more probable, to the admixture of more irritant gases, such as chloropicrin or diphenylchlorarsine, or whether they were the result of the early bacterial invasion, can not be decided from the study of the verv limited material. In one of the cases (Case 3) there had already occurred, after 24 hours, a very massive invasion of streptococci in both trachea and lungs.

In the other viscera lesions were not found which could with any degree of probability be attributed to the gassing. The intense venous congestion found in the gross was evident also, of course, in the sections. Parenchymatous changes in liver and kidney may be ascribed to the concurrent infection rather than to a direct toxic action of the gas upon these viscera. There has been general agreement, indeed, that absorption of the gas (phosgene) is unlikely, in view of its rapid decomposition in the presence of water.



The pathology of the skin lesions caused by dichlorethysulphide was very thoroughly studied by Warthin and Weller in this country and by Mayer at the central laboratory of the French medico-legal service at the (College de France, Paris. The studies recorded here on the human cases coming to autopsy have been supplemented by the examination of pieces of skin excised at varying intervals after the application of mustard gas in alcoholic solution of known strength under standard conditions.

The gross changes which followed the application of this substance may be summarized as erythema, followed by vesication and a variable degree of pigmentation. The vesicles were usually superficial; only rarely was the superficial portion of the corium involved. The customary sites for the burns were the face, especially the scalp, the eyelids, nose and lips, neck, axillae, elbows and knees (from lying or kneeling on contaminated ground), the under surface of the penis and the contiguous anterior aspect of the scrotum, the inner surface of the thigh, the buttocks, and the hands and feet.

It is now a familiar fact that after exposure there was a definite latent period which varied from a half hour to six or eight hours or even longer. It was not uncommon for small sudaminal-like vesicles to appear in crops at intervals of days after the exposure. In some cases, presumably those exposed to low concentration of the vapour, there was produced a very diffuse, almost scarlatinaform erythema, changing in colour from a somewhat dusky pink, through purple, to brownish. In other cases the vesicles were surrounded by an erythematous zone, which later became brownish. The duration of the pigmentation was not definitely known and probably varied in different individuals. Pigment flecks at the site of a small experimental burn were apparent after a year. The contents of the vesicles were usually clear, although the fluid might contain a few filmy fibrin clots. If infection occurred the fluid might become purulent, but ordinarily relatively few leucocytes were found on examination of the fresh fluid. It was frequently reported that a burn over the sacrum had developed into decubital ulcers.

Since deaths usually occurred from the fifth day to the end of the third week after exposure, the pathologist was apt to see only raw surfaces, blebs, ulcers, scars, or pigmentation, rather than erythematous lesions. These were also more apt to be in locations subjected to long friction during travel to the hospitals and thus, to a certain extent, the regional distribution of the burns, when tabulated from autopsy reports, differed from that of the lighter burns sustained by those who recovered.

The earliest lesion examined histologically was excised 3 hours and 25 minutes after the application of a 10 per cent solution of dichlorethysulphide in absolute alcohol, applied for 5 minutes over an area of 1 mm. in diameter in the skin of the back; at the time of excision there was a slightly elevated area of erythema about 3 mm. in diameter.

The only change noted in the epidermis was a slight thinning with a flattening out of the papillary processes; over the summit of the lesion they were completely obliterated. Under the high power a few individual cells showed vacuolated cytoplasm with displacement of the nucleus to the periphery of the cell. There was slight edema of the corium; the sweat glands, hair follicles, and sebaceous glands showed no recognizable injury; the blood vessels, especially the superficial capillaries, were dilated and filled with red cells and a few were surrounded by a loose mantle of lymphoid cells; the lymphatic vessels were distended, containing a fine granular coagulum; the deeper portion of the corium and the subcutaneous tissue showed no injury.

Another lesion was examined after 45 hours, having been produced at the same time, in the same individual; the lesion consisted of a small elevated vesicle filled with clear fluid and surrounded by an area of erythema 4 to 5 mm. in width.

