Article Text


Antagonist: Should we eradicate Helicobacter pyloribefore long term antireflux therapy?
  1. J W Freston
  1. University of Connecticut Health Center, Farmington, Connecticut, USA
  1. Dr J W Freston, Office of Clinical Research, University of Connecticut Health Center, 263 Farmington Ave, Farmington CT, 06030-1111, USA.Freston{at}

Statistics from

Patient management recommendations that might become health care policy should be based on the highest level of evidence. The evidence supporting the recommendation to “test and treat” forHelicobacter pylori infection before starting long term antireflux treatment falls well short of this standard. This recommendation arose directly from the report by Kuipers and colleagues1 in 1996 that the prevalence of corpus glandular atrophy increased during chronic omeprazole treatment in patients infected with H pylori. We were pleased to have another reason to eradicate H pylori; the prevailing wisdom was that no H pylori should go unpunished. This aggressive approach has been challenged, particularly in the case of patients with gastro-oesophageal reflux disease (GORD). Approximately 42% of adults in the USA suffer from GORD.2 Nearly 7% experience heartburn on a daily basis,3 making them potential candidates for proton pump inhibitor (PPI) therapy and, therefore, testing for H pylori infection. Apart from the obvious economic and antimicrobial resistance implications, growing evidence indicates that H pylori infection protects against the complications of GORD, including Barrett's oesophagus, dysplasia, and oesophageal adenocarcinoma.4The death rate for oesophageal cancer now exceeds that of gastric cancer in males in the USA.5 Given these considerations, the evidence supporting the test and treat approach should be substantial. Instead, it is weak

Uncontrolled studies

The study of Kuipers et al is often misrepresented as a controlled trial. It began as an uncontrolled descriptive study of corpus biopsy changes during chronic omeprazole treatment in H pylori infected GORD patients living in the Netherlands. After the fact, the authors identified a group of Swedish patients treated with antireflux surgery and used their gastric biopsy data for comparison. Lack of randomisation to an appropriate control group introduced bias. Omeprazole treated patients were older than surgery patients by an average of nine years. Glandular atrophy increases with age: its prevalence is twice as high inH pylori infected subjects aged 57–68 years than in those aged 46–56 years.6 A comparison of glandular atrophy in two different populations may also be significant because of the influence genetic, socioeconomic, nutritional, and dietary factors have on the development of atrophy. The most troubling aspect of Kuipers' study was a total absence of atrophy at baseline in the omeprazole cohort combined with the absence of any increase in atrophy over a period of five years in the surgery cohort. These findings differ from all other studies that show at least some atrophy in populations with H pylori gastritis and at least some progression over time if the infection is untreated (see below). This combination of unusual findings was responsible for the difference reaching statistical significance.

Other uncontrolled studies have yielded conflicting results: some reported an increase in atrophy in infected GORD patients treated with PPIs7-9 while others did not.10 11 The recent study of Geboes and colleagues11 from the Netherlands was particularly revealing because they found no increase in the prevalence of atrophy after five years of PPI therapy in a Dutch population that was similar to the PPI cohort in Kuipers' study. This issue clearly cannot be resolved on the basis of uncontrolled studies.

Controlled trials

Lundell and colleagues12 randomised GORD patients to either chronic omeprazole treatment or antireflux surgery. They found no significant difference between the groups in the prevalence of glandular atrophy. The US FDA reviewed the data linking PPI therapy to glandular atrophy. The review included the studies of Kuiperset al and Lundell et al, as well as the unpublished randomised controlled trials of PPI maintenance therapy for 12–60 months. The comparator groups were placebo, ranitidine, and antireflux surgery. The prevalence of atrophy did not differ significantly between the groups. The FDA concluded that PPIs did not accelerate the development of gastric atrophy, intestinal metaplasia, or cancer.13 The FDA also concluded that PPI labels should not contain a recommendation to test and treat before undertaking PPI therapy.

Commentary on quality of evidence

Among their myriad shortcomings, uncontrolled trials are particularly limited in their ability to discern causes of time dependent changes, such as glandular atrophy. Only the study of Lundellet al and those analysed by the FDA were controlled. They must be given more credence, despite their shortcomings, which are dwarfed by those of uncontrolled trials. Kuipers' study framed a hypothesis which has now been tested and rejected on the basis of controlled trials. The effect of PPIs on glandular atrophy is either non-existent or so trivial that it does not deserve the attention the issue continues to receive. In any case, the evidence hardly justifies a broad recommendation to test and treat the huge pool of GORD patients who will benefit from long term PPI therapy.

Reasons for not eradicatingH pylori   before long term antireflux therapy

  • Patient management recommendations, especially those that might become health care policy, should be based on the highest level of evidence.

  • The recommendation to eradicate H pylori infection before embarking on long term antireflux therapy is based on an uncontrolled observation that corpus glandular atrophy appeared to increase during PPI therapy. Evidence from controlled trials does not support the observation. As the effect is either non-existent or so small as to be clinically meaningless, the recommendation is inappropriate.


View Abstract

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.