QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cheskin, L. J. Search for Related Content PubMed PubMed Citation Articles by Cheskin, L. J.

---

Supplement

The Pathogens Are Speaking: Are We Listening?¹

Johns Hopkins School of Medicine, Department of Medicine, Division of Gastroenterology, Baltimore, MD 21224

²To whom correspondence should be addressed. E-mail: cheskin{at}JHMI.edu.

	An overview of the symposium

TOP An overview of the...

 
We have presented provocative data in this symposium that support an important role for viral or bacterial pathogens in the etiology of some of the most prevalent health conditions that afflict humans. These conditions are also notable for their chronicity and the importance of diet in their pathogenesis and/or treatment.

Before delving into a summary of the specifics and the complications of these discoveries, please permit some reflections on how one establishes a relationship between a pathogen and chronic conditions and why it may be that such relationships escape detection for so long and often face an extremely skeptical reception until they are proved beyond any doubt.

The classic method of demonstrating that a microorganism is pathogenic is via satisfying Koch’s postulates. Essentially, one must show that the disease or condition occurs in association with the organism, can be transferred from an infected to a noninfected host via transfer of the microorganisms, and can then again be isolated from the newly infected host.

The challenge in the modern era of satisfying these time-honored standards is that we are studying not acute infectious diseases, but chronic conditions that are not clearly infections, for which there is often no animal model, in an age in which ethical considerations proscribe cutting to the chase and directly assessing the pathogenicity of our suspects in humans.

The basis for suspecting that there is an infectious cause of a chronic condition is, as in much of medical discovery, often anecdotal to begin with. If I notice that a number of my patients with X infection (say Helicobacter pylori to provide a concrete example) are, say, obese, it is certainly to me a striking association, but perhaps it is a coincidence. How would I go about proving that H. pylori actually causes obesity? Further case-finding is often the first step, leading to case-series and, at a more refined epidemiologic level, case-control studies. These are usually retrospective in design by necessity, and are subject to a variety of criticisms even if they are supportive of the association. For example, there may be ascertainment bias, and a given is the vexing problem that epidemiologic studies detect associations, not causation. Epidemiologic studies may not correct for or even know of the existence of intervening variables that explain the association. If it turned out, e.g., that H. pylori–infected patients also eat at fast food chains more because they are poorer, this finding could explain obesity occurring at a high frequency among H. pylori–infected individuals without any direct causative mechanism for the infectious agent. The bottom line is, no matter how suggestive, observational and epidemiologic studies provide, at best, indirect evidence of causation.

To address causation directly, we must turn to the suspected agent itself. This necessitates laboratory work and usually an animal model (though as Dr. Mason argues with much persuasiveness, an in vitro cellular model may suffice for certain conditions marked by pathogenic cellular changes in vivo). Each of Koch’s postulates in turn are paraphrased below, and the pitfalls and potential ways around them.

Agent is present in affected host (and not in unaffected host).

For an acute infection, this is not a problem. For complex chronic conditions, rarely is there a single etiology, so that the evidence will not be clear-cut even if an infectious agent is a cause of the condition. Clearly, only a subset of obesity is caused by infectious agents, so that presence of the agent is not a sine qua non of the condition. Further, the host may exhibit varying degrees of susceptibility to the agent, so that there is a mixed picture, in both the proportion affected who bear the infection and the population of the infected who exhibit/express the condition. This is akin to finding both a low sensitivity and low specificity for a test, which is not a desirable combination. Further, the agent may prove to be ubiquitous and to be a commensal rather than a pathogen.

Isolate agent in pure culture from affected host.

There are difficulties in the isolation and culturing of fastidious organisms, particularly viruses; yet, even in the case of such agents of more traditional lineage, such as H. pylori, isolation and culture were quite challenging and protracted.

Inoculation of a pure culture into uninfected hosts causes condition.