The overlying superficial horny layer was still intact, but the remainder of the epithelium covering the vesicle was entirely necrotic. The contents of the vesicle consisted of interlacing fibrin strands inclosing a homogeneous coagulum loosely infiltrated with polynuclear leucocytes. At the margin of the vesicle the epithelial cells were dissociated for a short distance but rapidly became normal. No mitoses were found and there was no indication of increased proliferative activity. The base of the vesicle was formed by edematous corium which, in its most superficial portion, was devoid of connective tissue nuclei and was loosely infiltrated with round cells and polynuclear leucocytes. These were most densely aggregated about the blood vessels. The deeper connective tissue showed no edema or inflammatory reaction.

The perivascular infiltration extended for a considerable distance beyond the area of the vesicle; the sweat glands and hair follicles in the vicinity of the lesion showed no significant changes.

Although it would appear from a study of this experimental lesion that the vesicle was formed between the epidermis and the superficial corium, a study of numerous other accidental lesions showed that this was by no means always the case; the vesicle may be formed within the epidermis itself by a dissociation of the epithelial cells with the accumulation of fluid between them and their subsequent necrosis. In this way a cleft may be formed, the epidermis and some of the basal cells remaining visible.

Not only might fluid accumulate between the cells, forcing them apart, but the cells themselves might undergo hydropic changes, so that the cytoplasm would contain a single large vacuole which pushed the nucleus in crescentic form to the periphery of the cell. This type of hydropic degeneration was frequently seen at all stages, especially amongst the less severely injured cells at the periphery of the lesion.

In other cases individual cells, or the entire epidermis, would undergo a hyaline necrosis, the nucleus becoming shrunken and pycnotic and the cytoplasm taking on a dense retractile appearance and staining deeply with eosin. This mummified or hyalinized epidermis, constituting often the cap of the vesicle, might persist for a long time, showing in a shadowy way the outline of the individual epithelial cells. It might persist until a new growth of epithelium proceeding from the margin of the vesicle had completely undermined it and recovered the base of the vesicle. This was one of the common methods by which regeneration took place. There was no doubt that the sheaths of the hair follicles also played an important part in reinvesting the base of the vesicle or ulcer, although this was even more striking in the lesions experimentally produced in horses.

The opportunity was presented to examine histologically experimental lesions produced in the way indicated above in Negroes. Although it has been demonstrated by the studies of Marshall, Lynch, and Smith, cited below, that the Negro is relatively insusceptible to dichlorethylsulphide, the lesion produced by a 10 per cent alcoholic solution differed in no respect from the lesions in white individuals. One of the preparations excised after six hours showed very clearly the early edema in the papillary layer of the corium which preceded vesiculation. The tissue was very loose and foamy, and occasionally the entire epidermis was elevated by a granular coagulum. The same type of degeneration of the cytoplasm was found in the connective tissue and endothelial nuclei. There was hyperaemia and a moderate inflammatory reaction in which polynuclear leucocytes played the leading part.

Many of the preparations obtained from autopsy cases showed definite evidence of infection, with the typical inflammatory response. Since these may be regarded as incidental lesions and are in nowise distinctive they need not be considered in detail.

The pigmentation which was so striking a feature of the later stages was due to an increased production of melanin pigment by the cells of the rete mucosum and was not to be attributed to the deposition of blood pigment following capillary haemorrhages. The pigment production was often irregular, individual cells being loaded with coarse clumps, while adjacent cells might be wholly pigment free. Moreover, pigment-containing cells could be present in abnormal situations-in the stratum granulosum, for example, or even amongst the cells of the keratin layer. Usually, numerous chromatophores were seen in the capillary layer of the corium.

Where the burn had been a severe one, and the subepithelial tissue had been involved, definite fibroblastic growth was found in the later stages. The leucocytes which, in the early stages, were chiefly polymorphonuclear, and which tended to become fragmented as they approached the surface, gave way to lymphoid cells; but at no stage of the process was the cellular reaction, in uninfected cases, a very intense one.

