helicobacter pylori - The BMJ

10 downloads 20 Views 94KB Size Report
Medicine, Mayo Clinic Florida, Jacksonville, Florida, FL32224,. USA; visiting ... College London Medical School, London N20 8AS [email protected]

head to head

Does Helicobacter pylori really cause duodenal ulcers? The link between duodenal ulcer and Helicobacter pylori has revolutionised treatment. Alexander Ford and Nicholas Talley argue that the association is causal, but Michael Hobsley and colleagues believe acid secretion is the key

Alexander C Ford lecturer in medicine, Department of Academic Medicine, St James’s University Hospital, Leeds LS9 7TF alex[email protected] Nicholas J Talley professor of medicine, Department of Medicine, Mayo Clinic Florida, Jacksonville, Florida, FL32224, USA; visiting professor of medicine, Department of Medicine, University of Sydney, Nepean Hospital, Sydney, Australia,

Helicobacter pylori infects the human stomach and is acquired predominantly in childhood.1 Infection is from person to person, and evidence points to transmission through the gastrooral route.2 Gastric biopsy shows that all infected people have either antral or corpus predominant gastritis.3 Which of the two develops probably depends on the person’s parietal cell mass at the time of infection. Evidence Although causation can never be proved (hypotheses can only be dispelled), evidence for the causative role of H pylori in duodenal ulcer is remarkably compelling and fulfils the Bradford Hill criteria—for example, showing a strong, consistent, specific, and temporal association (see box on bmj.com). The National Institutes of Health judged H pylori to be the main cause of the condition in 1994. A causal association between H pylori and chronic

Michael Hobsley emeritus professor of surgery, University College London Medical School, London N20 8AS [email protected] Frank I Tovey honorary senior research associate, University College London Medical School, London N20 8AS Karna Dev Bardhan consultant physician and gastroenterologist , Rotherham General Hospital, Rotherham John Holton reader in clinical microbiology, Windeyer Institute for Medical Sciences, University College London


Helicobacter pylori and duodenal ulcer are linked. However, association does not prove causation. An association between A and B may mean that A causes B, B causes A, or both B and A are caused by another factor. Without detracting from the Nobel prize winning investigation that first drew attention to the role of H pylori, we think that H pylori infection does not cause duodenal ulcer but prevents healing of an ulcer produced by hypersecretion of gastric acid. Acid diminishing treatment reduces the principal barrier against gastric H pylori infection and so the patient becomes infected. All references are in the version on bmj.com

WHERE DO YOU STAND ON THE ISSUE? Tell us on bmj.com 542

No link with prevalence If H pylori were the primary cause, we would not see regional variation in the prevalence

gastritis was proposed at the time of its discovery, and Koch’s postulates have since been fulfilled for this histological lesion.4 It has been more difficult to confirm the same link in duodenal ulcer as in gastric ulcer because it would be unethical to inoculate people with the bacterium to induce an ulcer. However, when Mongolian gerbils are injected with H pylori, they develop duodenitis, gastric metaplasia, and duodenal ulcer.5 In humans, pre-existing infection with H pylori is significantly associated with subsequent development of duodenal ulcer.6 This, together with the fact that most infection is acquired during childhood,1 suggests that presence of the bacterium predates and predisposes to ulcer formation, rather than that existing ulceration provides ideal conditions for H pylori to exploit. In patients who go on to develop duodenal ulcer it is proposed that H pylori infection induces antral gastritis, which leads to hypergastrinaemia, and hypersecretion of acid by the corpus. Excess acid washing over the duodenal bulb causes gastric metaplasia in the duodenum, allowing colonisation by H pylori, induction of an inflammatory response, and focal ulceration. A metaof duodenal ulcer within areas of high prevalence, particularly developing countries.1‑5 This does not refer to Holcombe’s “African enigma,”6 the alleged finding that duodenal ulcer was uncommon in Africa despite near ubiquity of the organism; that suggestion has been consigned to oblivion by Agha and Graham’s elegant systematic review showing that difficulties in reporting and in delivering medical care were responsible.7 However, in rural areas in developing countries, there are marked differences in prevalence of duodenal ulcer. These differences seem to be related to the staple diet and the local climate. For instance, in both India and China duodenal ulcer is commoner in the south than the north despite a lack of corresponding differences in H pylori infection.1‑4 This regional effect is visible in Agha and Graham’s review: the percentages of endoscopic duodenal ulcer in areas of low rainfall are about one half of those in other areas, supporting earlier findings.5 Moreover, H pylori infection has been present for many centuries, but duodenal ulcer emerged only around 1900.1 Can different virulent strains explain these discrepancies? Reports from 19 ­developing BMJ | 5 SEPTEMBER 2009 | Volume 339

head to head

analysis of randomised controlled trials has shown that eradication of H pylori infection in people with duodenal ulcer confers significant benefits, in terms of both facilitation of ulcer healing and prevention of relapse, compared with a course of acid suppression therapy alone.7 Successful eradication of the organism also leads to normalisation of acid production by the stomach.8 Chicken and egg Before the discovery of H pylori, duodenal ulcer was thought to arise as a result of excessive acid production. However, gastric acid secretion is proportional to lean body mass,9 which is increasing in the general population. If duodenal ulcer were predominantly acid related, its incidence should be increasing in the Western world, in line with that of other acid related disorders such as gastrooesophageal reflux disease.10 Instead, the reverse is the case, with studies showing a fall.10 This is mirrored closely by trends in H pylori infection, with prevalence falling by increasing year of birth over the past 20 to 30 years.1 11 The advent of treatment to eradicate H pylori infection has greatly reduced the frequency with

