Stress and peptic ulcer: life beyond helicobacterBMJ 1998; 316 doi: http://dx.doi.org/10.1136/bmj.316.7130.538 (Published 14 February 1998) Cite this as: BMJ 1998;316:538
- Susan Levenstein, adjunct research physiciana ()
- Correspondence to: Dr Levenstein Via del Tempio 1A, 00186 Rome, Italy
- Accepted 21 July 1997
The discovery that Helicobacter pylori is a cause of peptic ulcer has tempted many to conclude that psychological factors are unimportant. But this is dichotomised thinking. There is solid evidence that psychological stress triggers many ulcers and impairs response to treatment, while helicobacter is inadequate as a monocausal explanation as most infected people do not develop ulcers. Psychological stress probably functions most often as a cofactor with H pylori. It may act by stimulating the production of gastric acid or by promoting behaviour that causes a risk to health. Unravelling the aetiology of peptic ulcer will make an important contribution to the biopsychosocial model of disease.
For this review of the role of psychological stress in the aetiology of peptic ulcer disease, I undertook conventional journal tracking and reference tracing, supplemented by Medline searches using Paperchase. The important keywords used in this search included peptic ulcer; duodenal ulcer—psychology; stress; life change events; and personality.
Summary points Psychosomatic factors in the aetiology of peptic ulcer have become unfashionable since the discovery of Helicobacter pylori
Most people harbour H pylori so the organism cannot serve as the sole explanation for ulcer disease
Psychological stress has an impact on the onset and course of ulcer disease
Psychological stress probably interacts with H pylori and other risk factors in causing ulcer disease
Peptic ulcer is an important example of the biopsychosocial model of disease
H pylori is not enough
When H pylori burst on the scene a few years ago, it revolutionised views on the aetiology and treatment of peptic ulcer. Psychosocial factors were quietly but firmly escorted off the stage, and gastroenterologists in particular banished psychological considerations with something approaching relief.
While this surge of biological reductionism is understandable, it risks throwing the baby out with the bath water. H pylori is inadequate as a sole explanation for peptic ulcers. Most people who harbour the organism never have ulcers, while a few who have never been infected with it or taken non-steroidal anti-inflammatory drugs develop ulcer disease.1 This testifies to the role of factors additional to infection in peptic ulceration (the situation is similar for Mycobacterium tuberculosis and, indeed, most infectious organisms). H pylori alone does not explain fully the epidemiological patterns of upper gastrointestinal disease.2 In addition, Marshall (who first identified H pylori on the surface of the gastric antrum) found that Koch's postulates for establishing the aetiological relation between a micro-organism and disease are not met in the case of H pylori and ulcers.3 Again, the same might be said for many other established causes of disease.
Explaining the aetiology of a disease is often like assembling a jigsaw puzzle. By the time H pylori had been discovered, many pieces of the ulcer puzzle, from cigarette smoking to type O blood, had already been found and fitted together, although we did not know exactly where they belonged in the larger picture. The subsequent discovery of an important and central piece may mean that other completed sections have to be moved around—not that they have to be discarded.
The discovery of H pylori, far from eliminating interest in that older assortment of physical risk factors for ulcer disease, has spurred some researchers to re-establish and reinterpret the importance of older factors in the new context. Clinical studies have examined these physical risk factors in relation to their association with or independence from H pylori in ulcer patients.1 4 Other workers are combing large databases to see how these same factors relate to H pylori in the general population.5 6
Why are psychological factors being ignored?
Why has there been no corresponding flurry of articles on the relation between H pylori and psychology?7 For one thing, researchers interested in psychosocial factors tend to have little interest or expertise in microbiology, and vice versa. But I suspect the reason is deeper. The current generation of gastroenterologists has lost patience with psychosomatic explanations, at least where these apply to peptic ulcer. The methodological inadequacies of some of the older published reports on psychosocial factors in ulcer disease are partly to blame. These studies were often flawed by contamination from disease effects, diagnostic uncertainty, and recall bias; did not pay enough attention to adjustment for confounding factors; and sometimes left readers distracted by side issues such as the nature or the measurement of stress.