The vascular lesions throughout appeared to be of minor importance. Thrombosis was exceptional, being found only in very severe burns in which there was a direct necrosis of the corium. These findings, therefore, are not in accord with the view of other observers, who hold that the injury to the blood vessels is an important factor in delaying the repair. The persistence of the injury and the retarded healing appeared to be due, rather, to the continued action of the dichlorethylsulphide.

To summarize briefly the lesions produced in the skin: There was caused a hydropic degeneration, or a hyaline necrosis of epidermal cells, often with the formation of vesicles within the epidermis. There was produced, also, edema of the corium beginning in the loose tissue of the papillae and leading to the separation of the epidermis from the underlying connective tissue and the formation of the vesicle.

In a few of the cases the lesions of the upper respiratory tract were gangrenous rather than diphtheritic in character. The walls of the trachea and bronchi showed Greenish discolouration. and the exudate had a characteristic foul odour. There were usually associated gangrenous areas in the lung itself. In two such cases the pathologist recorded the presence of extensive dental caries, indicating that, in his opinion, infection of the air passages had occurred, with the putrefactive bacteria from the mouth. This seemed a plausible idea, and in one of the cases it was supported by the finding of numerous fusiform bacilli in the necrotic exudate. Spirilla were not demonstrated but were doubtless present.

The histological lesions found in the trachea and bronchi were studied in a severe case, dying two days after exposure. The trachea and large bronchi presented about the same picture . There was a thick, fibrinous membrane with coarse laminated threads running parallel to the surface. In the meshes lay scattered nuclear fragments and a few better preserved wandering cells; masses of mucus, also, sometimes showing a curious concentric arrangement, were incorporated in the fibrinous membrane. Adherent to the under surface were the detached epithelial cells, many of which were surprisingly little altered, still conserving their cilia and showing good nuclear stain. The membrane lay loosely upon the exposed basement membrane, being attached only here and there by bridges of fibrin. The basement membrane itself was swollen and less sharply outlined than the normal structure. The subepithelial tissue was edematous: in places there was a definite fibrinous exudate between the connective tissue bundles. There was a moderate inflammatory infiltration with both mononuclear and polymorphonuclear leucocytes. As these approached the surface they appeared to undergo caryorrhexis, and in the superficial portion there was much nuclear debris. The nuclei of the connective tissue cells also showed the effects of the injury, their nuclei being shrunken and pycnotic. The blood vessels were enormously distended, and there were small capillary extravasations. The endothelium showed no definite alteration . The mucous ducts were widely distended with plugs of mucus. Near the surface their epithelial cells were apt to be cast off and more or less degenerated.

In the deeper portion the cells lost their orderly alignment and tended to take on a squamous type. The mucous glands themselves at this stage were in a state of hypersecretion; only except where there was a layer of flattened cells showing many mitotic figures, often highly atypical in character, interposed between the fibrinous exudate and the basement membrane. Often it seemed as if the epithelial growth had outstripped the other processes of repair, so that the growing cells themselves. because of being improperly nourished, degenerated. Some of the preparations showed the new epithelium lifted up from the membrana propria by a granular coagulum, as if a new vesicle had been formed. This appearance leads one to ask whether there may not have been a persistence of the poison and whether this late and secondary vesicle formation may not be comparable to the late appearance of skin vesicles four or five days after gassing.

One of the preparations showed the duct epithelium growing beneath the still preserved basement membrane. This anomalous growth was due to the proliferation of the duct epithelium beneath the partially loosened membrana propria. In one of the late cases, scattered through the subepithelial tissue, were solid nests of cells which resembled very closely squamous celled carcinoma. The cells were arranged as in epitheliomatous pearls; they were definitely squamous, highly atypical, often multinucleated, and showed numerous mitotic figures. Intracellular fibrils even were to be seen between them. This appearance was brought about by the proliferation of the duct epithelium within the lumen.

