INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Helicobacter pylori infection has been documented in adults and children worldwide, with a prevalence of 20 to 60% depending on age, geographic location, and socioeconomic conditions (12, 19). The majority of affected persons are asymptomatic despite evidence of chronic antral gastritis, and the relevance of this pathology to chronic abdominal pain is questionable (18, 29). Researchers have reported consistent evidence of an association of H. pylori infection with duodenal ulcer (23). Although H. pylori infection is neither necessary nor sufficient for ulcer development (13), its eradication nevertheless markedly reduces ulcer recurrence (29).

A potential virulence determinant of H. pylori is the cagA gene product, which is found in approximately 60% of isolates from adults (5). The antigen, a hydrophilic surface exposed protein of 128 kDa, is itself devoid of cytotoxic activity but is strongly associated with it, possibly by the transcription, folding, export, or other function of the toxin. The cagA gene, which encodes this protein, has been cloned and sequenced (5), and the in vivo expression of cagA mRNA in gastric mucosal biopsy samples has been noted (22). In addition, mucosal immunoglobulin A (IgA) recognition of the protein (7) and the presence of serum antibodies to the CagA protein are strongly associated with the presence of active gastritis and duodenal ulcer (5, 6, 31), and they may pose an increased risk for the development of atrophic gastritis (17) and intestinal metaplasia and gastric cancer (2). Thus, H. pylori strains may be divided into at least two subgroups based on the expression (type I) or nonexpression (type II) of CagA and the cytotoxin. Type I strains are variable, with about 30% of isolates possessing either CagA or cytotoxin activity (32).

The aim of the present study was to determine the relevance of the presence of CagA to the clinical picture and outcome of H. pylori infection in children.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

The study population consisted of 104 consecutive children with endoscopically diagnosed H. pylori infection who presented to our center between June 1989 and June 1995. Indications for endoscopy were recurrent abdominal pain (RAP) (defined as at least three episodes of abdominal pain over a period of at least 3 months of sufficient severity to interrupt normal daily activities), ulcer pain (defined as hunger pain, nocturnal pain and pyrosis), gastroduodenal bleeding, and suspicion or follow-up of gluten-sensitive enteropathy (GSE). The relevant clinical information included a family history (first-degree relatives) of peptic disease and recent antibiotic and antacid therapy.

Diagnosis. Endoscopy was carried out under sedation (intravenous Demerol HCl [1 mg/kg], midazolam [0.1 to 0.2 mg/kg]) with a GIF P3 or XQ20 fiberscope (Olympus). Findings were categorized as gastric or duodenal ulcer, nodular gastritis, or normal. One duodenal bulb and one gastric body biopsy sample were obtained for histologic study, and three antral biopsy samples were taken, one for histologic study, one for a 24-hour urease test, and one for culture. For histology, the samples were fixed in Bouin's solution and stained with hematoxylin-eosin, Giemsa stain, and Gram stains. Gastritis was classified as described by Whitehead (30). The diagnosis of H. pylori infection was based on the presence of typical bacilli on histology and a positive urease test and culture.

Treatment. Patients were treated before the serology findings were known. Up to September 1994, H. pylori infection was routinely treated in our department with a combination of amoxicillin (50 mg/kg/day) for 3 weeks and metronidazole (20 mg/kg/day) for 2 weeks and, in addition, with bismuth subcitrate (De-Nol) (120 to 240 mg four times per day) for 6 weeks or H2 blockers (cimetidine) (20 mg/kg/day) for 6 weeks. Thereafter, the protocol was changed, and patients received a combination of clarithromycin (15 mg/kg/day), amoxicillin (50 mg/kg/day), and omeprazole (20 mg/day) for 2 weeks. Those with persistent infection and severe symptoms received more than one therapeutic trial. Asymptomatic patients with GSE were not treated.

Outcome. Outcome was defined according to clinical and bacteriological parameters. Clinically, patients were considered to be either symptomatic or asymptomatic (no symptoms or occasional discomfort of less than one episode per month). No attempt was made to define the severity of symptoms. Healing of duodenal ulcers was confirmed by endoscopy. Bacteriologically, patients were considered cured (eradication of infection) when endoscopical H. pylori-related tests were negative. For children who refused to undergo repeated endoscopy, ¹³C-UBT (urea breath test) was used to determine the final infectious status. Breath samples were collected before and 30 min after the intake of 75 mg of ¹³C-labeled urea, and the ratio of ¹³CO₂/¹²CO₂ was measured by isotope ratio mass spectrometry. A value of >5 per million was considered positive for H. pylori infection.

