Abstract
Helicobacter pylori infection is one of the most common in man. The bacterium primarily resides in the human stomach, where it plays a significant role in gastric disease. If the spread of H. pylori is to be prevented, an understanding of the transmission process is essential. The oral cavity has been proposed as a reservoir for gastric H. pylori, which has been detected by culture and PCR in both dental plaque and saliva. This review will discuss the evidence for the role of the oral cavity in the transmission of gastric H. pylori. Moreover, the difficulties encountered in addressing this topic, possible directions for future research, and the implications for the dental profession are discussed.
Helicobacter pylori and Gastric Disease
The significant role of Helicobacter pylori in the etiology of gastric disease is now undisputed. In addition to causing type B gastritis, H. pylori has a well-documented role in the development and recurrence of gastric and duodenal ulcers (Balaban and Peura, 1997), which afflict more than 10% of individuals during their lifetime. Moreover, H. pylori has been designated a type I, or definite, carcinogen by the World Health Organization (International Agency for Research on Cancer, 1994), through its association with the development of gastric adenocarcinoma, the second most-commonly diagnosed fatal cancer worldwide. While the evidence is less convincing, H. pylori has also been shown to have been associated with non-Hodgkin’s lymphomas of gastric mucosa-associated lymphoid tissue (Wotherspoon et al., 1991).
Although prevalence is decreasing, H. pylori infection remains one of the most common in man. Based on the results of serological tests that detect the presence of antibodies to H. pylori antigen, the carriage rate of H. pylori is reported to be 20-80% for adults in the developed world, and this figure may rise to more than 90% in the developing world (reviewed by Taylor and Blaser, 1991). These prevalence figures are primarily dependent upon age of the population under study, since, once H. pylori is acquired, it persists into old age unless eradicated by treatment. Crowded living conditions are also associated with increased carriage rates of H. pylori (Mendall et al., 1992), and, consequently, the socio-economic status of study subjects will also influence findings. While gastric H. pylori colonization appears prevalent, it is important to note that full-blown clinical disease develops in a minority of affected patients, where as-yet-undefined conditions, whether host- or bacteria-related, are conducive to disease. However, given its potential to cause disease and its widespread presence among human populations, the impact of H. pylori on health worldwide is still immense. The substantial literature relating to H. pylori since its discovery in 1983 (Marshall and Warren, 1983) serves to emphasize this; yet there still remains considerable controversy on a fundamental issue of the disease process—that is, How and from where is this bacterium acquired? Without such knowledge, we are surely unable to take a crucial step toward instigating measures to interrupt the spread of infection.
One would have expected that, should a significant non-human reservoir for H. pylori exist, it would have already been discovered. However, to date, there has been little success in identifying a consistent non-human source of infection, and one may then conclude that H. pylori is likely to be spread directly from person to person. Certainly, this is supported by increased prevalence in crowded living conditions and clustering of disease within family units (Drumm et al., 1990; Bamford et al., 1993). Nonetheless, the mode of transmission remains something of a mystery. Since blood-borne infection seems improbable, H. pylori must reach the stomach, its primary residence, via the oral cavity. This was most notably demonstrated by the pioneering efforts of Marshall et al.(1985), who, in an attempt to convince the then-many skeptics of the role of H. pylori in gastric disease, ingested H. pylori orally to confirm that it could indeed cause gastritis. The recognition of the oral cavity as the only feasible entry point for gastric H. pylori has led to a considerable interest concerning its role in the transmission process, with speculation of both oral-oral and fecal-oral transmission, with or without an intermediate step in this process. This article will discuss the evidence for a definitive role of the oral cavity in transmission and, more specifically, the role of the oral cavity as a reservoir for gastric H. pylori infection.