countries, where H pylori infection is almost ubiquitous (70-90%) and 77-88% of the strains carry the virulence factors cagA and vacA, show no relation between these factors and clinical outcome.8 In developed countries, duodenal ulcers occur in people without H pylori infection, even if we exclude factors such as non-steroidal anti-inflammatory drugs and Crohn’s disease. Ulcers are proportionately more common (up to 75% of all cases) in areas of low H pylori prevalence.1 Duodenal ulceration can also recur after eradication without re-infection.9 Again, half of patients with acute perforations of a duodenal ulcer (that is, with only a brief period of previous indigestion) are H pylori negative.10 Three papers report patients with a short history of duodenal ulcer being H pylori negative and that infection rates increase with length of history. These results are more consistent with duodenal ulcer causing infection with H pylori than with the reverse.11‑13 Acid secretion These arguments preclude H pylori infection as the primary cause, but there is no doubt that treatment of H pylori infection does lead

which duodenal ulcer is encountered at endoscopy.11 Taken together, these data provide further support for a causative role of H pylori in duodenal ulcer. Some researchers argue that H pylori is a bystander in duodenal ulcer, and cite a high prevalence of infection with the bacterium in the developing world, coupled with a low prevalence of duodenal ulcer, as evidence against causality. This has been termed the African enigma. A systematic review that examined this issue identified 40 studies reporting prevalence of endoscopic findings in 17 African nations, with duodenal ulcers occurring in 20% of individuals.12 It concluded that the frequency of complications of H pylori infection in Africa are similar to those seen in Western countries, with no dissociation between prevalence of H pylori infection and H pylori related diseases. Arguments for eradication Even if H pylori is not the causative agent in duodenal ulcer, eradication of the organism leads to ulcer healing and significantly reduces the likelihood of ulcer relapse, unless infection recurs (in contrast to acid suppression therapy where disease usually recurs after discontinuation). It

to quicker and more stable healing of duodenal ulcer. How can this be explained? Before Warren’s epic paper in 1984 it was generally accepted that duodenal ulcer was due to high acid secretion. Effective measures to reduce acid output, surgical or medical, led to long term healing, despite (presumably) persistence of H pylori infection; ineffective measures did not. We conclude that a high acid output is the primary cause, and that H pylori infection is secondary, delaying healing and leading to chronicity. It delays healing by impairing angiogenesis,14 15 diminishing the local blood supply to the ulcer area, and by interfering with the healing of damaged duodenal epithelial cells.16 These effects explain how eradicating the organism converts a chronic relapsing disease into one that can be cured. Our suggestions explain the fact that H pylori infection is more common in people with duodenal ulcer than those without ulcers. The ability of H pylori to colonise the stomach is pH dependent.17 18 At a low pH the patient is likely to be uninfected. Later, as a result of treatment suppressing acidity, H pylori infection occurs, starting initially in the antrum because of its higher pH. The antral infection results in hypergastri-

BMJ | 5 SEPTEMBER 2009 | Volume 339

is also cost effective, dominating all other management strategies in Markov modelling.7 H pylori has been linked definitively to the development of distal gastric cancer,13 and the World Health Organization classifies it as a human carcinogen. Searching for and eradicating the bacterium in high risk populations reduces incidence of gastric cancer.14 Critics argue against adopting this approach as gastric cancer is becoming rare in western countries. However, population screening and treatment programmes conducted in Europe have shown significant reductions in the prevalence of dyspepsia in the community. This strategy is therefore likely to be cost neutral to the health service in the long term.15 When we encounter a duodenal ulcer at upper gastrointestinal endoscopy, we call in our registrars because it is now a rare occurrence. The explanation that by far best fits all of the known facts is simple: H pylori causes peptic ulcer disease directly. We know of no compelling argument disproving this link. Competing interests: None declared. Cite this as: BMJ 2009;339:b2784

naemia, which causes an increased output of acid. Many think that this increase is sufficient to cause duodenal ulceration. However, duodenal ulcer occurs only in patients whose maximal acid output in response to continuous intravenous histamine stimulation exceeds a certain level; above that level ulcers increase with secretion rate, and above the upper 95% tolerance limit of the population they become inevitable.19 Accurate techniques20 show that H pylori infection reduces maximal histamine stimulated acid output in people with and without duodenal ulcer.21 A bacterium that depresses maximally stimulated gastric secretion is unlikely to cause a condition associated with hypersecretion of maximally stimulated gastric acid. Advocates of eliminating H pylori from a population to reduce gastric cancer incidence should be aware that they may be increasing the likelihood of duodenal ulcer by removing a brake on gastric acid secretion. Competing interests: KDB has received partial reimbursement from Nycomed, with whom he has worked for many years on oesophageal reflux, for attending two American Gastroenterological Association Meetings and an unconditional grant to fund a PhD programme in the history of medicine at Manchester University. Cite this as: BMJ 2009;339:b2788 543