As regards the lesions in the submucosa, it has been stated above that in the majority of cases the cellular response was not very marked, and in some cases, in the trachea at least, practically wanting. Bacterial stains showed that the organisms rarely penetrated below the surface. In one of the cases, however, in which there was an extreme edema not only of the submucosa tissue but also of the areolar tissue external to the bronchus, a Gram stain showed unrestrained growth of Gram-positive cocci throughout the edematous area.

With the beginning of repair the connective tissue cells in the submucosa took on the character of fibroblasts. New blood vessels were formed which were sinusoidal in character and which formed wide channels extending to the basement membrane, where this was still intact. The inflammatory cells then became predominatingly lymphoid in character, and numerous plasma cells and other large mononuclear elements were present. This formation of a very vascular granulation tissue was found not only in the trachea and large bronchi, but in the medium-sized and smaller bronchioles, where it sometimes led to a pronounced thickening of the wall, with narrowing of the lumen. As we have stated above, the granulation tissue might or might not be reinvested by epithelium according to whether the initial injury involved the epithelium of the mucous ducts or not.

In a few of the cases in which the mucous glands were injured it was possible to study a regeneration of these structures. The sinus was filled with a pink-staining, more or less hyaline, necrotic mass, embedded in which were nuclear particles. This represented the remains of the necrotic gland cells, and at the periphery the new secretory epithelium was seen pushing its way between the necrotic cells and the basement membrane of the gland.


The appearance of the lungs at autopsy presented a bewildering variety, and it is quite impossible to describe a composite picture which would distinguish the lungs of mustard-gas poisoning from those of the various types of pneumonia. Perhaps the most distinctive cases were the very early ones, in which the pulmonary lesions were largely confined to the bronchi and their immediate neighbourhood. The lung was voluminous and did not collapse readily after removal. Occasionally the pleura was smooth, but in most cases there were patches of early fibrinous pleurisy. Darker sunken area of atelectasis associated with the occlusion of bronchi were visible on the surface. On section the most conspicuous features were the thickened bronchi filled with plugs of creamy fibrinopurulent exudate, or in the case of the smaller bronchi a droplet of creamy pus might exude. Each bronchus was surrounded by a dark red sunken areola 2 or 3 mm. in width. This last feature is regarded as particularly characteristic; the red peribronchial zone histologically was found to be composed of the adjacent alveoli, which were filled with red blood cells and more or less collapsed. Further outward from the bronchus the haemorrhage gave place to a fibrinous exudate in which only a few desquamated alveolar cells and occasional leucocytes were included. This peribronchial reaction did not appear to be due to a direct extension of the infection within the bronchus through the bronchial wall. It would seem that the haemorrhage was caused by the direct action of the toxic agent diffused through the wall of the bronchus. The collapse may perhaps be explained by the distension of the small bronchus with compression of the adjacent air spaces. Frequently, at least, the alveoli directly adjoining the bronchus appeared flattened.

Aside from these bronchial and peribronchial lesions the lung, on section, showed emphysema and darker areas of partial collapse. Edema was present to a greater or less degree in about two-thirds of the cases. It was never so extensive as in the lung of phosgene or other asphyxiating gas and was often patchy in its distribution, being much more marked in some portions of the lobe than in others. From the autopsy records, which naturally vary greatly in detail and accuracy, it appears that excessive edema was noted in only three cases occurring two and three days after gassing, and in one case in which the period of survival was not established; moderate general edema was present in 28 cases and slight patchy edema in 43 cases. In 18 cases the absence of edema is specifically reported, and in 6 cases no record was made.

As with so many of the pulmonary lesions, it was difficult to decide whether this edema was the direct result of the action of the mustard gas upon the alveolar capillaries or whether it was due to secondary infection, to the failing circulation, or to other obscure factors. It is true that even in the early cases bacteria abounded in the bronchial and infundibular lesions. They were not usually present, however, in the edematous areas at a distance from the bronchi. This, and the fact that the most intense edema was recorded in relatively early cases, is perhaps an argument in favour of the direct edema producing action of the mustard gas.