Detection of Oral H. pylori
The detection of H. pylori within dental plaque or saliva would provide a foundation for a role of the oral cavity in H. pylori transmission, and the literature addressing this issue is considerable. Before a critical review of the data, however, a word of caution related to methodology is needed. Detection of H. pylori in gastric samples has invariably used a rapid urease test, or the CLO (Campylobacter-like organisms) test, which allows for confirmation of the presence of H. pylori bacteria based on its production of large amounts of urease. Since H. pylori is the only urease-positive bacterium known to reside in the stomach, this provides a suitable detection method for use on gastric samples. However, the oral cavity is residence to several urease-producing species, including Streptococcus spp., Haemophilus spp., and Actinomyces spp., and it is inappropriate to conclude that high urease activity in dental plaque is indicative of the presence of H. pylori. The potential for false-positive results is, thus, considerable, and this rapid urease test alone cannot be recommended for definitive identification of oral H. pylori. Consequently, results of previous studies where this test was used for the detection of oral H. pylori (Gill et al., 1994; Pytko-Polonczyk et al., 1996; Desai et al., 1998; Butt et al., 1999; Avcu et al., 2001) should be treated with caution. On a similar note, while diagnosis of H. pylori in gastric samples may be based on microscopic appearance (Gram-negative, curved or spiral-shaped rods), this is likely to provide low specificity in the case of oral samples, where spirochetes, including Treponema spp., are routinely found. Thus, results of such studies (Butt et al., 2001) should again be considered prudently.
The currently accepted ’gold standard’ for the diagnosis of gastric H. pylori is culture, and this could provide a definitive method for the detection of oral H. pylori. The first documentation of the presence of H. pylori in the oral cavity was reported in 1989, when the bacterium was cultured from the dental plaque of one of 29 patients with H. pylori-associated gastric disease (Krajden et al., 1989). Since then, culture of oral H. pylori has also met with limited success. While the bacterium has been isolated from dental plaque and saliva samples, detection rates have been consistently low, as illustrated in Table 1, although there are exceptions to this (Majmudar et al., 1990; D’Alessandro and Seri, 1992). In both of these studies, confirmation of isolate identity by means of molecular methods was not performed, and, as a result, there is a potential for false-positives (Namavar et al., 2001).
Investigators’ inability to culture oral H. pylori consistently may be attributable to the complexity of the oral microflora together with the fastidious nature of H. pylori, which requires a microaerophilic environment, supplemented media, and up to seven days’ incubation for growth. Under these conditions, overgrowth by other oral species is likely, and direct growth inhibition of H. pylori by oral species in vitro has also been reported (Ishihara et al., 1997). Overcoming this potential difficulty will require a more appropriate selective medium, since current media appear unsatisfactory for the isolation of H. pylori from specimens harboring an abundant flora. Difficulties in culturing H. pylori from oral samples have also been ascribed to the absence of the culturable oral H. pylori and, instead, the presence of a viable coccoid form that is unculturable by conventional techniques (Bode et al., 1993). The coccoid form is believed to predominate in environmentally unfavorable conditions, and, given the inhibitory effects of oral species on H. pylori (Ishihara et al., 1997), this could potentially apply to colonization of the oral cavity. While this coccoid form is not believed to be associated with disease, it has been demonstrated to revert to the infectious rod-shaped form under appropriate conditions (Cellini et al., 1994), and so its presence can still be construed as clinically significant. Finally, it is conceivable, and to some more likely, that H. pylori is rarely cultured from oral samples since it is present in insufficient numbers. If this is indeed the case, as recently reported (Song et al., 2000a), the subsequent ability of oral H. pylori to infect the stomach may be questionable.
With the advances in molecular technology, the potential difficulties with culture have been circumvented by the use of the polymerase chain-reaction (PCR), which permits the amplification of a species-specific region of DNA for subsequent detection by agarose gel electrophoresis. PCR allows for the rapid detection of even small numbers of specific bacteria within a sample and obviates the need for viable organisms. By this method, H. pylori has been detected more frequently in oral samples, but results have been very variable, with a detection rate ranging from 0% to more than 90% (Tables 2, 3 ). Naturally, these results leave the scientific community as confused as ever, and there is a need to establish why such discrepancies exist. In this context, methodological differences between studies can most simply be divided into three areas: study population, oral sample collection, and laboratory detection procedure.
If the status of oral H. pylori reflects that of the stomach, then one would naturally expect that the oral carriage rate would greatly depend upon gastric infection. Consequently, results of studies have been separated as far as possible in Tables 1 and 2 based on this criterion. If H. pylori has a preferred oral niche such that it is unevenly distributed within the oral cavity, the chosen oral sample, whether supra- or subgingival plaque or from oral mucosa, as well as the number of sites sampled, is also likely to influence the results significantly.