Serology. Serum samples were obtained at diagnosis and follow-up endoscopies and stored at 20°C. For those patients who did not undergo follow-up endoscopy, serum samples were taken within three days of the ¹³C-UBT. Quantitative anti-H. pylori IgG antibodies were determined by electroimmunoassay with the COBAS-CORE (Hoffman-La Roche) fully automated methods. Samples from each patient were tested concurrently. The method was validated in our laboratory by a pilot study which yielded a sensitivity of 94%, specificity of 90%, and positive and negative predictive values of 100 and 90%, respectively.

Statistical analysis. The distributions of demographic, clinical, and laboratory data for the patients with positive and negative CagA results were compared with the two-sample t test for normally distributed continuous variables and the Mann-Whitney test for nonparametric distribution. The chi-square test was applied for categorical data. Multivariate logistic regression models were used to determine which factors were independently related to persistence of infection.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

CagA status. Of the 104 patients included in the study, 45 (43%) were seropositive for the CagA antigen (group I) and 59 (57%) were seronegative (group II). The relevant clinical data for the two groups are given in Table 1. No between-group differences were observed for patient age and family history of peptic ulcer disease. The percentages of asymptomatic (GSE) patients were not significantly different between the groups. When these patients were excluded, group I showed a significantly longer duration of symptoms (² = 4.0; P = 0.02) and more ulcer pain (P = 0.05); there was no difference in RAP.

Endoscopic and histologic findings. Duodenal ulcer was noted in 21 patients (20%), 17 in group I and 4 in group II (² = 20.88; P < 0.01). Seven of the group I patients also had gastrointestinal bleeding. Nodular gastritis was equally present in the two groups, and none of these patients had evidence of gastric ulcer or erosion. Chronic active gastritis on histology was significantly more frequent in group I (82 versus 32%; ² = 24.65; P < 0.01), and only 5% of group I patients had normal gastric histologic findings, compared to 18% of group II patients.

Anti-H. pylori therapy. Eighty-three children were treated (Table 2). As the treatment regimen was not uniform, we can conclude only that most of the children had at least one treatment protocol. Significantly more patients in group I required two or more courses of therapy (11 versus 3; ² = 8.2; P = 0.02). The indication for additional therapy was recurrent duodenal ulcer or severe symptoms, along with failure to eradicate the infection.

CagA status and eradication of H. pylori. The relevant outcome data are shown in Table 2. The status of H. pylori infection (at the end of follow-up) was determined by endoscopy in 60 patients (32 in group I and 28 in group II), and by ¹³C-UBT in 33 patients. Eradication of infection was achieved in 46 of 93 of patients (50%). Significantly more patients in group I than in group II had evidence of ongoing H. pylori infection (70 versus 36%; ² = 20.0; P < 0.01).

CagA status and symptoms. Ninety-three patients, 40 in group I (89%) and 53 in group II (90%), were available for follow-up; 11 patients could not be traced or refused clinic visits. The follow-up time in both groups was 27.1 ± 18.1 months. Forty-five percent of the patients were monitored for 24 months or more (Table 3).

CagA status and anti-H. pylori antibody titer. The mean anti-H. pylori antibody titers of the 104 patients at diagnosis were significantly higher in group I than in group II (74.3 ± 58.7 versus 38.7 ± 50.5 U/ml; P < 0.01). False-negative results occurred in 7% of the whole study group; all of these results were positive by the immunoblot method.

Effect of CagA status on immunoassay bands. Patients were considered CagA positive or negative according to the presence of the 116-kDa band in immunoblotting. The responses to the other H. pylori antigens on immunoblotting were similar in the two groups, with the exception of the 89-kDa band, which was present in 50% of group I patients but in only 8% of group II patients.

CagA status and duration of infection and symptoms. Comparison of the serum immunoblotting studies between diagnosis and end of follow-up showed no change in the CagA (116-kDa) reactivity in individual patients with persistent infection in both groups (i.e., none switched groups).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our study demonstrates a strong association between the presence of serum anti-CagA antibodies and severe clinical and histologic H. pylori disease. Seropositive patients had prolonged abdominal pain of greater clinical relevance (ulcer pain) and a higher prevalence of duodenal ulcer; 40% also had complications with at least one episode of bleeding, and 80% had histologically active gastritis.

An association between CagA-positive infective strains and the presence of severe gastritis and gastric or duodenal ulcer has previously been documented in adults (5, 6, 7, 22, 31), but no such findings were noted in children (3, 9). Some authors found an association only with histologically severe gastritis and vomiting (15). Recently, Queiroz et al. (24) found that in Brazilian children who underwent endoscopy for abdominal pain, CagA-positive status was associated with a more marked macroscopic gastritis and a greater inflammatory infiltrate compared to CagA-negative status.