Resistance to an integrated approach
The main feeling, however, seems to be that psychosomatic reasoning can be discarded as soon as another explanation becomes available. This dichotomised thinking, focused on the possibility of moving peptic ulcer from a stigmatised “psychosomatic” cubbyhole into a more dignified “infectious” one, reflects an ingrained resistance to the difficult but essential task of examining disease aetiology in an integrated manner that incorporates both psychological and biomedical elements8—a task begun for peptic ulcer long ago.9
Negative case-control studies
In fact, while all references to psychology have been removed from the chapter on ulcers of the 20th (1996) edition of Cecil and Loeb,10 this is not solely due to H pylori. In the 17th edition (1982), published before the discovery of helicobacter, 18 lines were allotted to psychology, but all were dismissive. Around that time a spate of negative case-control studies was sabotaging interest in psychosomatic factors in ulcer disease.11 One reason for the difficulty these studies had in showing a causal role for psychosocial factors is that patients with peptic ulcer are a heterogeneous group. Stress is probably an active factor in only some patients—perhaps those with less exposure to ulcerogenic substances12 or those with higher pepsinogen concentrations.13
Stress is fashionable elsewhere
Ironically, while the gastroenterological community seems to view those who have continued to support a psychosomatic aetiology for ulcer disease as stubbornly clinging to obsolete views, psychosocial factors have the glitter of novelty for researchers in other specialties, who are happily exploring whether you can die of fright or get the sniffles from stress.
Stress and ulcers: the evidence
Since the 1980s there has been a modest resurgence of research interest in the ulcer-stress question, and the resulting body of evidence has generally, though not always, been supportive.14 A solid series of methodologically sound studies now supports an aetiological effect of aspects of “stress” (a convenient term to cover both life stressors and subjective distress) ranging from depression to war. Considerable prospective evidence has been gathered alongside studies that have found an excess of life stressors in ulcer patients compared with matched controls.15 In one large longitudinal population study, the occurrence of self reported ulcer over a nine year period was more likely in subjects who reported any of several concrete life stressors or psychological distress at baseline.16 17 In another study, self reported stress predicted the occurrence of ulcer disease (diagnosed by a doctor) over the next 13 years.18 Family and job difficulties increased the risk of ulcer over five years in one sample of Israeli men.19 Major societal disasters, including German air raids in London and economic collapse in Sofia, have been associated with documented increases in acute ulcers.20 21 The ulcerogenic effects of stress have been shown to be robust enough to survive adjustment for behavioural and physical confounding factors.16 17 18
Stress and prognosis in ulcer disease
Prospective clinical studies have reported that psychological distress impedes ulcer healing (as seen with endoscopy) after H2 agonist treatment—even when stringent recruitment criteria exclude confounding in relation to disease chronicity.22 23 Life stress continued to worsen the prognosis over several years in two prospective case series.23 24 This effect seems to be reversible, however: a psychologically stable person who develops an ulcer during a stressful period is likely to remain free of symptoms for years after a short course of treatment, even without medication to eradicate H pylori.24
Psychological stress and gastric acid secretion
Psychological stress is not only empirically associated with ulcers, but is a very plausible risk factor for ulcer disease. Gastric acid output is correlated with psychological distress in patients with and without ulcers,25 and increased enormously during intense military training.26 Compared with healthy people, patients with duodenal ulcers are particularly likely to respond to laboratory stressors by secreting more acid.27 In two patients with duodenal ulcers, extraordinary life circumstances resulted in a 10-fold to 20-fold increase in the basal acid output.28
Under stress, the amount of acid reaching the duodenum may increase further because gastric motility has changed or meals have been missed. People affected by stress may also smoke more, sleep less, and take more non-steroidal anti-inflammatory drugs, thereby increasing their susceptibility to ulcer by mechanisms that are not related to acidity. Evidence that prospective epidemiological associations between life stressors or psychological characteristics and ulcer are reduced by adjusting statistically for patterns of eating, sleeping, and substance use support a mediatory role for these health risk behaviours in the psychogenesis of ulcers.16 17
An association between stress and H pylori infection?
Despite empirical support and biological plausibility, there could in theory still be a reason for abandoning psychogenic causes of ulcer disease. Stress might be associated strongly with H pylori infection, inducing the illusion that psychosocial factors can influence ulcer formation. For many of the other classic risk factors, this possibility has been examined directly and discarded. Smoking, alcohol consumption, non-steroidal anti-inflammatory drugs, and type O blood are not correlated with H pylori in populations without ulcers.5 6 The story is somewhat different for socioeconomic status. An excess of H pylori infection in poor people5 suggests that low social class may partly owe its reputation as an ulcer risk factor to confounding by H pylori.29
No direct data are available as yet for stress, but there is no reason to expect that infection with H pylori is related to any particular psychological state. It is true that life stress, psychological distress, and H pylori may all be associated with low socioeconomic status, but epidemiological evidence argues against socioeconomic status being a principal cause of the association between stress and ulcers.17
How do H pylori and psychosocial stress interact in ulcer disease?