The great majority of the cases showed, in addition to the bronchial and peribronchial lesions, areas of focal pneumonia, sometimes small and nodular, often large and confluent. There was no constancy in the appearance of these pneumonic patches, nor could one find sharply cut differences in the appearances in cases dying early after gassing and in those which survived for weeks or even months. A fresh bacterial infection might develop at any stage after the initial chemical injury, and it was not uncommon to find in the same lung recent and older organizing lesions.

During the height of the influenzal epidemic in October and November, 1918, many of the gas pneumonias exhibited the characteristic gross features which had come to be associated with the pulmonary lesions of influenza. The lungs were heavy and voluminous and often a dusky red, especially in the posterior portions. There were fresh pleural haemorrhages and more or less fine fibrinous pleurisy which became organized in the later cases. Only twice was the pleurisy suppurative in character; in one of these cases (Case 99), the empyema was due to the extension of a traumatic liver abscess; in the other (Case 25), which occurred in August, the pleurisy was not extensive, and the resemblance of the pulmonary lesions to those found in influenza were less striking than in many other cases.

On section these lungs showed, in addition to the characteristic membranous bronchitis and bronchiolitis and peribronchial areola, diffuse, often incomplete areas of consolidation associated with much haemorrhagic edema. The greater portion of one or several lobes was frequently affected. In some of the later cases there are described also opaque grayish areas of necrosis and groups of small abscesses. Organization was not infrequently recognized in the gross by some of the more experienced pathologists and confirmed histologically.

In describing the histological changes associated with this type of pneumonia attention is called again to the extraordinary resemblance of the finer changes to those observed in the primary influenzal cases. These lesions can not be ascribed wholly to the influenzal or postinfluenzal infection, in as much as they may be reproduced in animals by exposure to the gas alone, and have been present also in human cases dying at times when the epidemic was not active.

Leaving aside the larger bronchi, which have been described and in which the membranous character of the necrosis is usually more extreme than in the primary influenzal cases, the terminal bronchioles were found usually with their ciliated epithelium more or less conserved. There was an acute suppurative inflammation, which was in nowise distinctive. The atria, however, were often widely dilated; they might or might not themselves contain exudate. The wall was lined by a wavy hyaline band, which sometimes, but not regularly, gave a faint fibrin stain with the Gram-Weigert-safranine method. It was a little difficult to be sure of the composition of this hyaline band. It did not appear to be composed solely of the necrotic lining epithelial cells, although these probably took part in its formation. The continuity of the membrane with definite bands of fibrin in the walls of the atria or alveoli indicated that the fibrin was at least the chief constituent. Later, as it became more and more swollen and hyalinized, the specific staining was less readily obtained. In part it might have been due to the condensation of a highly albuminous material about the wall of the infundibulum.

The lesions found in the alveoli were manifold. The alveolar exudate in the vicinity of the bronchioles and atria was apt to be fairly cellular, containing, in addition to a variable number of red cells, leucocytes of various types, but predominantly polymorphonucleal. But there were large areas in which the exudate was characteristically poor in nucleated cells and was rather of the nature of a haemorrhagic edema. The fluid either contained a loose fibrin net or appeared merely as a homogeneous or fibrinous coagulum. There would be merely diapedesis of red cells or more profuse haemorrhages, leading to disruption and necrosis of the lung tissue and distinguishable from infarcts only by the absence of thrombi within the larger arteries. The alveolar capillaries in these portions of the lung were tremendously distended, bulging into the alveolar spaces. Sometimes it appeared as if the membrane upon which the endothelium rests was swollen and thickened. Fibrin thrombi were fairly often found within the capillary lumen, and in some cases were quite abundant. It was very common also to see coarse fibrin threads deposited in the septa between the capillary wall and the epithelium. Edema of the alveolar wall was often striking, and the still intact alveolar epithelium could be elevated at times as a continuous sheet of cells.

The alveolar epithelium itself in the early acute cases often appeared swollen and vacuolated, although it was only occasionally that degenerative changes could be recognized in the still adherent cells. There was always more or less exfoliation, the cast-off cells becoming rounded, taking up red cells, pigment granules, leucocytic nuclear fragments, losing their nuclear staining eventually, and becoming degenerated.

