H. pylori, a gram-negative bacterium with a curved, spiral, or gull-wing shape, is 2.5 to 3.5 µm long and 0.5 to 1.0 µm in diameter and has a periodicity of 1 to 2µm. It has smooth surfaces, and one to six polar-sheathed flagellae emerge from one of its rounded ends. Since it is morphologically similar to C. jejuni, it was initially named "pyloric Campylobacter" and subsequently C. pyloridis and C. pylori before finally being named H. pylori. This organism colonizes only the non-acid-secreting mucosa of the stomach and is not found where parietal cells are numerous. Thus, it may be observed in the gastric antrum and the cardia, but also in the corpus, when atrophic gastritis is present, and attached to the gastric epithelial cells found in the duodenum, when gastric metaplasia is present.

G. hominis (H. heilmannii) is tightly spiraled, and is 3.5 to 7.5 µm in length and 0.9µm in diameter; it has a periodicity of 0.8 to 1 µm and up to 12 flagellae at each pole. 16S rRNA indicates that this organism belongs to the genus Helicobacter, and is more closely related to a Helicobacter sp. isolated from the stomach of cats (H. felis) than to H. pylori (17). The name H. heilmannii was proposed in honor of the late German pathologist Heilmann. However, the subsequent examination of the rRNA of different clinical isolates indicates that there is enough heterogeneity among isolates tentatively identified as H. heilmannii that it is premature to propose an official name (17) . This bacterium colonizes only the parietal cell area of the gastric mucosa and may be found within parietal cells (18,19).

H. pylori infection may be diagnosed by harvesting gastric biopsy specimens during endoscopy, by culturing and isolating the bacterium under microaerobic conditions (90% N2, 5% O2, and 5% CO2), and by characterizing the enzymes (urease, catalase, and oxidase) it produces. Visualization of the bacterium by light microscopy on slides stained with hematoxylin and eosin, Gram, Giemsa, Genta, or Warthin-Starry stain is also of great benefit since it allows the concurrent diagnosis of the extent of the antral chronic-active gastritis that H. pylori causes. However, because H. pylori colonization is focal, negative biopsy results do not exclude the possibility of infection in areas not sampled. Infection also may be diagnosed by determining plasma and salivary immunoglobulin (Ig) G or IgA levels with enzyme-linked immunosorbent assays (20,21). This latter technique is noninvasive, specific, and sensitive and is believed to reflect the mucosal and systemic immunity induced by H. pylori infection.

Two other tests, which rely on the production of urease, also can be used to identify H. pylori. One is the CLO (for Campylobacter-like organisms) test, which is performed by placing a mucosal biopsy specimen in medium containing urea and a pH- sensitive dye that changes color in the presence of OH- ions. The second test is the noninvasive 14C or 13C breath test following the oral administration of 14C- or 13C-urea. Neither of these tests is specific for H. pylori since G. hominis, which generates urease, also gives a positive reaction. Until specific methods based on the polymerase chain reaction (PCR) amplification of 16S rRNA (17) become widely available, the diagnosis of G. hominis infection must rely on histologic morphologic characteristics; histologic identification must be confirmed by transmission electron microscopy since other spiral organisms, e.g., Flexispira rappini, also may be present in the stomach of humans (22).

The seroepidemiology of H. pylori has been extensively studied in the United States and in other countries (23). The high frequency of seropositivity (up to 100% in some age groups in Albania) and acquisition of the infection during infancy are characteristic of disadvantaged socioeconomic groups living in crowded or poor hygienic conditions and appears to be independent of gender and ethnic origin. In adults of higher socioeconomic groups, the rate of seroconversion is estimated at 0.5% per year, although the frequency of seropositivity increases with age and may be as high as 40%. A longitudinal study has indicated that the high frequency of seropositivity in older adults might be due to a higher rate of H. pylori infection in Western countries in the years between the two world wars than during recent years (cohort effect) (24). Alternatively, the increase in frequency of infection in older adults might be due to years of low but cumulative risk for infection. Although the route of transmission for this infection is not known, the contamination of drinking water may play a role in certain developing countries (25). In the United States and in other regions, direct contact and/or consumption of food or water contaminated by saliva (26), gastric contents, or feces (27) may be major factors. The recent observation that H. pylori can be isolated from cats (28) suggests that transmission from pets to humans (or humans to pets) is also possible.

The epidemiology and route of transmission of G. hominis are largely unknown. The frequency of this infection appears to range from less than 1% of the population in industrialized countries (29) to 3% to 8% in developing countries (30). Although the detection of spirilla in the stomach of cats and dogs suggests possible transmission from pets, marked morphologic differences exist between these spirilla and the organism found in the stomach of humans.

