William H. Welch, MD, and his colleagues performed an autopsy at The Johns Hopkins Hospital in October 1891 on a 38-year-old man and discovered a new bacterium, Bacillus aerogenes capsulatus. During the postmortem examination, gas bubbles were noted within many of the patient's blood vessels. Welch's laboratory personnel determined that a previously unknown bacterium was the source of the gas. Through a series of experiments, the organism's characteristics were described and its pathophysiology was detailed, findings that proved accurate in explaining gas gangrene during World War I. Welch never followed up these initial investigations with more experimentation. His subsequent writings regarding the bacterium that came to be known, appropriately, as Bacillus welchii consisted mostly of case reports from other medical institutions and summaries of previous data.

William H. Welch, MD (Figure 1), became director of the pathology laboratory of The Johns Hopkins University and Hospital in 1886. Although the hospital would not open for another 3 years, and the medical school for another 6 years, Welch and a small cadre of students began performing research. Clinical duties, such as the performance of autopsies, were added to Welch's responsibilities after the opening of the hospital. The first autopsy was performed by Dr Welch on May 28, 1889. According to Ella H. Oppenheimer, MD, a pathologist who followed Welch at Johns Hopkins, Dr Welch performed approximately the first 300 autopsies.1 The emphasis on research during Welch's tenure at Johns Hopkins before clinical responsibilities began created an atmosphere of investigation in which interesting autopsy cases were studied with controlled experiments. In this article, we describe the case of an autopsy performed and reported by William Welch that led to the discovery and characterization of a new bacterium, perhaps the most significant discovery to result from Welch's collaborative and experimental environment.

Figure 1.

William Welch, MD, sitting at his desk at Johns Hopkins. Note the picture of Sir William Osler over Welch's right shoulder

Figure 1.

William Welch, MD, sitting at his desk at Johns Hopkins. Note the picture of Sir William Osler over Welch's right shoulder

Close modal

In October 1891, Welch performed an autopsy on a 38-year-old man, J.M., who had died on the service of William Osler, MD, 8 hours earlier. The patient initially presented to the hospital in October 1890 with “a large, hemispherical, pulsating tumor in the right infraclavicular and mammary regions” that gradually increased in size.2 His past medical history was significant for alcoholism, syphilis, and tuberculosis. J.M. was discharged in July 1891 and experienced no problems until a month prior to his death, when he observed a small ulcerated opening over the tumor. He noted 3 hemorrhagic episodes from the ulceration, each occurring approximately 1 week apart from each other. The patient was readmitted to Dr Osler's service and his condition did not appear alarming. J.M. died suddenly without further hemorrhage. Significantly, the physicians caring for him did not observe any emphysematous swelling of the neck or escape of gas from the ruptured openings of the tumor throughout the patient's final hospital course.

Eight hours after death, the body still warm and the weather cool, Welch performed the autopsy. He ultimately diagnosed the patient with chronic pulmonary tuberculosis, acute miliary tuberculosis, and a large sacculated aneurysm of the ascending arch of the aorta, which ruptured in 2 places through the anterior thoracic wall.2 The body also exhibited extensive emphysematous crackling with diffuse, symmetric swelling. There was near circumferential involvement from the lower portion of the neck to the superior border of the thyroid cartilage. Moreover, areas including both axillae, the inner side of the arms, over the pectoral muscles, over the buttocks, in the groin, and on the inner side of the thighs demonstrated the same crepitus.

Superficial veins along most of the body and extremities also showed emphysematous crackling. Gas bubbles were noted to be escaping from 2 circular openings through the skin over the tumor on the anterior, right side of the thorax. Welch and his associates dissected out the jugular, femoral, and brachial veins; the superficial veins of the abdominal wall; and the femoral and temporal arteries. The presence of gas bubbles was determined in all of them without making an incision into the vessels' lumens. In an effort to determine the nature of the gas, a flame was placed over the open vessel lumens. The gas (presumably with hydrogen as a component) made a slight explosive sound at the moment of ignition and burned pale blue.

The large aortic aneurysmal sac producing the tumor responsible for the patient's hemorrhage was exhibited. The walls of the aneurysm were intact, excluding the 2 openings through the skin. Within the sac, a thrombus was found that contained gas bubbles. Examination of the heart revealed gas bubbles within the pericardial sac, the cardiac arteries and veins, the right and left atria and ventricles, the pulmonary arteries and veins, and the myocardium. In addition, the aorta was filled with thin, watery, transparent blood with gas bubbles. Pressure placed on the surfaces of the spleen, kidneys, and liver led to the seepage of gas bubbles. The liver had crepitus with large numbers of gas bubbles within the lumens of its vessels. Notably, the peritoneal cavity was dry and was not distended with gas.