Necrosis of the capillary wall was observed repeatedly, with fragmentation of the endothelial nuclei and of the leucocytes, as recently described by Le Count in cases of primary influenzal pneumonia, and regarded by him as highly characteristic

Just as the bacteriological studies in influenzal pneumonia showed in the lung a varying flora, so the bacteriological data in the series of gas pneumonias studied failed to throw any light on the difference in anatomical types. Both culturally and in bacterial-stained sections a variety of forms was found. In a few cases, and these unfortunately were uncontrolled by cultures, the sections showed enormous numbers of minute Gram-negative rods as the predominating organism in the alveolar exudate. Many other cases, grossly and histologically similar, showed only Gram-positive cocci. So confusing were the findings that it seemed not worth while to attempt an analysis, particularly as no systematic study could be carried on.

The regenerative changes in the trachea and larger bronchi have already been described. Although bacteria were always present, the destruction produced by the chemical irritant was of so gross a character that the role of the bacteria may well be regarded as altogether secondary, particularly as they rarely invaded the deeper tissues. The healing process also may be looked upon as a repair of the chemically injured tissue. But in the parenchyma of the lung, where the original chemical injury was lost or overshadowed by the bacterial infection, it became quite impossible to say whether the reparative and organizing processes which were present in the later cases were in response to the chemical or the bacterial poison. But it was these late and permanent changes,

however brought about, which were of the greatest practical interest, and it is necessary to describe them in some detail in order to form an approximate idea of the damage which may be expected to ensue upon the gassing.

The thickening of the walls of the small bronchi was an alteration which was quite evident in gross sections of the lung. The section showed the bronchial wall replaced by a vascular granulation tissue, with lymphoid and plasma cells predominating. Where the epithelial lining had failed to regenerate this granulation surface lay exposed, and it can not be doubted that the further contraction of this tissue would lead to narrowing or complete occlusion, with the formation of bronchiectasis distal to the stenotic area.

Bronchial stenosis, also might result, though apparently not so frequently as one might expect, from the organization of the exudate within the bronchiole. This seemed to occur more often in the ductus alveolaris, where the organizing fibrinous plugs extended into the contiguous alveoli.

The zone of atelectasis, haemorrhage, and fibrinous edema which so often encircled the medium-sized and smaller bronchi has been described. In this zone, where the fibrin offered support to the growing cells, active organization was regularly found in progress, even when the pneumonic process at a distance from the bronchi was in full blast. It was in this peribronchial zone also that there was found, as in the later stages of the primary influenzal pneumonias, a remarkable proliferation of the alveolar epithelium. The new cells, distinguished by their deeper staining, their cuboidal form, and often by numerous kinetic figures, not only reinvested the cavities of the alveoli but grew over and even into the plugs of fibrinous exudate, and in some instances formed solid, carcinoma-like nests of cells. The alveolar septa also were thickened by the growth of fibroblasts, the new-formed connective tissue being continuous with the granulation tissue about the small bronchi, and also with the new-formed fibrous tissue which invaded the edematous tissue about the arteries and interlobular septa.



The study of human cases afforded little evidence that mustard gas, through cutaneous application or by inhalation, could be absorbed and could produce systemic effects in other organs. Clinically there has been observed a group of cases with diffuse pigmentation, marked apathy and asthenia, low blood pressure, and often pronounced mental disturbances, symptoms which it is difficult to correlate with the obvious lesions of the skin and respiratory tract. Indeed, in some of these cases, respiratory damage may be quite insignificant. Such cases certainly suggest a systemic intoxication of some sort, and Satre and other French clinicians have attributed the symptoms to an acute adrenal insufficiency. No anatomical evidence has been brought forward to incriminate the adrenals, and the observations are not sufficiently definite to justify any far-reaching conclusions. The lassitude and asthenia, and even the mental disturbances, in men freshly returned from combat, are perhaps more easily explained in other ways.