H. pylori is considered a pathogen because its presence is always associated with chronic active gastritis, and eradication of the bacterium is always followed by resolution of gastritis. In addition, nearly all patients with duodenal ulcer disease have H. pylori gastritis, and ulcer relapse is exceptional after H. pylori eradication. Thus, the presence of H. pylori seems necessary for the production of duodenal ulcers, with the exception of ulcers attributed to the use of nonsteroidal antiinflammatory agents or to the Zollinger-Ellison syndrome (10). The association with gastric ulcers is not as strong, although H. pylori infection is present in 80% of patients with gastric ulcers who do not consume nonsteroidal anti-inflammatory agents (10). However, most H. pylori-infected persons do not report any clinical symptoms. This may be because these persons are colonized by less virulent strains or because other host or bacterial cofactors are required for overt disease.

In addition, three prospective cohort studies have demonstrated that H. pylori-infected persons have an increased risk of developing intestinal-type, but not undifferentiated, gastric adenocarcinoma (10). In fact, the association of H. pylori with either gastric ulcer or gastric cancer may be underestimated in these studies: the atrophic gastritis that follows long-term infection makes the gastric niche less hospitable for the bacterium, which may either eliminate H. pylori or make it difficult to detect. Nevertheless, atrophic gastritis per se is believed to be a precancerous lesion that leads to carcinogenesis without the presence of H. pylori.

The pathogenicity of G. hominis is unclear. The organism has been associated with upper gastrointestinal complaints, and its carriage is generally accompanied by gastritis, although the inflammation and gastric atrophy are less than noted with H. pylori (31,32). In addition, G. hominis was observed in gastric cancer patients (3) as well as in patients with only minimal gastritis (29). In this relatively small number of cases, the frequent concurrent infection with H. pylori makes interpreting the respective pathogenic role of either bacterium difficult. It is probable that G. hominis will turn out to be at least somewhat pathogenic, as it makes urease and products of urease action that have been implicated in inflammation.

H. pylori multiplies with great efficiency in the hostile environment within the stomach but survives poorly in the gastric lumen; it is mainly found where the pH ranges between 4 and 7, i.e., under the mucous layer and in close proximity, or even attached, to gastric superficial epithelial cells. The virulence and the ecologic niche of G. hominis are unknown, although its presence within parietal cells of patients with gastrointestinal complaints (18,19) suggests that it is even more resistant to acid than H. pylori.

The production of urease was the first putative colonization or virulence factor studied. The production of this enzyme is shared by the two organisms, and it may explain their extraordinary ability to survive in an environment previously considered sterile because of the presence of proteolytic enzymes, as well as the low pH of gastric contents. Because the ecologic niches of these bacteria are rich in urea, urease generates OH- ions that neutralize gastric acid. Although the neutralization of gastric acid benefits the two bacteria, the production of hydroxide ions also is toxic to gastric epithelial cells in vivo, as indicated by in vitro experiments (33).

Two other important virulence factors shared by H. pylori and G. hominis are their spiral shape and the motility of their flagellae, which render them resistant to peristaltic flushing of the gastric contents and enable them to persist in the mucous layer. Because G. hominis appears to infect fewer persons than H. pylori, a more important role might be attributable to characteristics that are unique to H. pylori; these include the production of other enzymes (catalase, oxidase, protease, and phospholipase), as well as the synthesis of specific adhesin proteins that enable them to adhere to mucous and epithelial cells, both in vivo and in vitro (34-36).

The putative virulence factor of H. pylori that has commanded the most attention during the past few years has been its vacuolating cytotoxin (vacA gene product). Intragastric administration of the toxin to mice causes some (but not all) of the tissue damage seen in H. pylori-infected persons (37). In addition, cytotoxin production is highly correlated with the production of a high molecular weight (120 to 128 kilodaltons) major protein antigen that is called cytotoxin-associated protein (cagA) and is not the toxin itself (38).

H. pylori isolates may differ with respect to each of the virulence factors described above; this diversity is likely to contribute to variation in colonization or disease. For example, urease-negative strains have been isolated, and the vacuolating cytoxin is produced by only a subset of H. pylori strains (vacA+ or tox+ strains) (39-41). This observation is probably clinically relevant because most or all strains from duodenal ulcer patients, and many strains from gastric cancer patients, produce cytotoxin, whereas only a fraction of strains from patients with gastritis alone produce the cytotoxin (42,43). This phenotypic diversity is mirrored in great diversity on the DNA level. Thus, only cytotoxin-producing strains contain the gene for this cytotoxin-associated protein (cagA) (38,42), although genetic tests have shown that cagA protein is not needed for toxin production (44). Strains that do not produce the 128-kDa cagA protein generally lack the entire cagA gene and additional neighboring genes. Although the function of the cagA region is unknown, its presence or absence is easily scored by hybridization or PCR and thus serves as an easy marker for probable cytotoxin production and possible virulence of H. pylori strains. Additional virulence factors are likely to be present. For example, another recently discovered region constitutes at least 21 kilobases of the H. pylori genome in hybridization experiments, and its presence is highly correlated with the presence of cagA: 39 of 40 strains lacking cagA also lacked this region, and 50 of 52 strains containing cagA contained this region. This newly discovered region is being called cagII, and the effort to sequence it is nearly complete (D. E. Berg, pers. comm.). Preliminary searches have identified several open reading frames with strong homologies to virulence functions from other microbes (45).