In addition to the difficulties mentioned in reference to postulate 2, there may be limitations because of the absence of an animal model, the ethics of human inoculation experiments, and the possibility that only certain susceptible hosts will be affected by the agent or that transmissibility is poor.

Recovery of agent from infected host.

As with the isolation of the agent in pure culture from the affected host (postulate 2), the initial isolation is quite challenging.

To get around the problem of chronic disorders with multiple etiologies, describing and defining the model as narrowly as possible is important. For H. pylori and ulcers, for example, this meant not considering cases of ulcers caused by irritants such as nonsteroidal antiinflammatory agents.

To get around the problem of susceptibility and incomplete expression of the condition by the infected, one must work to understand the nature of the susceptibility, including immune and genetic factors. This is arguably the most important limitation ultimately in our ability to explain a chronic condition with an infectious model, and it is a worrisome sign that we may be barking up the wrong tree. This problem held up the acceptance of H. pylori as the etiology of most ulcers because it could be argued for years that the H. pylori was an innocent bystander, a commensal, and that acid was still causing the ulcers. Only extensive, carefully controlled treatment trials can start to overcome this challenge.

To overcome isolation/pure culture/infection problems, we must turn to cell cultures, transfection models and animal models. These are less convincing than direct infection studies, but they are again best supported by bottom-line treatment trials.

So, where do we stand in our evaluation of the role of infectious agents in the chronic conditions covered in this symposium? (For the details, the reader is referred to the excellent coverage of these topics in the articles published in this supplement.)

In summary, H. pylori is well established as causative in most cases of peptic ulcer disease and gastric carcinoma. The history of the acceptance of this agent is instructive of the hurdles faced in establishing an infectious etiology in a chronic condition, as widespread acceptance has come only recently, after years of clinical treatment trials.

In obesity, adenovirus-36 has been associated with, but not proven causative of, increased visceral adiposity. Viral agents are exciting although they will likely account for only a subset of obesity. The mechanism of the action in the case of adenovirus-36 appears to be upregulation fat cell differentiation by the virus.

In autoimmune liver disease of the primary biliary cirrhosis type, the evidence for a retroviral (non-HIV) etiology is mounting, but most observers will undoubtedly await controlled clinical treatment trials before accepting this potential cure for a relatively rare but devastating condition.

Finally, in atherosclerosis, both a bacterial agent, Chlamydia pneumoniae, and a viral agent, human cytomegalovirus, seem to play an important role in acceleration of atherosclerosis. The ubiquitous nature of both of these agents means that, if treatments are someday proven to be effective, a substantial majority of patients with atherosclerotic cardiovascular disease could be helped.

While in no way would it be desirable for an emphasis on infectious etiologies of chronic conditions to undermine the well-established dietary and other lifestyle factors known to be important for prevention and sometimes treatment, in concert, these studies and traditional risk factor reduction have the potential to do what neither can do alone—provide a lasting treatment attainable by most people who suffer from these increasingly more important chronic conditions.

We eagerly await future developments in this provocative field.

	FOOTNOTES

 
¹ Presented as part of the symposium entitled "Emerging Role of Pathogens in Chronic Diseases Requiring Nutritional Intervention" given at the Experimental Biology 2001 Meeting held March 31–April 4, 2001 in Orlando, FL. This symposium was sponsored by the American Society for Nutritional Sciences and was supported in part by educational grants from Novartis Nutrition and Ortho McNeil. The proceedings of this symposium are published as a supplement to The Journal of Nutrition. Guest editors for the symposium publication were Nikhil V. Dhurandhar, Wayne State University, Detroit, MI, and Lawrence J. Cheskin, Johns Hopkins School of Medicine, Baltimore, MD.

This article has been cited by other articles:

				S. de Ferranti and D. Mozaffarian The Perfect Storm: Obesity, Adipocyte Dysfunction, and Metabolic Consequences Clin. Chem., June 1, 2008; 54(6): 945 - 955. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cheskin, L. J. Search for Related Content PubMed PubMed Citation Articles by Cheskin, L. J.