Helicobacter pylori and Nonulcer Dyspepsia: Is There a Relation?

In this issue of the Journal of Pediatric Gastroenterology and Nutrition, Uc and Chong (1) touch on a controversial topic in the clinical management of Helicobacter pylori, i.e., the association of H. pylori and nonulcer dyspepsia. The current recommendation from the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) is that H. pylori–infected children with nonulcer dyspepsia or recurrent abdominal pain, or both, do not require treatment (2). This recommendation is based on the lack of proof that infection with H. pylori is a significant cause of abdominal pain. In the NASPGHAN review of 2,715 children with H. pylori, 5% to 17% had abdominal pain. However, 5% to 29% of patients without abdominal pain also had H. pylori infection. Two types of studies have sought to address the relation of H. pylori to pain and dyspepsia in children and adults. The first approach seeks simply to demonstrate an association between H. pylori infection and abdominal pain. There are numerous problems with the existing studies of this type. To prove an association there should be a higher incidence of H. pylori infection in the dyspeptic group than in a comparable group without dyspeptic symptoms. Matching the comparison groups for age and socioeconomic class is important. Recent use of antibiotics may also be an important factor but is rarely mentioned or assessed. Patient selection in children is important to isolate the group most likely to have the symptoms in question. For example, the largest study seeking to relate H. pylori to abdominal pain in childhood evaluated a group of preschool children and may not adequately address the issue because recurrent abdominal pain is usually found in older children (3). Given the high frequency of the symptoms and the low frequency of H. pylori infection in North America, it is difficult to accumulate sufficient cases and controls to obtain statistically significant results in studies such as these. Adding to the difficulty is that widely varying definitions of abdominal pain have been used. Case-controlled studies using day-surgery populations or clinic populations as controls may not be appropriate because the day-surgery populations tend to be younger and the clinic populations are not truly “healthy” children. Even without these reservations the existing studies generally do not support the conclusion that there is an association between H. pylori and dyspepsia. The second type of study attempt proves an association between H. pylori and abdominal pain or dyspepsia by showing that eradication of H. pylori results in resolution of symptoms. Reliable studies of this type are also hard to find. By definition, all patients in these studies have H. pylori infection. To be reliable, documentation of H. pylori should not rest on serologic data, but on demonstration of active infection by endoscopy or urease breath test. Patients should be randomized to treated and untreated groups and should be matched for significant variables, such as age and socioeconomic status. The definitions of abdominal pain or dyspepsia must be uniform and reliable. Patients with confounding diagnoses, such as acid reflux, should be eliminated from the study. Endoscopy should be performed to rule out ulcer disease. The use of antibiotics and acid suppression before treatment needs to be taken into consideration. Patients receiving other treatments, such as antacids or bismuth, that may decrease bacterial load and pain even if H. pylori is not eradicated should be eliminated from the study. To relate improvement in abdominal pain to eradication of H. pylori, eradication must be reliably documented by use of endoscopy or breath test. The timing of posttreatment assessment of symptoms is another problem in these studies. Although eradication of H. pylori may be fairly rapid, resolution of symptoms may require healing of gastritis, which may take longer. Assessing symptoms too quickly after treatment may decrease the apparent effectiveness of therapy. In addition, despite a fairly high rate of treatment failure, simply decreasing the level of infection without full eradiation may have a beneficial impact on pain. Because most investigators are uncomfortable not treating the control group, the comparison group in most studies is composed of treated subjects in whom H. pylori was not eradicated. Most studies report an improvement in abdominal pain or dyspepsia, or both, with treatment, but the studies have been small without a control group, and eradication has not been documented (4–6). The study by Uc and Chong is a treatment study that comprising 13 patients with nonulcer dyspepsia (1). The diagnosis of H. pylori was confirmed histologically in all subjects. Patients were treated with triple therapy for 1 week. A standardized questionnaire designed to quantitate dyspeptic symptoms was used before and after treatment. There was significant improvement in all patients 2 to 4 weeks after therapy. At follow-up evaluation 4 to 24 months after therapy, 8 of the 13 patients had sustained remission of symptoms. This study shares