Vision without action is a daydream.
Action without vision is a nightmare.
-Derived from The Analects of Confucius

All over the world, physicians are treating patients who carry Helicobacter pylori in an attempt to eradicate the organism. The discovery of H. pylori (1) and its relationship with peptic ulcer disease and gastric cancer (2) has led to a new era. In a relatively short time, the field of gastric microbiology has moved from obscurity to a central stage in medical practice. In this editorial I review the fundamental associations between H. pylori and human health and disease. The issues of how best to diagnose the presence of the organism and how best to eliminate it must follow from the larger question: In whom is elimination of H. pylori beneficial?

Consensus Issues

In the past 15 years, much has been learned about H. pylori (2). We know that once H. pylori is acquired, colonization continues for life unless the organism is eliminated by antimicrobial treatment or by the usually late-in-life development of atrophic gastritis. We also know that essentially everyone who carries the organism in the gastric mucous layer has evidence of tissue reaction (termed chronic active gastritis), yet most colonized persons remain asymptomatic for life. In the absence of treatment, the presence of H. pylori can be determined with a high degree of confidence by endoscopy (with culture, histologic examination, or urease testing of gastric biopsy specimens), by serologic testing, or by urea breath tests. After successful treatment, specific antibody levels decrease so slowly that serologic testing cannot be used to document success for at least 6 months. We know that colonization with H. pylori is associated with an increased risk for developing peptic ulcer disease, noncardia gastric adenocarcinomas, and gastric mucosa-associated lymphoid type (MALT) B-cell lymphomas (2, 3). We know that in most patients, elimination of H. pylori changes the natural history of peptic ulcer disease and of gastric MALT lymphomas (2). It is now recommended that patients with ulcer disease be treated to eliminate H. pylori because the benefits seem to substantially outweigh the risks and costs. However, patients with ulcers represent only a small subset of all those who carry H. pylori.

Emerging Issues

Increasing microbiological and epidemiologic evidence indicates that H. pylori was once more common, perhaps nearly universal in humans, than it is in our postmodern society (4). Clearly, H. pylori has been disappearing as industrialization has proceeded (5), reflecting changes in living conditions. Its decline has probably been accelerated by the widespread use of antibiotics in recent decades and by the now-concerted efforts by physicians to remove it from large numbers of people. If H. pylori once was so common, as it remains in most developing countries, and its presence generally benign, could colonization have benefits as well as costs to humans?

Helicobacter pylori strains are highly diverse, and emerging evidence suggests that the type of strain with which an individual patient is colonized affects risk for disease. In western countries, carriage of cagA ⁺, vacA s1, or iceA 1 strains is associated with an increased risk for duodenal ulceration, atrophic gastritis, and noncardia gastric adenocarcinomas (6-8). However, because strains differ among geographic regions, it has not been possible to generalize about virulence genotypes on a global basis (9). This is particularly germane to the United States, where our ancestors brought their H. pylori with them from their European, African, Asian, and American origins. The relation between bacterial and host genotypes and clinical outcomes is one of the most dynamic areas in current H. pylori investigation.

As H. pylori has been disappearing, peptic ulcer disease and noncardia gastric cancers have predictably been decreasing. However, maladies such as gastroesophageal reflux disease, Barrett esophagus, and adenocarcinomas of the lower esophagus and gastric cardia have been dramatically and progressively increasing. In white men in the United States, adenocarcinoma has surpassed squamous cell carcinomas as the main type of esophageal tumor, and among gastric tumors, cancers at the cardia have become nearly as frequent as those located more distally. Are these phenomena in any way related to the disappearance of H. pylori? This is a critical area for investigation, but preliminary studies already suggest that the presence of H. pylori may protect against these diseases (10-13), especially cagA ⁺ strains; in a seeming paradox, these strains are those most highly associated with risk for lower gastric diseases (7). If these findings are confirmed, as I believe they will be, then the clinical evaluation of H. pylori will become increasingly complex. By eliminating H. pylori to reduce risk in one group of diseases, could we be increasing risk for others? Such ideas suggest another possibility: Might H. pylori have other beneficial features, not readily apparent today, the loss of which might subject humans to unknown perils? A protective role of H. pylori against orally ingested pathogens is attractive, and experimental support for that hypothesis is emerging (14).