In addition to these extensively studied genes, genetic diversity of various H. pylori strains can be demonstrated by the use of two sensitive, efficient, and reliable PCR-based methods (46,47). This approach is particularly useful because it allows tracing of strains in epidemiologic studies.

One of the most puzzling aspects of gastric infection with H. pylori is its persistence despite intense local and systemic immune responses. These immune responses are extremely complex and vary among infected humans. The systemic response is characterized by a marked increase in plasma IgG, which remains present for months after the infection has been cured. The local response includes the production of IgA, which binds to the surface antigens of H. pylori in vitro and coats the bacterium in vivo. In addition, infection is consistently associated with an intense inflammatory response and the infiltration of cells into the gastric mucosa. Although polymorphonuclear cells are often present, most cells in such infiltrates are mononuclear cells. Both B and T cells are present, and recent studies have indicated that the natural killer activity of peripheral blood lymphocytes can be increased by H. pylori, possibly by its stimulating the production of interferon and other cytokines (48). Thus, the long-term carriage of the infection may be related to the ability of the bacterium to influence the T-cell response. Fragmentary evidence also suggests that this infection can be abortive and cure spontaneously without the use of antibiotics (A. Dubois and D. E. Berg, unpublished).

On the other hand, the mucosal response may promote colonization, as indicated by the observation that patients with acquired immunodeficiency syndrome (AIDS) tend to have a lower rate of infection than aged-matched subjects who are negative for human immunodeficiency virus (49,50). The latter study (50) also demonstrated that AIDS patients had a different pattern of gastritis, characterized by greater mononuclear cell responses, fewer lymphoid follicles, and a greater prevalence of intestinal metaplasia. The immune response may also prevent the invasiveness of H. pylori, as suggested by the anecdotal but puzzling observation of invasive H. pylori infection in a patient with AIDS (51).

Although H. pylori is sensitive to many antimicrobial drugs in vitro, it is difficult to eradicate from the stomach. This may be ascribed to antibiotic breakdown by gastric acid, clearance by gastric emptying, and the difficult-to-penetrate mucous layer in which the bacterium resides. Resistance of H. pylori to specific antibiotics, especially metronidazole, is also frequent. Therefore, it is generally accepted that a combination of at least two, and possibly three, antimicrobial agents should be given for a minimum of 1 week. The regimen found to be most effective is the administration of amoxicillin (or tetracycline) plus metronidazole and bismuth subsalicylate 2 to 4 times a day for 2 to 3 weeks (52). The use of one antibiotic associated with an antisecretory agent, such as a histamine H2 receptor antagonist, has given disappointing results. In contrast, the combination of a proton pump inhibitor (H+-K+ ATPase antagonist) with amoxicillin or acid-stable macrolides (clarythromycin or roxithromycin) appears more promising; a number of studies are being conducted to determine the optimal dose, duration, concomitant therapy, and cost-effectiveness of these compounds (53,54). Recently, it was shown that at least a 7-day course of any of these regimens is required to obtain a high (90%) cure rate, but that continuing treatment for more than 10 days does not significantly improve its efficacy. Finally, topical therapy for 1 h was recently tried with excellent results, albeit in only one center at this time (55). This treatment involves a 2-day administration of a mucolytic agent to dissolve the mucous layer and of a proton pump inhibitor. On the third day, a balloon is introduced into the second portion of the duodenum under fluoroscopic control, and a solution of pronase, amoxicillin, metronidazole, and bismuth subsalicylate is injected into the stomach, where it is left for 1 h. The presence of the duodenal balloon appears to prevent emptying of the antibiotics and the mucolytic agent, thus ensuring maximum efficacy of the therapy.

The past 12 years have seen extensive progress in research on H. pylori as a cause of chronic active gastritis, duodenal ulcer disease, and gastric cancer. This has been largely due to an unusual collaboration among gastroenterologists, pathologists, molecular geneticists, bacteriologists, and immunologists. However, our understanding of how H. pylori colonizes and causes diseases is far from complete, and it will benefit from studies performed in animal models that can be experimentally infected with H. pylori (56-59). In addition, no easily administered treatment leading to eradication of this bacterium in all patients is yet available, although a better knowledge of its physiology may lead to the development of such a "silver bullet." Studies in animals that are not naturally infected with H. pylori suggest possibilities for vaccines (56,57), and ongoing trials in nonhuman primates are exploring the possibility of immunizing hosts that can be naturally infected with this organism. Although the elimination of peptic ulcer disease and of certain forms of gastric cancer will require extensive and coordinated efforts from public health authorities, this goal now appears to be within the reach of the scientific and medical community.