More direct proof of the absorption of dichlorethysulphide are the changes in the bone marrow and in the circulating blood. Zunz, Stewart, and later Krumbhaar have shown that severely gassed cases, after an initial leucocytosis, may develop a marked leucopenia. Pappenheimer and Vance have also shown that intravenous injections of small doses in rabbits brought about a profound leucopaenia, with destruction of the granulocytes in the bone marrow, an effect comparable in its specificity with that of benzol. This has been confirmed by Warthin and Weller, Krumbhaar and Krumbhaar have brought confirmation of their clinical evidence by a study of the bone marrow in human cases. It may be taken as proved, therefore, that when introduced into the body dichlorethylsulphide is a specific poison for the haematopoietic tissue, and there are both clinical and experimental reasons for believing that a similar effect follows the inhalation of massive doses.



Abdominal tenderness, anorexia, nausea, vomiting, and less frequently diarrhoea were observed clinically in a large proportion of mustard-gas cases. The experiments of Warthin and Weller, Lynch, Smith, and Marshall, Pappenheimer and Vance, and others, have shown that the intravenous or subcutaneous injection of dichlorethylsulphide in animals may be followed by a haemorrhagic enteritis. Norris refers the submucous haemorrhages of the stomach and duodenum to the swallowing of contaminated saliva, but the possibility of the elimination of the absorbed substance or its hydrolysis products through the alimentary tract has not been disproved. This possibility is suggested on clinical grounds be Ramon, Petit, and Carrie.

In the human protocols studied, the alimentary tract was not examined in about half of the cases; in about 25 per cent of the cases the stomach and intestines are specifically stated to be normal; in the remainder there are noted injection or hyperaemia, sometimes of the stomach, more commonly of the small intestine; haemorrhages in stomach, small or large intestines, erosions or small ulcers, in three cases in the stomach, once in the small intestine, once in the colon.

Too much weight should not be placed on these fragmentary references. The injection may be attributable to the general visceral stasis, which is the rule in the fatal cases. Haemorrhagic erosions of the stomach are so frequent a finding in any large series of autopsies that their occasional presence in these cases does not seem very significant. On the other hand, a more careful scrutiny might have shown a higher incidence of gastrointestinal lesions.


No changes were found apart from those common to all acute infections. Patients dying early have not shown degenerative changes which might be ascribed to a specific effect of the mustard gas.


No anatomical alterations of the myocardium were found which might throw light on the late circulatory disturbances (effort syndrome) noted in a certain proportion of gassed individuals. Dilatation of the right side of the heart is not infrequently mentioned and may be regarded as secondary to the pulmonary lesions. It was not always evident at autopsy. Suppurative pericarditis as a secondary infection with hemolytic streptococci is recorded in one case. Nor was there found a typical or characteristic vascular lesion in the lung or elsewhere, which could not be of infective origin. That the frequent occurrence of rather marked pulmonary edema in the severe early cases implies an alteration in the permeability of the blood vessels may be assumed, but the cellular changes in the endothelium are not sufficiently definite to warrant description.


Apart from incidental lesions, obviously antecedent to the gas poisoning, these organs showed only an intense venous and capillary congestion. In one case only were there haemorrhages into the capsular spaces and tubules. No alterations which would suggest a toxic effect upon the renal epithelium were noted. The intense congestion is probably sufficient to explain the diminished urinary output, albuminuria, presence of casts and red blood cells described by Hermann as typical urinary findings during life.


Unfortunately there were available no data upon possible finer alterations in the central nervous system. In very few cases was the brain examined, and in none of these were significant gross changes detected, nor was material for histological study preserved. Stewart has described ring haemorrhages in the brain associated with swelling and degeneration of the endothelium, thrombosis, and slight leucocytic emigration.


No facts of value have been deduced from the incomplete study of the adrenal gland and other organs of internal secretion which have been occasionally included in the material studied.


Return to Medical Aspects Of Gas Warfare or to the WWI Medical Front index