While choice of study population and sampling procedure are likely to contribute to the discrepancies found in the prevalence of H. pylori, it is important to use these differences to assist with our understanding of the role of the oral cavity as a reservoir for gastric H. pylori. Perhaps the most critical consideration in scrutinizing study results is that pertaining to the analysis of the oral sample. The lack of uniformity of laboratory procedures is likely to play a crucial role in the reported inconsistencies, and methods require standardization—for example, with respect to DNA extraction and purification and, most notably, choice of primers for PCR. To date, a host of primers has been used, including those based on the urease genes, 16S ribosomal RNA genes, a gene encoding a specific 26-K protein, and randomly selected DNA fragments. Unfortunately, the sensitivity and specificity differ among the primers (Lu et al., 1999; Song et al., 1999), and the most appropriate method for detection of oral H. pylori has yet to be established. Sensitivity and specificity can be increased by the use of a nested PCR method (Bamford et al., 1998; Jiang et al., 1998), in which a second set of primers specific for an internal region of the primary amplicon is used, and by Southern blot hybridization (Jiang et al., 1998), whereby the PCR amplicons separated by gel electrophoresis are blotted onto a membrane and detected by means of a labeled H. pylori-specific probe, but these procedures have not been consistently used. Most importantly, the use of appropriate positive and negative controls in the PCR reactions is essential, particularly with these heavily contaminated samples where low specificity is a potential problem.
The Oral Cavity as a Reservoir for H. pylori
While there appears to be little doubt that H. pylori can be detected in the oral cavity by PCR, one may still question the importance of this finding. First, PCR allows for the detection of low numbers of bacteria, which may be too few to influence gastric health. Moreover, PCR can permit detection of non-viable H. pylori, which is, by definition, non-infectious. If, however, we are to believe that H. pylori detected in the oral cavity is viable and is in sufficient quantity to have pathogenic consequences, then there should be a microenvironment capable of supporting the growth of H. pylori. That is, there should be a niche with, for example, the appropriate pH, redox potential, and nutrient availability to sustain H. pylori. In line with this thinking, it is also being increasingly recognized that the growth of specific species is dependent upon the presence of other species within that microenvironment, such that there is a unique microflora enabling certain species to survive at certain sites. It is unlikely that H. pylori exists in isolation, and, while initial reports have shown it to adhere selectively to Fusobacterium nucleatum and Porphyromonas gingivalis (Ishihara et al., 1997; Andersen et al., 1998), work is needed to characterize the oral microflora associated with those sites in which H. pylori is detected.
If a specific niche does exist for H. pylori, then one would expect to find the prevalence to differ between sites. H. pylori has been detected in supragingival plaque (Allaker et al., 2002; Kim et al., 2000; Song et al., 2000a,b), subgingival plaque (Pustorino et al., 1996; Dowsett et al., 1999; Riggio and Lennon, 1999), saliva (see Tables 1, 3 ) and oral lesions (Mravak-Stipetic et al., 1998; Birek et al., 1999), and on oral mucosa (Mravak-Stipetic et al., 1998; Dowsett et al., 1999; Allaker et al., 2002), but it is difficult to compare prevalence rates between these sites due to variations in methodology among studies. The prevalence of H. pylori in supragingival plaque has been shown to decrease from the molar to the premolar to the incisor region (Song et al., 2000b), suggesting that H. pylori has a distinct distribution within the oral cavity, but, again, this topic needs to be addressed more fully.
Finally, if the oral cavity is a reservoir for H. pylori, it is unclear whether it is a permanent or transient reservoir, but this is likely to have implications in terms of the interpretation of study findings as well as the understanding of its role in gastric infection.
Is the Mouth Significant in Gastric Infection?
Whether permanent or transient in the mouth, the fundamental question is, "Can oral H. pylori be a reservoir for gastric H. pylori?" One might first address this by elucidating whether there is an association between oral and gastric H. pylori carriage. Results of several studies have suggested a positive association between oral and gastric H. pylori detection (Mapstone et al., 1993a; Oshowo et al., 1998; Parsonnet et al., 1999), although, conversely, there are studies that have failed to demonstrate such an association (Olsson et al., 1993; Song et al., 2000b). Interestingly, within these studies, different methods were used for the diagnosis of gastric infection and for the detection of oral H. pylori; hence, with likely differences in sensitivity and specificity between methods, conclusions should be treated cautiously.
