Whither Koch’s Postulates?

S pecific etiology was one of the most powerful and productive ideas in the entire history of medicine. Replacing a shadowy intellectual landscape of humors and miasmata with the notion that particular diseases had particular causes, it marked the beginning of the end for both diagnostic imprecision and clinical impotence. Within the domain we now recognize as microbiology, the prime mover was, of course, Louis Pasteur. He had the genius to discern that the characteristic presence of distinctive types of microbe in different “diseases” of beer, and other natural processes, might be paralleled in human maladies. But it was Robert Koch who outlined the postulates according to which an organism could be conclusively implicated as the agent of a specific disease. The now-familiar form of the four postulates closely followed Koch’s own work on anthrax. An organism under suspicion as a disease agent should be present in all cases of the disease. It should be cultured in pure form in the laboratory, cause the same disease when inoculated into a healthy animal, and be isolated again from the lesions of the disease. There have in fact been significantly different versions of these tenets over the years, as observed by K. Codell Carter (Medical History, 29:353, 1985). He also claimed that Edwin Klebs, rather than Jacob Henle, should be credited alongside Koch as their creator. Whatever the historical truth, and however useful they have undoubtedly been, the postulates have also spawned many problems. Rene Dubos highlighted one such in Mirage of Health (Allen & Unwin, 1960) while discussing Koch’s work on tuberculosis. “Most of the persons present in the very room where he read his epoch-making paper in 1882 had been at some time infected with tubercle bacilli and probably still carried virulent infection in their bodies,” he wrote. “At that time, in Europe, practically all city dwellers were infected, even though only a relatively small percentage of them developed tuberculosis.” In a population where Mycobacterium tuberculosis infection is universal, Dubos suggested, the real cause of the disease is not the bacterium but the malnutrition and exhausting work which, for some unfortunate persons, convert infection into pathology. Half a century later, there are additional grounds for asking whether the tenets of specific etiology can be considered as realistic any more. As well as more sophisticated comprehension of the ecological context explored by Rene Dubos, our understanding of pathogenesis is being continually modified by insights from disciplines such as microbial population genetics and molecular ecology. We now recognize conditions triggered not by single, defined microorganisms but by consortia coexisting in biofilms. Another type of difficulty is posed by organisms that cannot be cultivated in the laboratory. A third stems from our considerably wider perspective on a disease such as cholera, whose epidemics cannot be fathomed on the traditional, simplistic model of pathogen and host but only through analysis ranging over fields as diverse as climatic change and human social behavior. Among several attempts over the years to brush up Koch’s postulates to take account of new knowledge, Stanley Falkow’s paper (Rev. Infect. Dis. 10(Suppl 2):S274, 1988) was an elegant restatement in light of modern molecular genetics. Though conceptually updated, his