The dark psychosomatic history of peptic ulcer part III

Peptic ulcer was long seen as one of the prime examples of psychosomatic disease. From the 1930s to the 1980s, repressed emotions and stress were considered its main cause. “That psychic factors play a prominent role in the causation of ulcer is doubted by no one”, a 1952 JAMA review proclaimed.

The most popular theory was advanced by Franz Alexander, the father of psychosomatic medicine. He argued that ulcer patients had repressed desires to be loved and nurtured like a child but that they could not give in to these wishes because they wanted to appear strong and independent. The body, however, usually finds a way of expressing repressed emotions. In the worldview of a baby, the wish to be loved and cared for is closely associated with feeding. Therefore, Alexander advanced that unresolved dependency needs are expressed by the stomach. As peptic ulcer patients craved affection, their stomach was constantly preparing for food that would never come. This resulted in an overproduction of acid that over time resulted in the formation of an ulcer.

The common treatment of hospitalization, bed rest, and frequent feeding with milk and cream was seen as a confirmation of Alexander’s theory that peptic ulcer patients had a repressed desire to be nurtured. Multiple scientific articles proposed psychotherapy as “the only etiologic treatment in most cases of peptic ulcer.”

Then, in the 1980s a paradigm shift occurred. Two Australian doctors discovered that peptic ulcer was caused by a germ, earning them the 2005 Nobel Prize in medicine. Ulcers, it turned out, could easily be treated by a combination of antibiotics and acid inhibitors.

In this three-part series, we will explore the dark psychosomatic history of peptic ulcer.  In part I, we looked at how psychosomatic explanations of peptic ulcers became popular in the first place. In part II we explored the main psychosomatic theories in depth. This episode will look at the discovery of Helicobacter pylori and current evidence linking stress and peptic ulcers.

Critical voices

In the previous episode, we looked at some of the psychosomatic theories that were found in the scientific literature from the 1930s to the 1980s. In that period, stretching half a century, relatively few researchers spoke out against this paradigm. Many scientists went along with it and suppressed findings that contradicted it.

The epidemiologist Richard Doll, who famously helped establish a causal relationship between smoking and cancer, provided insight on how things were at the time. In 1951 Doll conducted a large epidemiological study on peptic ulcer which found a high incidence in foremen in factories compared to other workers. This fitted into the stress hypothesis quite nicely but other findings that contradicted it were never published. Reflecting on the issue many decades later, Doll recounted

“I suppose at one time I did contribute to the idea that stress might play a part, […]  The only research I tried to do to test whether stress had any effect was something that I regret I never published. We drew up a list of all the things that we thought might be most worrying to people, like loss of job, members of their family dying, divorce, a whole list of some dozen objectively determinable factors that might cause stress. […]  We then looked at how often these events had occurred prior to a patient’s admission to hospital for haematemesis, which was frequently described as due to stress on top of an ulcer, and how often they had occurred before the admission of patients in a control group. We found that the proportion was exactly the same in both groups.

One of the few colleagues who openly questioned the psychosomatic view on peptic ulcer was another eminent British physician named Richard Asher. Asher was witty and liked to challenge accepted wisdom in provocative essays. His paper “The Dangers of Going to Bed”, for example, criticized the idea that bed rest could be prescribed as the basis of treatment for almost any medical condition.

But Asher also criticized the lack of empirical support for psychosomatic theories. He once gave Doll a paper linking duodenal ulcer and psychological stress, asking if it was an appropriate reflection of what researchers think. When Doll said yes, Asher revealed that the paper was actually written in 1850 about general paralysis of the insane. He had merely substituted the name peptic ulcer with ‘general paralysis of the insane’.

Other critical voices are hard to find* but there was an anonymous letter in the BMJ from 1959 which stated presciently:

“When future work has solved the riddle of chronic peptic ulcer, we may find that the facile explanation sometimes given to-day that it is a disease of civilization due to mental stress will seem as remote from the truth as the view that malaria is caused by foul vapours ascending from the swamps.”

A paradigm shift

Barry Marshall, one of the Australian discoverers of H. pylori as the main cause of peptic ulcers, was also skeptical of stress theories when he studied the illness in the early 1980s. He was appalled by how patients were treated if their symptoms could not easily be explained.

