how a doctor proved bacteria cause ulcers by drinking them

The Mechanism

*Barry James Marshall* (born Kalgoorlie, Western Australia, *30 September 1951*) was, in 1984, a *32-year-old internal-medicine registrar* at *Fremantle Hospital* in the port suburb of Fremantle, twenty kilometres south of Perth, Western Australia. He had completed his MBBS at the *University of Western Australia* in 1974 and had begun his internal-medicine training at *Royal Perth Hospital* in 1981, where in *June 1981* he had been assigned a six-month rotation in gastroenterology and had been asked by the hospital pathologist *J. Robin Warren* (born Adelaide, *11 June 1937*; died Perth, *23 July 2024*, aged 87) to investigate a series of *biopsy specimens* in which Warren had been noting an unusual *curved, gram-negative, motile bacterium* in the gastric mucosa of patients with chronic gastritis. Warren had been seeing the bacterium since *June 1979* — silver-staining gastric biopsies and finding the same characteristic small spirochete-like organisms in the *gastric pits* of patients with chronic active gastritis — and had become convinced that the organism was a real biological entity, not a contaminant, but had been unable to interest any of the gastroenterology consultants in pursuing the question. Marshall, in his early career and looking for a research project, agreed to do prospective clinical work to test whether the bacterium was associated with disease. Through 1981-83, Marshall and Warren collected biopsy specimens from *100 consecutive patients undergoing diagnostic upper endoscopy at Royal Perth Hospital* and correlated the silver-stain finding of curved bacteria with each patient's clinical diagnosis. *77 of 22 patients with gastric ulcer* and *100% of 13 patients with duodenal ulcer* showed the bacterium present; *87 of 100 patients with chronic gastritis* showed it present. Healthy controls — biopsies of normal gastric mucosa — showed the bacterium absent. The correlation was overwhelming. The community's response was uniformly dismissive. The *consensus view of peptic ulcer disease*, in 1981-84, was that ulcers were caused by *excess gastric acid secretion driven by psychological stress, personality type, dietary factors, and smoking*; the dominant pharmaceutical paradigm was the *histamine H₂-receptor antagonist class* of acid-suppressing drugs (Tagamet/cimetidine, introduced 1976, and Zantac/ranitidine, introduced 1981), which were among the most commercially successful drugs of the 1980s and made $1 billion per year in worldwide sales for SmithKline. The H₂ blockers worked to suppress acid and heal active ulcers, but ulcers recurred in the majority of patients within a year of stopping the drug — the standard recommendation was *lifelong daily maintenance therapy*. The proposition that a *bacterial infection treatable with a 10-day course of antibiotics* could permanently cure the same disease was deeply unwelcome to the dominant clinical-pharmaceutical paradigm and was nearly unanimously rejected by the gastroenterology community in 1982-84. Marshall and Warren submitted their initial findings as a letter to *The Lancet* in October 1983, where it appeared on *4 June 1984* as Marshall, B.J., Warren, J.R., *"Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration,"* *The Lancet* 1(8390): 1311-1315. The two had also separately presented *case reports* of bacterium-positive patients whose chronic ulcer disease had been cured by 14-day courses of *bismuth subsalicylate plus metronidazole* — but the case-report evidence was widely dismissed as anecdotal. The dominant criticism was that *Koch's postulates* — the four classical criteria for establishing that a microbe causes a specific disease — had not been met. The third Koch postulate requires that the suspect microbe, when introduced into a healthy host, produces the same disease. Marshall and Warren had been unable to produce the disease in laboratory animals: attempts to infect rats, pigs, and dogs with the bacterium had failed. Without a satisfied third postulate, the bacterial-cause hypothesis was experimentally incomplete. *On a Tuesday morning in July 1984* (the exact date was either *24 July* or *31 July 1984*, recorded in Marshall's lab notebook), at *Fremantle Hospital*, Marshall walked