Microscopic examination of the patient's blood and organs revealed bacilli. Blood from the heart and vessels was reported as being rich with bacilli. Welch described the bacilli as “about 3 to 5 mm in length, about the thickness of anthrax bacilli, with ends slightly rounded, sometimes almost square-cut, occurring chiefly in pairs and in irregular masses and not in long chains.” 2 No other species of bacteria were observed on any of the slide preparations made from this patient's tissues. The bacilli were not motile, they possessed a capsule, and they stained with aniline dyes.

Frozen and alcohol-hardened sections of the organs were prepared. Sections of the liver showed small cavities, visible to the naked eye, that were surrounded by liver cells and that were much disintegrated with fatty degeneration. The bacilli were present throughout the liver; they were most abundant in and near the small cavities.2 Furthermore, inflammation was seen within the liver parenchyma along the rim of the cavities (Figure 2). The myocardial sections also showed the same cavities with many bacilli and disintegrating muscular fibers containing fine fatty granules in their walls. Sections of the kidney showed darker areas that were more opaque than the rest of the tissue, corresponding to groups of perhaps a dozen cortical tubules. In these areas, the epithelial cells were finely fatty and desquamated. There were also many bacilli here, whereas in other areas the epithelium was not altered and the bacilli were fewer.

Figure 2.

A, Liver. A cavity formed secondary to gas produced by Bacillus welchii is on the left and is surrounded by necrotic liver tissue. An inflammatory cell response lies between the normal-appearing liver on the right and the necrotic zone. The inflammation shows that the organisms had spread into the liver during the patient's life (hematoxylin-eosin, original magnification ×64). B, Bacillus aerogenes capsulatus. The dark blue–stained bacteria are located within the cavities as Welch described: cylindrical, with rounded ends, singly or grouped in pairs, and not in long chains. The rounded structures are laked red cells (hematoxylin-eosin, original magnification ×400)

Figure 2.

A, Liver. A cavity formed secondary to gas produced by Bacillus welchii is on the left and is surrounded by necrotic liver tissue. An inflammatory cell response lies between the normal-appearing liver on the right and the necrotic zone. The inflammation shows that the organisms had spread into the liver during the patient's life (hematoxylin-eosin, original magnification ×64). B, Bacillus aerogenes capsulatus. The dark blue–stained bacteria are located within the cavities as Welch described: cylindrical, with rounded ends, singly or grouped in pairs, and not in long chains. The rounded structures are laked red cells (hematoxylin-eosin, original magnification ×400)

Close modal

Dr Welch first reported discovery of the new bacterium in November 1891 at a meeting of The Johns Hopkins Hospital Medical Society,3 and by August 1892, Welch and Nuttall had published a full report of the case, their subsequent research, and conclusions.2 The new bacterium was named Bacillus aerogenes capsulatus, despite the fact that a capsule was not consistently present in all observed bacteria.

It came to be known, fittingly, as Bacillus welchii. Other possible names Welch considered were Bacillus sanguinis aerogenes, Bacillus aerogenes, Bacillus aerogenes cadaveris, and Bacillus pneumathaemia.

Welch and Nuttall concluded that the bacillus developed rapidly in the blood after death, forming gas, but was not pathogenic under ordinary conditions. Doses of up to 2.5 mL of fresh bouillon cultures could be injected into healthy rabbits under ordinary conditions without resulting mortality. The lone unexpected condition involved a pregnant rabbit; both the mother and the 2 embryos died soon after injection. Welch later noted that the pregnant rabbit contained dead embryos at the time of injection.2 The bacillus and gas formation were found to be abundant in the blood vessels and organs of rabbits if the animals were killed immediately or soon after intravenous injection of bacteria. In contrast, if the rabbits were killed several hours to 2 days after intravenous injection of bacteria, then a longer time was required before bacilli and gas developed in vessels and tissues. Welch and Nuttall hypothesized that the gas found in the vasculature and organs was not atmospheric and must have been produced by the bacillus.

In concluding his first article,2 Welch asked 3 questions that he could not fully answer. Were these bacilli distributed throughout the circulation before the death of the patient? At what point did the bacilli enter the circulation? Were they concerned in causing death? Based on his experiments, Welch affirmed the first question. Deceased rabbits injected with the bacillus culture required a much greater time period for gas to spread throughout the vessels and organs than when the injected bacillus was carried throughout the body by the circulation prior to death. A mere 8 hours passed between the patient's death and the autopsy, not enough time for the bacteria to multiply and release copious amounts of gas. Welch allowed that special conditions in this patient, such as anemia secondary to hemorrhage and tuberculosis, may have led to decreased oxygen in the body, therefore providing the anaerobe with a suitable environment in which to grow in the body prior to death.