While there may well be a positive correlation between oral and gastric H. pylori carriage, more convincing support for the role of the oral cavity in gastric infection would come from demonstrating that the oral cavity and stomach harbor identical or closely related strains of H. pylori. Today, this is invariably addressed by molecular typing, which involves comparison of isolates at the DNA level to look for differences that would indicate different strains. The H. pylori genome displays considerable diversity (Taylor et al., 1992), so that isolates from unrelated patients are likely to be differentiated by these methods. Consequently, it is expected that identification of identical or closely related strains in one or more sites or subjects would suggest transmission between these sites/subjects, or a common source of infection. More specifically, identification of the same strain in the mouth and stomach would support the role of the oral cavity as a reservoir for gastric H. pylori. Molecular typing methods for H. pylori have included restriction fragment length polymorphism (RFLP) analysis, whereby extracted chromosomal DNA is digested with high-frequency cutting restriction endonucleases, and the digested fragments are separated by agarose gel electrophoresis to reveal distinct digest patterns between strains. Since interpretation of the complex band patterns is difficult, this method is often combined with Southern hybridization—for example, using a labeled ribosomal RNA gene probe (ribotyping).
Typing of both oral and gastric H. pylori was first performed by Shames et al.(1989), who used RFLP analysis to determine if oral and gastric isolates from the same subject were identical or closely related. The group reported that H. pylori previously isolated from the dental plaque of one gastric patient (Krajden et al., 1989) was indistinguishable from the gastric isolate, but that strains differed between patients. These results were later confirmed by others (Khandaker et al., 1993; Cellini et al., 1995; Parsonnet et al., 1999). More recently, however, Song et al.(2000c) used nested PCR followed by DNA sequencing of the amplicon to demonstrate that, in the three patients studied, the H. pylori strains detected in the mouth differed from those detected in the gastric samples.
Ferguson et al.(1993) isolated H. pylori from saliva of one of nine patients with gastric H. pylori and again found that it was the same strain as that isolated from the stomach. In this case, both soluble-protein electrophoresis and PCR-restriction endonuclease analysis (PCR-REA) were used for H. pylori typing. In the latter case, the PCR amplicon, rather than chromosomal DNA, is digested with restriction endonucleases, and the simpler pattern is then observed by agarose gel electrophoresis. PCR-REA thus obviates the need for culture, and, with the previously addressed difficulties in culturing oral H. pylori, this facilitates matters immensely. More recently, PCR-REA has been used to type H. pylori detected in both dental plaque and the stomach, and results revealed that, in 13 of 15 cases, the restriction digest patterns of the PCR amplicon from both sites were identical (Oshowo et al., 1998), although only one restriction endonuclease was used, thus limiting the ability to discriminate between strains.
Overall, there is indeed support for the existence of the same strain in both the oral cavity and stomach, but this evidence is far from strong, and additional data are needed. Moreover, there are further issues that should be considered with respect to finding identical or closely related strains of H. pylori in both the oral cavity and the stomach. Since there is evidence that more than one strain of H. pylori may be present in the stomach (van der Ende et al., 1996), the H. pylori strain derived from the oral cavity is not necessarily that responsible for gastric pathology. Furthermore, while finding the same oral and gastric strains in the same subject may suggest that the oral cavity is a reservoir for gastric infection, it may be equally likely that this results from a common source of infection. Alternatively, the source of oral H. pylori may be the stomach, with gastric H. pylori reaching the mouth via vomitus or gastric reflux. This gastro-oral transmission route has certainly been reported, and the subject has been reviewed (Axon, 1995).
How is Oral H. pylori Transmitted?
If we take the next step and assume that the oral cavity is a reservoir for H. pylori, then we can still speculate as to whether transmission is oral-oral or fecal-oral.
Oral-oral transmission would most likely involve transfer of H. pylori between individuals via infected saliva. In support of oral-oral transmission, Megraud (1995) found an increased prevalence of H. pylori in children of African mothers who pre-masticate the infants’ food, and gnotobiotic beagle puppies infected with Helicobacter felis, again with a primarily gastric niche, have been demonstrated to transmit the bacteria by licking uninfected animals (Lee et al., 1991). The use of chopsticks in Chinese immigrants in Australia has been associated with a higher prevalence of H. pylori infection, as diagnosed by serology (Chow et al., 1995), and this is presumed to be the result of sharing of chopsticks and H. pylori transmission through saliva.