into the hospital microbiology laboratory next to his ward, asked the lab technician to *scrape the contents of two petri dishes* — Skirrow plates incubated for 4 days at 37°C, supporting visible growth of curved bacteria from a 66-year-old male patient who had presented with chronic gastritis — *into a 30-ml beaker of beef-extract broth*. The technician complied without being told what the broth was for. Marshall returned to his office, swirled the cloudy yellow-grey liquid (an estimated *10⁹ colony-forming units* of the bacterium), drank it in a single swallow, and went back to work. He told only his wife, *Adrienne*, what he had done — he did not tell the hospital ethics committee, the chief of medicine, his colleagues, or his children, because he was sure the ethics committee would have refused permission and he did not want to risk being told no. The first symptoms appeared *three days after the ingestion*: a sense of fullness and abdominal bloating, foul breath that his family began to remark on, and *intermittent morning vomiting of clear gastric fluid without bile*. On day 5 Marshall had himself *endoscoped* by a colleague at Fremantle Hospital under conscious sedation; the colleague was told only that this was for "research purposes" and did not ask further. The endoscopy showed *acute gastritis with diffuse erythema and superficial erosions* of the gastric antrum and body; biopsies were taken. Marshall was endoscoped a *second time on day 10*, the biopsies of which grew the same curved bacterium that he had originally cultured from the donor patient — establishing that the same organism he had swallowed had colonised his own previously-healthy gastric mucosa and produced gastritis. He had fulfilled Koch's third postulate *in his own body*. He took a 14-day course of *bismuth and metronidazole*; the symptoms resolved within 3 days of starting treatment, and follow-up biopsy showed clearance of both the bacterium and the gastritis. The self-experimentation case was published as Marshall, B.J., Armstrong, J.A., McGechie, D.B., Glancy, R.J., *"Attempt to fulfil Koch's postulates for pyloric Campylobacter,"* *Medical Journal of Australia* 142(8): 436-439 (15 April 1985). The community's response was, on the whole, *still skeptical* — the absence of formal animal-model confirmation, the small sample size, and Marshall's unconventional approach all worked against rapid acceptance — but a small minority of clinicians began running pilot bismuth-and-metronidazole trials in their own ulcer patients, and through the late 1980s the cumulative weight of randomised trial evidence — particularly the *1988 McNulty et al.* New England Journal of Medicine study of 100 patients showing that *Helicobacter pylori* eradication achieved permanent cure of duodenal ulcer disease in 92% of treated patients versus 9% of placebo controls — slowly shifted the consensus. In *1989*, *Goodwin et al.* renamed the bacterium *Helicobacter pylori*, reflecting *Helix* (the spiral shape) and *pylori* (the pyloric antrum of the stomach where the organism is concentrated). The *NIH Consensus Conference of February 1994* formally adopted bacterial-eradication therapy as the standard of care for peptic ulcer disease, and the *2005 Nobel Prize in Physiology or Medicine* was awarded jointly to *Barry J. Marshall* and *J. Robin Warren* "for their discovery of the bacterium Helicobacter pylori and its role in gastritis and peptic ulcer disease." Marshall, age 73, is in 2026 the Brand Marshall Professor of Clinical Microbiology at the University of Western Australia. Warren died on 23 July 2024 at age 87. The original *beaker* from which Marshall drank the broth is now on permanent display at the *Nobel Museum in Stockholm*. The pharmaceutical industry's H₂-receptor-antagonist franchise collapsed within five years of the consensus shift; *over a million peptic-ulcer surgeries per year worldwide* in the 1980s had fallen to nearly zero by the early 2000s. The current estimate is that *half of all humans alive in 2026 carry H. pylori* in their stomachs; in most carriers, the bacterium is asymptomatic, but in a small fraction it causes peptic ulcer disease and a smaller fraction develops *gastric adenocarcinoma*, for which H. pylori is now classified by WHO as a Group 1 carcinogen.