Regarding the second question, Welch hypothesized that the bacilli entered the patient's circulation through the aneurysmal sac's external openings in the chest wall. The thrombus discovered within the aneurysmal sac at autopsy was riddled with gas and may have provided an environment of decreased oxygen for foci of bacteria to grow and seed the body. The experiments by Welch and Nuttall demonstrated that the bacillus could grow in fluid cultures with Buchner's method of anaerobic culture before the oxygen was completely absorbed.

Indeed, the clot possessed an “enormous number of bacilli.”2 Although he could not exclude the possibility, no other site in the patient's body was found that offered a site for the initial infection by the bacterium.

Based on his experiments, Welch knew the bacilli were pathologically impotent in a living organism. Therefore, could the bacilli have killed this man? The patient had other conditions sufficient to cause death. Welch hypothesized that the presence of dead tissue, thrombi, or cavities like intestines or the uterus might allow a reduced oxygen concentration. In such a protected environment, the bacilli might grow and gas could accumulate until a sudden release of air into the vasculature caused death, a process described previously.4 Regarding the pregnant rabbit with dead embryos, Welch noted “the extent of the development of the bacillus and of gas was much greater than is observed at so short a period after death (six hours) when the animal is killed immediately or shortly after intravenous injection of the cultures and is kept at the temperature of the room.” 5 

This observation foreshadowed conclusions of the United States Surgeon General's report on the US Army's medical preparedness and response to medical problems during World War I. “They [anaerobic bacteria, of which B welchii, by 1929 called Clostridium welchii, claimed first place as a pathogen during the war] thrive on decaying animal and vegetable matter, but not on living tissue, and through their resistant spores are able to retain their vitality through long periods of conditions unfavorable to their multiplication.” 6 Highly explosive shells produced wounds well-suited to the bacteria's growth and led to gas gangrene. Bacillus welchii could only create such pathology if tissues were first injured by some chemical or mechanical means, be it either a battlefield injury or dead rabbit embryos. We can only theorize that the sacculated aortic aneurysm in Welch's original case offered a similarly hospitable environment for this bacterium.

In their biography of William Welch, Simon and James Thomas Flexner stated that Welch's discovery of B welchii opened a new chapter in medicine and established a new field of investigation, namely, pneumopathology.7 Unfortunately, Welch never performed original experimental work on his bacilli—arguably his most important scientific contribution to medicine—after the publication of his initial investigations. Welch, however, continued to write and lecture on the discovery that provided him with such a significant opportunity.5,8 Although prolific, these articles lacked original contributions to the scientific understanding of B welchii. An article published in 1896 by Welch and Simon Flexner, for instance, offered a review of the Welch and Nuttall paper and updated work done by other scientists since its publication. Then, the authors reviewed a series of 23 cases involving the new bacterium. These cases were drawn from the services of Dr William Osler, William Halsted, MD, and Howard Kelly, MD, at The Johns Hopkins Hospital, although cases from the University of Maryland and as far away as Western Reserve Medical College in Ohio were represented. No new experimental data were presented.

Oppenheimer
,
E. H.
Preface.
In: Index Books for the First and Second Thousand Autopsies. Baltimore, Md: The Johns Hopkins University School of Medicine; 1926
.
Welch
,
W. H.
and
G. H. F.
Nuttall
.
A gas-producing bacillus (Bacillus aerogenes capsulatus, Nov. Spec.) capable of rapid development in the body after death.
Bull Johns Hopkins Hosp Baltimore
1892
.
3
:
81
91
.
Welch
,
W. H.
Aneurism with demonstration of bacilli causing air in the tissues, and description of the bacillus.
Case report presented at: The Johns Hopkins Hospital Medical Society, November 2, 1891; Baltimore, Md
.
Nysten
,
P. H.
Recherches de Physiologie et de Chimie Pathologiques: Pour Faire Suite à Celles de Bichat sur la Vie et la Morte.
Paris, France: JA Brosson; 1811
.
Welch
,
W. H.
and
S.
Flexner
.
Observations concerning Bacillus aerogenes capsulatus.
J Exp Med
1896
.
1
:
5
45
.
Callender
,
G. R.
and
J. F.
Coupal
.
Pathology of the Acute Respiratory Diseases, and of Gas Gangrene Following War Wounds.
Washington, DC: US Army Surgeon General's Office; 1929. The Medical Department of the United States Army in the World War; vol XII
.
Flexner
,
S.
and
J. T.
Flexner
.
William Henry Welch and the Heroic Age of American Medicine.
Baltimore, Md: Johns Hopkins University Press; 1993
.
Welch
,
W. H.
Morbid conditions caused byBacillus aerogenes capsulatus: The Shattuck Lecture.
Boston Med Surg J
1900
.
143
:
73
87
.

The authors have no relevant financial interest in the products or companies described in this article.

Author notes

Reprints: Brendan P. Lucey, MD, Washington University Medical Center, Department of Neurology, Campus Box 8111, 660 S Euclid Ave, St Louis, MO 63110 (brendanlucey@hotmail.com)