An important question is whether we should refer to our interaction with H. pylori as colonization or as infection. Infection indicates a relation that is harmful to the host. Colonization is a broader term that indicates a persistent interaction without qualifying whether or not it is harmful. Favoring the concept of infection for H. pylori is the presence of a tissue response, and the fact that the organism's presence increases risk for disease. However, studies in germ-free animals show that their colonic lamina propria is nearly devoid of cells until their usual colonic bacteria are restored; the lamina propria then fills with immune and phagocytic cells, a phenomenon that we consider the normal response. I believe it likely that the tissue response to H. pylori in the stomach represents a parallel process, but one that we are calling chronic active gastritis. Viridans streptococci and Bacteroides species are persistent colonizers of humans that occasionally cause disease (endocarditis and peritonitis, respectively), but for the most part, our interactions with them are favorable to us. The interactions of H. pylori and humans have many parallels with those for these organisms. The presence of H. pylori increases risk for some diseases (such as peptic ulcer disease) but seems to decrease risk for other diseases (such as esophageal diseases). For these reasons, the broader term, colonization, appears most appropriate for describing the fundamental interaction of H. pylori and humans.

Clinical Dilemmas

Treatment to eradicate H. pylori is improving. Current optimal therapies involve three or four medications, usually including two antibiotics (such as amoxicillin, metronidazole, tetracycline, or clarithromycin), acid inhibitors (such as proton-pump inhibitors or H₂-blockers), or bismuth salts (15). With use of such combined regimens for 7 to 10 days, it seems possible to have eradication rates better than 90%. However, rigorously conducted studies have shown that the results in the United States are 10% to 20% lower than those in Europe. In addition, no head-to-head trials have been done to define optimal therapy. Resistance to macrolides (such as clarithromycin) and nitroimidazoles (such as metronidazole) substantially worsens therapeutic efficacy, and rates of such resistance are increasing (16).

A more important question is, Which persons who are colonized with H. pylori should be treated? Many studies show that there is little, if any, relation between eradication of H. pylori and substantial (greater than placebo effect) improvement in symptoms of nonulcer dyspepsia (17). Even studies with the most positive results show short-term benefit in only a small fraction of patients with nonulcer dyspepsia (18), and no methods are available to prospectively distinguish such patients from the others. Nevertheless, around the world, physicians are treating patients with nonulcer dyspepsia in an attempt to eradicate H. pylori. Whether they are doing more good or causing more harm will, in part, turn on the relation between H. pylori and esophageal diseases.

In certain persons, colonization with H. pylori contributes to the development of chronic atrophic gastritis, which may lead to the intestinal type of noncardia gastric cancer. An important question is whether early intervention can interdict this phenomenon. A treatment trial in Japan of patients who had early removal of gastric tumors showed that with H. pylori eradication, the likelihood of developing a subsequent gastric cancer was reduced (19). Although these data are preliminary, they suggest that in some patients, prophylactic eradication of H. pylori could decrease risk for noncardia gastric cancer. An important challenge will be to define the patients for whom the benefits exceed the costs and risks and to develop simple screening assays to identify these patients in the general population.

Challenges

Investigations of the relation of H. pylori to disease were begun by practitioners who focused on clinical symptoms and clinically defined outcomes. These approaches have had much benefit, especially among persons with ulcer disease. Yet analyses of clinical data often have undefined biases, and, until recently, the field lacked a viable paradigm of the biological associations between H. pylori and humans. In the absence of a paradigm, viewpoints are short-term and immediate results are favored. Yet for maximizing human health, a way of understanding our interactions with H. pylori that addresses both short- and long-term consequences is needed. The foundation of such a model must be ecological and be rooted in understanding the selective forces that have governed our relations with H. pylori since earliest times (20).

Although popularly believed to be so, Koch's postulates for a role of H. pylori in peptic ulcer disease have never been fulfilled. Similarly, ulcer disease was rare in the days when H. pylori was highly prevalent in what now are developed countries and is uncommon today in many developing countries where almost all adults are colonized (4, 5). Why did ulcer disease arise just as H. pylori was beginning to recede from humans? Why do ulcer rates differ between men and women even though both are colonized with nearly equal frequencies? With lifelong colonization, why is ulcer disease concentrated in particular decades of life? Even in the relation of H. pylori with ulcer disease, profound and ultimately clinically important questions about H. pylori remain.

Synthesis

It has been only 15 years since the presence of the gastric microflora was rediscovered. In our society, there is a technological imperative. Physicians are under great pressure—from their patients, their managed care organizations, the lay press, the pharmaceutical companies, and their own eagerness to help—to use the latest panacea. But penicillins arise only rarely, and even these, with overuse, carry their own problems.

What should physicians do about H. pylori? My advice is to search for it only in patients with peptic ulcer disease and gastric MALT lymphomas. In such patients, the benefits of treatment seem to outweigh the risks and costs. However, before treating other patients, we should wait to learn what ongoing research will show us. Fifteen years is a short time to reach definitive conclusions about eliminating organisms that have lived with us for millions of years. It is entirely possible that physicians in the future will be administering selected H. pylori strains to colonize selected patients to reduce risks for particular diseases.