Assigning Causation in Disease: Beyond Koch's Postulates

ASSIGNING CA USA TION IN DISEASE: BEYOND KOCH'S POSTULATES MORLEY C. SUTTER* Introduction Assessing and assigning causality is important in the treatment and prevention of disease. However, identifying causality has philosophical and practical difficulties, which often provoke heated discussion. The intent of this essay is to look at causality in disease: to explore some of the history and some problems involved; to suggest possible approaches to their solution ; and to discuss why it is important to try to assign causality. Koch's postulates frequently are used to establish causation of disease. Their fulfillment often is viewed as the "gold standard" for proofofcausality [I]. On the other hand, the philosopher Karl Popper (1902-1994) maintained that a hypothesis can not be proved, only disproved [2, 3]. There exists an apparent conflict between "proof" of causality and the views of Popper, and this has led to much discussion of causation in the epidemiological literature with criticism as well as defense ofPopper's views [4-8]. Popper's views raise questions about the adequacy of Koch's postulates . Are they the ultimate test in decisions concerning causality, and how do they relate to the Popperian approach to science? A Brief History ofAssigning Causality in Disease THE ANCIENTS A recent book by Evans and an article by Donham and Thorne succinctly trace the history of European attempts to identify causation in disease [9, 10]. Beginning with Empedocles in Greece in 500 b.c., and elaborated by Aristotle and by Galen, the belief was that diseases were due to imbalance The author acknowledges with gratitude helpful discussions with Malcolm Maclure and John Norris. *Department ofPharmacology and Therapeutics, University ofBritish Columbia, 2176 Health Sciences Mall, Vancouver, BC V6T 1Z3 Canada.© 1996 by The University of Chicago. All rights reserved. 0031-5982/96/3904-0968$01.00 Perspectives in Biology and Medicine, 39, 4 ¦ Summer 1996 581 between the four humours: phlegm, blood, yellow bile, and black bile. This view persisted well into the 1500s. There then followed a succession of contributors to the understanding ofcausality, including Paracelsus (14931541 ), William Harvey (1578-1657), and Linnaeus (1707-1778), who introduced , respectively, the concepts of dose, experimentation, and classification into the notion of diseases, their causes, and their treatment [10]. At least one other name should be added to this list: Sydenham (16241689 ), who described disease species [9] , whereas Linnaeus dealt only with botanicals. KOCH'S POSTULATES The germ theory of tuberculosis (TB) as enunciated by Robert Koch (1843-1910) had its origins in the microscopic observations of the botanist van Loewenhoek (1632-1722); in the theories of the histologist Henle (1809-1885) that living organisms caused disease; in the antisepsis ofLister (1827-1912); and in Pasteur's (1822-1895) demonstration of the involvement of germs in anthrax [10]. The transmissibility of tuberculosis between animals by inoculation with extracts from tuberculous lesions was first shown by Villemin, but he did not demonstrate a parasite in the inoculum or in the tuberculous lesions [H]. Koch performed meticulous experiments to identify the bacillus in, and to culture it from, the several types of tuberculous lesions of man: miliary tuberculosis of the lungs, caseous pneumonia, a solitary tubercle in the brain, intestinal tuberculosis, tuberculous lymph nodes, and tuberculous bone. He also isolated and grew the bacillus from the lungs of ten cattle, a chicken, three monkeys, nine guinea pigs, and seven rabbits, all of whom had died of tuberculosis. In a paper presented to the Berlin Physiological Society, Koch first enunciated his criteria for establishment of a causal role in tuberculosis for the bacillus he had isolated: "it was necessary to isolate the bacilli from the body; to grow them in pure culture . . . and by administering the isolated bacilli to animals, to reproduce the same morbid condition which, as known, is obtained by inoculation with spontaneously developed tuberculous material" [H]. These criteria—to isolate the bacilli from people and animals with the disease; to grow the bacilli in pure culture; and to produce the disease by inoculation ofthe bacilli into animals—are Koch's postulates as originally stated. They often are presented in more elaborate forms, e.g., requiring a second isolation of the bacillus from the inoculated animal [I]. Variants...