the problems of other studies of H. pylori and dyspepsia. There was no control group and therefore we do not know the percentage of patients who would have improved in the absence of treatment. The number of patients evaluated in this study is small. Eradication was supported by decreasing titers as seen during serology, but not confirmed by histology or breath test. A decrease in the number of symptoms could therefore be related to a decrease in the H. pylori colonialization without eradication. Despite these reservations, the care with which patients were evaluated and followed-up clinically, using a validated dyspepsia questionnaire, makes it an important contribution to our experience with non-ulcer dyspepsia in children infected with H. pylori. A recent meta-analysis of five adult studies emphasizes some of the problems in eradication studies (7). One hundred studies that addressed the topic were evaluated, but only five were adequate for the meta-analysis (a powerful statement about the sorry state of research in this area). The other studies were excluded because there was no description of randomization or blinding, no assessment of treatment compliance, and poor presentation of results. The meta-analysis only included studies that validated eradication (7). One problem is that three of the five selected studies used single therapy and only evaluated short-term outcomes. The other two studies used triple therapy and observed patients for 1 year. These latter two studies showed a less favorable impact of treatment on symptoms of nonulcer dyspepsia than the three short-term studies. In a study by McCarthy et al. (8) all dyspeptic patients with H. pylori improved at 3 months post-treatment whether they were treated with bismuth alone, metronidazole and amoxycillin, or bismuth, amoxycillin, and metronidazole; yet only triple therapy resulted in extended eradication of the bacteria. Blum et al. (9) recently published information from a randomized double-blind placebo controlled trial with a 12-month follow-up period. This study seems to provide level 1 evidence that eradication of H. pylori infection does not result in resolution of the symptoms of dyspepsia. The study included adults from 11 different countries. Dyspepsia was defined as pain or discomfort centered in the upper abdomen for more than 3 months. Patients were excluded if they had received antibiotics 4 weeks before the study or acid suppressors 7 days before the study. Symptoms were uniformly recorded for 7 days before randomization, and patients had to have had pain on 3 of the 7 days before randomization. Patients with acid reflux or ulcer disease were excluded. The positive [13C] breath test was used to document active infection. Patients were randomized to receive triple therapy with omeprazole, amoxicillin, and clarithromycin or omeprazole alone for 1 week. The groups were well matched for weight, race, history of smoking, and alcohol consumption. Compliance was monitored. The participants were followed-up at 1, 3, 6, 9, and 12 months, and eradication was documented at 3 and 12 months by endoscopy and the breath test. A weak antacid was given when necessary, and the amount of antacid consumed was recorded. Relief of symptoms and quality of life were determined. This study takes into account the majority of my concerns about treatment studies, and it showed no difference in quality of life or resolution of symptoms between the groups. Even if the data were reanalyzed with respect to H. pylori eradication, there would be no difference in quality of life or resolution of symptoms. However, there was a major difference in the outcome of gastritis. Gastritis continued in those patients with continued H. pylori infection. Gastritis healed in those patients in whom H. pylori was eradicated. Part of the reason an effect of eradication on symptoms was not detected may be that both groups experienced improvement in symptoms and quality of life over the course of the study. Even with this qualification, this study is fairly convincing that eradication of H. pylori does not cure nonulcer dyspepsia, although it does cure gastritis. The mean age of the patients in this study was 50 years and therefore the results may not be applicable to children. These patients had probably acquired H. pylori in childhood and may have had chronic gastritis. They may have had changes in their physiology that did not change after eradication of H. pylori. Clearly, more good studies in children are needed to address this issue. Even data from children suggest that chronic H. pylori infection is asymptomatic. Acute infection during childhood seems to be associated with an acute and transient gastroenteritis syndrome, which is well documented in a study from Texas by Opekun et al. (10). Epidemiologic data suggests that most people who acquire H. pylori infection in childhood probably have chronic gastritis, which in most does not cause symptoms. Rather than asking whether H. pylori infection causes pain, the important question perhaps should be, why do so many people with chronic gastritis remain symptom free? Acute infection with