“I was never satisfied with saying that by ruling out all these diseases, a person must have a fake disease”, he said in a 2010 interview with Discover Magazine. “So I accepted the fact that lots of times I couldn’t reach a fundamental diagnosis, and I kept an open mind.”

Marshall was still in training when he started working with Robin Warren, the pathologist who was finding mysterious bacteria in the gut of people with stomach problems. Marshall was assigned to investigate this further and decided to look deeper into the medical history of some of the cases where these bacteria were found. He remembered one patient, a woman in her forties, who suffered from stomach pain and nausea, but her tests all came back normal, so she was sent to a psychiatrist and put on antidepressants. He thought this was rather odd.

In his view, stress is often used to mask our lack of understanding. As he explained in an interview with Slate:

“You can always find stress in someone’s life if you want to. You ask a few questions and eventually it’s,’Yes, I admit, I was worried about something recently.’ So they tried to find evidence for stress causing ulcers, and whenever they had an experiment which worked, it would just be blown out of all proportion, and everyone would get so much publicity out of it that you would think,’Ah, at last, it’s proven.’ But the data was very bad. And in fact there was plenty of evidence showing that stress didn’t make much difference.

The finding that peptic ulcer is mostly caused by H. pylori instead of stress has wide implications. Although the term paradigm shift is frequently overused, this seems one instance where it is appropriate. Marshall went on:

“What else is supposedly caused by stress that we can debunk? A lot of these things that are supposedly caused by stress, you try to track down the reason for that link, and there isn’t one, except the fact that we don’t have any better cause. Everything that’s supposedly caused by stress, I tell people there’s a Nobel Prize there if you find out the real cause.”

Birth cohort effect

Another interesting question is whether psychosomatic theories stalled progress in our scientific understanding of peptic ulcer. Although we acknowledge that this is nearly impossible to answer (we can’t wind back time), there are two interesting facts that are worth mentioning.

The first is that multiple researchers had already found and reported spiral bacteria in the gut and speculated that these might be related to gastritis and peptic ulcer, long before Marshall & Warren did in the 1980s. Their discovery was not due to some new technology that was unavailable in the past. Previous accounts were simply ignored or not followed up on. Researchers following the stress-hypothesis saw those bacteria either as an accidental finding or an opportunistic infection that followed rather than caused the ulcer.

A second interesting finding is the birth cohort effect, first reported by Mervyn Susser and his wife Zena Stein in 1962. Based on epidemiological data, they found that the risk of peptic ulcer and its socio-economic gradient differed strongly for different birth cohorts. The age you were born in affected the risk of peptic ulcer more than what happened or what you did afterward. This points to an environmental factor in childhood that determines who would develop peptic ulcer later in life. The data fits well with an infectious theory and changes in hygiene but at the time psychosomatic theories were still dominant. Therefore, the birth cohort effect was attributed to the stress caused by the World Wars and the economic depression in the 1930s.

Susan Levenstein

One might expect that psychosomatic theories remained popular until suddenly H. pylori was discovered as its main cause. The reality is a bit more complex.

Psychoanalysis went out of fashion in the 1970s, so the theories of Alexander and Szasz fell out of favor. Stress was still a popular explanation but in the 1980s a number of studies appeared that reported null results (examples here, here, here, here, and here). The psychosomatic approach was already losing traction at the time when Marshall and Warren made their great discovery. And curiously enough it remained quite popular even when bacteria became accepted as the main cause of peptic ulcer.  

The main reason is that there are a lot of asymptomatic infections: almost half of humankind has H. pylori in their stomach but far fewer develop peptic ulcer. Stress was proposed as the missing link that would explain why some go on to develop an ulcer after an infection with H. pylori while others don’t.

The most vocal proponent of this view is internist Susan Levenstein. She published several opinion pieces in influential medical journals such as JAMA and the BMJ, claiming that stress is still important in the development of peptic ulcer. Levenstein uses remarkably strong language to make her point. She claims for example that “dichotomised thinking” and a “surge of biological reductionism” risk “throwing the baby out with the bath water”. In another paper she writes that despite manifestations of bacteriologic missionary zeal, the evidence linking psychosocial stress to peptic ulcer is too strong to be ignored. And those negative results in the 1980s, finding no link between stress and peptic ulcer? According to Levenstein these were sabotaging interest in psychosomatic factors in ulcer disease.”

