How do we know bacteria cause disease? A simple possible experiment is to isolate the bacterium, infect yourself through the suspected portal of entry, and see if you become ill. Unacceptable, you may say. But this is exactly what Barry Marshall did in 1984 to prove the link between the newly discovered Helicobacter pylori bacterium and gastritis. His experiment paved the way for use of antibiotics to treat gastritis and gastroduodenal ulcers.

There is, of course, more to the story. In 1984 Barry J. Marshall and J. Robin Warren working in Western Australia reported the presence of “unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration”.¹ Most dramatically, all 13 patients they examined that had duodenal ulcers had the bacteria, and 87% of patients of all ulcer types harbored them. These observations suggested to Marshall and Warren that these newly discovered critters may be the causative agents of ulcers and other more mild forms of gastric inflammation.

But how to prove this? The physician Robert Koch (1843-1910) developed four postulates that if fulfilled essentially prove the link between an infectious agent and disease.² In brief, these postulates are as follows:

1. The suspected pathogen should be present in all cases of the disease and absent from healthy host organisms.
2. The suspected pathogen should be isolated from diseased hosts and cultured to purity outside the host.
3. Pure pathogen cultured outside a host should cause disease when introduced into a healthy host.
4. The pathogen should be re-isolated from the host infected in step 3 and shown to be identical to the original isolate from step 2.

The first two of Koch’s postulates are, at least in principle, possible to fulfill in the investigation of most pathogens. However, on ethical grounds, the third and fourth postulates are often impossible to satisfy for diseases that are unique to human beings. So what are the alternatives? One approach that has been used in AIDS studies, for example, is to treat affected individuals with a curative agent developed in the laboratory against the pure pathogen, and assess, on an epidemiologic scale whether the agent prevents disease. Such studies, by their very nature, require many years for drug development. Furthermore, because of confounders unbeknownst to the investigator, it is difficult to know whether an agent really works unless disease prevention succeeds with very high efficiency.

Marshall, faced with significant opposition from the medical community to his ideas about H. pylori and ulcers, probably realized these limitations and opted to meet Koch’s postulates directly, by self-experimentation. His experiment³ satisfied Koch’s third and fourth postulates. The first two postulates were already met.^1 4

Marshall’s unorthodox approach to treating disease has had a world-wide impact on treatment of ulcers, and has provided a cure for millions suffering from this disease state. In recognition, he has been the recipient of numerous awards, including the Lasker award (1995). However, his experiment certainly toyed with an ethical boundary. True, he could have only harmed himself. Thus, this was different from Edward Jenner who infected an 8-year-old boy with cowpox virus as a trial vaccine for smallpox in 1796. But does Marshall’s experiment represent the way we would like all research to proceed? In this era when emerging infectious diseases are a reality, these are issues that not only ethicists, but probably all scientists should discuss.

¹ Marshall, BJ and Warren, JR (1984). Lancet 1: 1311-1315.

² Grafe, A (1991). A history of experimental virology. Heidelberg: Springer-Verlag, p. 343.

³ Marshall, BJ et al. (1985). Med. J. Aust. 142: 436-439.

⁴ Marshall, BJ et al. (1984). Microbios 25: 83-88.

Related Articles:

• Microbial Pathogenesis
• Political Science