Fecal-oral transmission is also feasible. Certainly, H. pylori has been cultured from stools, both of children in The Gambia (Thomas et al., 1992), and of adults in England (Kelly et al., 1994) and the USA (Parsonnet et al., 1999). However, as with oral samples, culturing H. pylori from these samples harboring an abundant flora has met with only limited success, and this problem is only exacerbated by the potentially toxic effects of feces on the bacterium. H. pylori has also been detected by PCR in fecal samples (Mapstone et al., 1993b; Shimada et al., 1994; Li et al., 1996; Namavar et al., 2001). Again, detection rates have been consistently low, generally even lower than those reported for oral samples, and, while this could challenge the potential significance of fecal-oral transmission, it has also been attributed in part to inhibition of the PCR reaction by fecal contaminants (Bamford et al., 1998). Fecal-oral transmission is indirectly supported by results of studies performed in South America, which have suggested that water supplies may harbor H. pylori (Klein et al., 1991; Hulten et al., 1996), and consumption of raw vegetables fertilized with human feces is associated with an increased risk of infection (Hopkins et al., 1993). The potential for flies as vectors for enteric disease has been known for years, and, if fecal-oral transmission does occur, the housefly could play a role in this capacity. Indeed, PCR has detected H. pylori in the housefly (Grubel et al., 1998), although not consistently (Osato et al., 1998).
Since there is support for both oral-oral or fecal-oral transmission, it has been proposed that H. pylori can be spread in both manners, and that the principal mode of transmission is likely to be population-dependent. Whether oral-oral or fecal-oral, there is potential for an intermediate step in H. pylori transmission. The possible roles of chopsticks in oral-oral transmission and of water or the housefly in fecal-oral transmission have been mentioned. H. pylori has also been detected, by PCR, beneath the fingernails of indigenous Indians from rural Guatemala (Dowsett et al., 1999), and, more importantly, there was a significant positive association between tongue and fingernail carriage, suggesting a role of finger-mouth contact in the transmission process.
The Potential Significance of Oral H. pylori
Not only is the oral cavity a potential reservoir for H. pylori infection of the stomach, but also it is a potential reservoir for re-infection, so that even if oral H. pylori is initially acquired from the stomach via a gastro-oral route, it could still be detrimental to gastric health. Relapse following treatment of gastric infection is common, with re-isolation of the initial infecting strain. It has been proposed that this is due to re-colonization of the gastric mucosa by H. pylori originating from dental plaque, an environment that may be inaccessible to systemic antibiotics. In support of this premise, it has been demonstrated that cure of gastric infection does not consistently promote disappearance of H. pylori from the oral cavity (Dore-Davin et al., 1999), and the success of gastric eradication is significantly lower in oral H. pylori-positive than -negative cases (Miyabayashi et al., 2000).
The impact of H. pylori oral carriage on the gastric patient has been discussed, but how do these data influence the dental profession? One might hypothesize that those working close to the oral cavity may be at increased risk of H. pylori infection. However, while studies are few, there appears to be little evidence in support of this. Japanese dentists have been reported to be at higher risk for H. pylori infection than age-matched controls, although, for indiscernible reasons, the risk appears to be greater in younger dentists (Honda et al., 2001). By contrast, data reported by Banatvala et al.(1995) indicated that dentists do not appear to have a higher carriage rate than pre-clinical dental students. In terms of patient care, H. pylori seropositivity has been significantly associated with higher plaque levels (Peach et al., 1997) and less frequent visits to the dentists (Gasbarrini et al., 1995), although this has not been consistently found (Nguyen et al., 1993). However, one can do little harm in reinforcing oral hygiene measures and encouraging dental attendance, particularly given the recent interest in an association between oral and general health.
In summary, there is increasing evidence for a role of the oral cavity in transmission of H. pylori, and new methods of detection continue to support this (Young et al., 2001). However, this subject requires considerably more investigation before definite involvement of the oral cavity can be confirmed and preventive measures can be tailored toward the prevention of oral spread. More specifically, more extensive studies to confirm identical or closely related strains in the mouth and stomach are required. Also, a specific niche for H. pylori should be identified in the oral cavity, and the association of H. pylori with other members of the oral microflora needs to be addressed. More successful culturing techniques for oral H. pylori are also needed to confirm its viability and thus infectious status.
For those skeptics among us, if H. pylori were a typical member of the oral flora, one would expect it to be recovered more frequently than has so far been reported, but it may be that oral carriage is population-dependent or is only transient. This will need to be addressed in appropriate population-comparison studies and longitudinal investigations, respectively. However, if the oral cavity is a reservoir for gastric infection, in even a minority of individuals, this should be sufficient to warrant a preventive approach that encompasses consideration of the oral reservoir.
Footnotes
Acknowledgements
We are grateful to Dr. Don LeBlanc for his constructive comments and suggestions.