The diagnosis of infection with CagA-positive strains is based on the detection of the gene in bacterial isolates or the presence of specific IgG antibodies against the CagA protein. Serum testing is more sensitive for the detection of CagA-positive strains than random selection of bacterial isolates (6), because patients may be infected by multiple strains (1). The Helicoblot 2.0 method used in our study provides qualitative information on the IgG response to various H. pylori antigens. Studies with both adults (11) and children (28) with H. pylori infection have found the Helicoblot to be equivalent or superior to other serological tests, with the advantage of providing information on CagA-positive strains. Using this method with 68 children, Oderda et al. (20) demonstrated a strong association of CagA seropositivity with duodenal ulcer (30%), nodular gastritis (53%), and active chronic gastritis (60%). These results are similar to those observed in the present study.

Our CagA-positive children with newly diagnosed or persistent H. pylori infection had significantly higher titers of anti-H. pylori antibodies on electroimmunoassay than the seronegative children. This finding is most probably related to the more active gastric mucosal inflammation in the first group, although a specific CagA antigen-related reaction could be responsible for the more intense immune response.

The rate of CagA positivity in the children with duodenal ulcer, noted here and in the study of Oderda et al. (20), was similar to that found in adults (80 to 100%) by bacteriology or serology and in all (100%) CagA-positive Brazilian children (24). Yet, none of these studies reported a high incidence of peptic ulcer in affected patients. This supports the notion that genetic factors strongly influence not only the acquisition but also the outcome of gastroduodenal disease. Further evidence was provided by Van den Borre et al. (28), who found an 85% prevalence of CagA-positive infective strains in children of Belgian origin (91% with duodenal ulcer) but only a 35% prevalence in Moroccan children (30% with duodenal ulcer) living in the same region under similar conditions (28).

In Japan, where CagA-positive strains are common in persons of all ages, Kato et al. (16) found that a high prevalence of CagA-positive H. pylori strains (80 to 90%) in children was not associated with nodular gastritis or peptic ulcer disease. Yet, in Brazil, children with and without duodenal ulcers had more intense macroscopic gastritis associated with a positive CagA status (24).

Our study is the first to correlate the long-term outcome of H. pylori infection with CagA expression. We found that patients positive for CagA had a significantly higher incidence of persistent infection (70 versus 36%; P = 0.002), irrespective of family history, duration of disease, length of follow-up, or therapy.

Our patients received identical treatment without knowledge of the CagA status. The regimens used were those accepted at the time for eradicating infection in 60 to 90% of patients in controlled trials (4). The overall eradication rate in this study was 50%. In the remainder of the patients, the infection persisted even after repeated therapeutic trials including new and effective drugs. This needs to be confirmed in a prospective controlled trial.

The failure of therapy to cure H. pylori infection is most frequently the result of low compliance (14), which is definitely related to the prescription of multiple drugs, some with side effects, for prolonged periods of time. This may partly explain our overall low eradication rate, although other factors are undoubtedly involved, considering the significant difference in the success rates between the CagA-negative group (64%) and the CagA-positive group (30%).

An additional risk factor is antibiotic resistance, specifically resistance to metronidazole and clarithromycin, which is known to have a substantial effect on the success of therapy (10). This problem was not addressed in our study, nor, to the best of our knowledge, is there any information available on the specific antibiotic resistance of CagA-positive bacterial strains.

Finally, the severity of the mucosal inflammation and the bacterial load need to be taken into account (8). This was true for our patients as well. A significant increase in active antral inflammations is consistently found in patients infected with CagA-positive strains. The quantitation of inflammation in both antral and body gastric mucosa and the degree of bacterial colonization are presently being investigated.

In examining the potential impact of H. pylori eradication on symptoms, it is important to distinguish between patients with duodenal ulcer and those with gastritis only. In the absence of peptic ulcer, H. pylori infection is not accepted as an important source of symptoms in children (18, 26, 27), although strong differences of opinion exist (21, 25). No correlation between symptom relief and CagA-positive serology was demonstrated in this study; however, we found that patients with persistent CagA-positive infection had a greater tendency to exhibit persistent symptoms (64%) than those with CagA-negative ongoing infection (42%). Thus, the degree of active mucosal inflammation may be an underlying factor in the persistent dyspeptic symptoms. Analysis of the results on the efficacy of therapy in relieving symptoms in H. pylori-infected individuals will clearly be improved when study of the CagA strains is included.

Presently, therapy of H. pylori infection is universally recommended only for patients with duodenal ulcer (19). However, confirmation of the association of CagA positivity and severe gastrointestinal pathology in further studies will highlight the clinical importance of determining the CagA status of H. pylori-infected individuals in order to target those who require therapy and follow-up.