H. pylori results in hypochlorhydria, which seems appears necessary to allow the organism to colonize and survive. There may follow years of normal acidity associated with a feedback increase in gastrin and a decrease in somatostatin. Long-term infection of the corpus of the stomach with H. pylori, however, via its associated gastric atrophy may result in decreased acid secretion (11). Decreased gastric acid production may be associated with decreased pain, but it could contribute to dyspeptic symptoms by promoting bacterial overgrowth. No studies in children have evaluated gastric acidity or motility in H. pylori gastritis by comparing patients with dyspepsia to those without. Hypergastrinemia has been documented in H. pylori gastritis, but it does not seem to correlate with recurrent abdominal pain (12). The data supporting eradication of H. pylori to treat nonulcer dyspepsia are weak at best. However, there are good data that treatment of H. pylori decreases the risk of duodenal and gastric ulcers in adults and children (2). There are also data that treatment of H. pylori promotes resolution of mucosal associated lymphoid tissue (MALT) lymphomas (2). Data just released from Japan on the role of H. pylori in the development of gastric cancer are convincing. Uemura et al. (13) observed a large group of patients with H. pylori and a group with ulcers and dyspepsia that did not have H. pylori infection for a mean of 7.8 years. Gastric cancer developed in 2.9% of the H. pylori–positive patients and in none of the patients who were H. pylori negative. Those patients with severe gastric atrophy, corpus predominant gastritis, and intestinal metaplasia were at a significantly greater risk for the development of gastric cancer. Gastric cancer was detected in 4.7% of patients with nonulcer dyspepsia, in 3.4% of patients with gastric ulcers, in 2.2% of patients with hyperplastic polyps, and in 0% of patients with duodenal ulcers. The patients with duodenal ulcers had antral predominant inflammation and no atrophy. Cancer did not develop in patients in whom H. pylori had been eradicated, although the duration of follow-up for this group was shorter. These results suggest that H. pylori infection associated with nonulcer dyspepsia can progress to gastric cancer. These data lend importance to studies such as the current one by Uc et al. (1). However, it is important to remember that the rate of gastric cancer is low and the number of children with dyspepsia with or without H. pylori infection is great. There is a need identify those patients with H. pylori gastritis who might be at risk for cancer. If we assume that patients with duodenal ulcer have increased gastric acidity, these results suggest that gastric acidity may confine H. pylori gastritis to the antrum and result in protection from gastric cancer. It is fairly well established that a decreasing amount of acid, especially with proton pump inhibitors, may promote infection of the body of the stomach with H. pylori. This development of corpus gastritis did not occur with treatment using omeprazole in the absence of H. pylori infection; therefore, it does not seem to be related simply to the use of PPI (14). Many patients with nonulcer dyspepsia are treated with acid suppressers for a prolonged period. I would suggest that, if we are considering prolonged acid suppression in a patient, we need to determine the H. pylori status of the patient at some point during therapy. If the patient is positive for H. pylori infection, H. pylori should be eradicated to avoid promoting infection of the corpus and possible subsequent intestinal metaplasia, gastric atrophy, or gastric cancer. So far there have been few reports of gastric atrophy in children, but this could be related to insufficient sampling of the body or insufficient follow-up. In one 2-year follow-up study of children with persistent H. pylori infection, progression to chronic gastritis occurred in a significant proportion (15). In a 10-year follow-up of adults with H. pylori gastritis, there was a significant incidence of gastric atrophy (16). These lesions apparently take time to develop. Current studies are investigating whether particular H. pylori cytotoxins are associated with an increased risk of gastric cancer. CagA-positive strains are not associated with nonulcer dyspepsia, but they may have some predictive value for the development of gastric cancer (17–19). It will take time to determine which of these indicators will be helpful in predicting risk for gastric cancer. Pediatric research in this field needs to define whether H. pylori gastritis is a disease that needs treatment. Data suggest that H. pylori infection results in chronic gastritis, which can progress in symptomatic and asymptomatic patients both. As pediatricians we want to prevent the possible long-term complications associated with H. pylori acquired in childhood. We need to define the risk factors for gastric cancer. Long-term, well-controlled, multicenter trials with large numbers of patients will be required so that we will be able to make informed decisions about treatment of this fascinating infection.