Levenstein stressed (no pun intended) that there is “solid evidence that psychological stress triggers many ulcers” and she estimated that psychosocial factors contribute to 30% to 65% of all ulcers. She hopes that scientific research will “breathe new life into psychosomatic concepts of peptic ulcer” because far from being obsolete she thinks that “the concept that psychosocial factors play a role in peptic ulcer presents exciting and varied research opportunities in the age of H. pylori.”

Current evidence

Let’s take a closer look at current evidence for a psychosomatic understanding of peptic ulcers. Levenstein conducted a large prospective study in Alameda, California. More than 4000 people answered various questionnaires with information on psychosocial variables and were then followed up for a couple of years. In that period, some had developed peptic ulcer, allowing researchers to test if these participants differed from those who didn’t develop an ulcer. Levenstein reported that depression, maladjustment, and hostility were related to peptic ulcer.

There are some major limitations to the study such as the fact that the diagnosis of peptic ulcer was self-reported and not confirmed by the researchers and that H. pylori wasn’t measured and taken into the equation. The main limitation of the Alameda study, however, is that it is not a scientific experiment but an observational study making it nearly impossible to establish causal relationships unless the effect is reasonably large.

This is an issue that regularly returns in psychosomatic research. Because a lot of things are correlated with each other, it is often hard to differentiate cause, effect and a spurious correlation. Suppose, for example, that people who earn less, smoke more. Anything correlated with lower socio-economic status will then also likely correlate with smoking and seem to be associated with lung cancer. This is called confounding and it is hard to avoid in observational studies where researchers cannot control for all risk factors.

The relative risk of lung cancer in cigarette smokers is approximately 20. That is a very large effect and unlikely to be a false signal. Effects sizes lower than an odds ratio or relative risk of 3, however, are tricky because it is difficult to know if such an effect is real or the result of confounding. As Robert Temple, a director at the Food and Drug Administration, once put it: “My basic rule is if the relative risk isn’t at least three or four, forget it.” Observational studies are a bit like a microscope that cannot provide enough resolution to see small effects clearly. 

The effect Levenstein found in her Alameda study was really small: an odds ratio of 1.7 with a confidence interval of 1.0-3.1. Other longitudinal studies have found similar effects with odds ratios usually below 2. Rather than providing “solid evidence”, these studies cannot show that there is a true effect of stress on peptic ulcer. They only suggest that if there is an effect, it is likely a small one.

Levenstein also points to an increase in peptic ulcer after dramatic historical events, such as the London air raids during the Second World War (discussed in part I of our series) or the Hanshin-Awaji Earthquake in Japan in 1995. We had a closer look at the latter study.

The number of peptic ulcers hardly increased after the Japanese earthquake because the number of endoscopies in the most afflicted area dropped from 6000 to only 3000.  The authors highlight that number of ulcers per endoscopy increased but this is difficult to interpret because it is influenced by many societal factors (such as the tendency of patients to go see a doctor). In other words, the data is too messy to conclude much.

Levenstein also refers to studies on Prisoners of War (POWs) where an increase of peptic ulcers has been noted by several reports (examples here and here). POWs, however, often have developed chronic pain conditions because of the torture and abuse they endured during captivity. Non-steroidal anti-inflammatory drugs (NSAIDs) such as aspirin or ibuprofen are common painkillers and the second most prominent cause of peptic ulcer, next to H. pylori. So, it is possible that the higher rate of peptic ulcer is due to a higher intake of NSAIDs among POWs.

In conclusion, the evidence linking stress and peptic ulcers remains unconvincing and problematic. Although not refuted, current data suggest that if there is an effect it is likely to be small and less important than risk factors such as H. pylori and NSAIDs. Almost a century of psychosomatic research on peptic ulcers has provided close to no progress.

Click here for an overview of all our articles on the dark history of psychosomatic medicine.


* An excellent modern critique of psychosomatic approach to peptic ulcer is found in the essay “The biopsychosocial approach: a note of caution” by epidemiologist George Davey Smith. It was published as a chapter in the book “Biopsychosocial Medicine: An integrated approach to understanding illness,”

Leave a Reply