If we conclude that the discovery of H pylori has not erased the role of psychosocial risk factors in ulcer disease, asking how the two interact becomes logical. In some cases there may be no interaction at all. Given the large increase in acid production under severely stressful conditions26 28 and the high secretion of acid30 and pepsinogen1 in patients whose duodenal ulcers are not related to either H pylori or non-steroidal anti-inflammatory drugs, stress may be capable of causing peptic ulceration even in the absence of H pylori.12 In most cases, however, stress probably functions as a cofactor with H pylori. Several possible mechanisms can be postulated (box).
Stress and H pylori as cofactors in ulcer disease
Stress could facilitate the evolution of H pylori infection into ulcer by producing gastric hyperchlorhydria31
Stress could disturb the equilibrium between H pylori and its host via psychoneuroimmunological mechanisms
Stress could reduce mucosal defences to H pylori invasion through behavioural mediators such as cigarette smoking4
Stress could increase the chances of ulceration in duodenal mucosa that have already been weakened by the effects of H pylori infection simply by increasing the acid load which flows past
Stress induced acid secretion could promote H pylori colonisation of the duodenal bulb by neutralising the inhibitory effect of bile32
Again, little empirical evidence exists on the relation between stress and H pylori, and that is limited to studies of psychological distress in gastroenterology patients. In subjects with undiagnosed dyspepsia, those who are depressed or anxious are less commonly infected with H pylori.33 Similarly, in patients with documented peptic ulcer, the higher the absolute titre of H pylori IgG antibody (a rough measure of the intensity of bacterial colonisation of the antrum), the less anxious is the patient likely to be.12 These findings suggest that H pylori infection and psychological stress promote ulcer pathogenesis via pathophysiological pathways that are largely additive and therefore independent and complementary rather than synergistic.34 Two different mechanisms of facilitating acid damage to the duodenum would result in a statistically inverse relation in ulcer patients, though not necessarily in the general population
Inverse relation between psychological stress and H pylori
In the classic view of the pathogenesis of peptic ulcer disease—that it is the sum of factors that increase acid secretion and those that lower mucosal defences—psychological stress probably acts mostly on the side of increased aggression and H pylori on the side of weakened defences.35 36 H pylori infection does not necessarily raise, and may even lower, acid secretion.30 34 Thus, though it has been suggested that the organism contributes to increased acid in the duodenum37 and to raised pepsinogen values,38 the striking increases in gastric acid found in many patients with duodenal ulcer seem to result from mechanisms that are at least partly independent of H pylori infection.1 30
Other possible explanations exist for an inverse relation between anxiety and H pylori antibody titres in ulcer patients. Stress may, for example, suppress serum antibody titres by stimulating cortisol production. Local processes in the gastric mucosa could also contribute, since a low pH tends to cause antigen and antibody to dissociate. Gastric hyperacidity resulting from stress could therefore suppress mucosal immunity, including H pylori antibody formation.
The way ahead
But all this is speculation. Now that the helicobacter earthquake has passed, those of us interested in the effects of psychosocial factors on peptic ulcer must begin investigating again in the new context. This will require scrupulous use of a variety of research tools if we are to have any impact on the many colleagues who think our efforts are a waste of time. We need to be careful to use appropriate measures in examining psychosocial data and to embrace the complexities of the biopsychosocial model. We need to pay great attention to avoiding recall bias, contamination of epidemiologically defined “ulcer” groups by subjects with non-ulcer dyspepsia, and statistical confounding by socioeconomic status.
Far from being obsolete, the concept that psychosocial factors play a role in peptic ulcer presents exciting and varied research opportunities in the age of H pylori. We can study the clustering between stress indicators and H pylori in the general population, we can look at the relation between H pylori and stress in case-control and prospective studies of ulcer patients, we can bring people whose H pylori status is known into the physiology laboratory and look at their physical reactions to stress, we can extract data on H pylori antibody titres from serum samples frozen for ongoing longitudinal studies, and we can study interactions of H pylori and stress in animal models. These kinds of investigations can serve not only to breathe new life into psychosomatic concepts of peptic ulcer, but also to develop a more general paradigm for applying the integrated biopsychosocial model to medical disorders both infectious and otherwise.
I am indebted to Douglas Drossman, Harold Bourne, and Cosimo Prantera for their encouragement and for many invaluable exchanges of ideas.
Conflict of interest: None.