modiﬁed from “an end to ulcers” by k. hannam and r...

A Case Study on Peptic Ulcers

Modified from “An End to Ulcers” by K. Hannam and R. Hagley. National Center for Case Teaching in Science.

Tuesday, January 24, 2012

What causes peptic ulcers?


Gastric Diseases


Design an Experiment

In small groups design an experiment to test the “excess acid” hypothesis.Be specific about how you would treat your experimental groups and what you would measure.


Which option is the best test of the excess acid hypothesis?


What step in the scientific method does this test represent?


A Possible Study Design

Patients suffering from ulcers were divided into two groups

Group 1 received antacids and were instructed to take them 3X/day

Group 2 received sugar pills (a placebo) and were instructed to take them 3X/day

After 3 months, the number of patients with ulcers was assessed.

Neither the doctor evaluating the patients nor the pathologist assessing the patient biopsy samples knew which patients had been given the antacids and which had been given the sugar pills. The study was “double-blind”.


How would you label the two groups?


Why is it important that the trial be “double-blind”?


PredictionPredict what the results of the trial would look like after 3 months if the excess acid hypothesis was correct.

No. o

f pat

ient

s wi

th u

lcer

s

ExperimentalGroup

ControlGroup


Which graph most closely matches your predicted results?

Ex C Ex C

Ex C Ex C

A) B)

C) D)


Anecdotal Evidence for the Excess Acid Hypothesis

Unfortunately, controlled, double-blind studies were assessing the number of patients with ulcers were not conducted

There was abundant anecdotal evidence and even controlled studies in which antacids relieved the SYMPTOMS of gastric ulcers.

However, the antacids did not cure the ulcers. The relapse rate was high.

Drug companies spend millions on developing and testing new antacids and chemicals that block stomach acid production.


A Technological Breakthrough

Fiberoptic endoscopy allowed gastroenterologists to take biopsies.


Dr. J. R. Warren’s Bacteria


Silver Stained Bacteria


Marshall cultivates a bacterium from gastric biopsies.

Collaboration with Dr. Barry Marshal


1273

UNIDENTIFIED CURVED BACILLI ON GASTRICEPITHELIUM IN ACTIVE CHRONIC GASTRITIS

SIR,-Gastric microbiology has been sadly neglected. Half thepatients coming to gastroscopy and biopsy show bacterialcolonisation of their stomachs, a colonisation remarkable for theconstancy of both the bacteria involved and the associatedhistological changes. During the past three years I have observedsmall curved and S-shaped bacilli in 135 gastric biopsy specimens.The bacteria were closely associated with the surface epithelium,both within and between the gastric pits. Distribution wascontinuous, patchy,’ or focal. They were difficult to see withhaematoxylin and eosin stain, but stained well by the Warthin-Starry silver method (figure).

I have classified gastric biopsy findings according to the type ofinflammation, regardless of other features, as "no inflammation","chronic gastritis" (CG), or "active chronic gastritis" (ACG). CGshows more small round cells than normal while ACG ischaracterised by an increase in polymorphonuclear neutrophilleucocytes, besides the features of CG. It was unusual to find noinflammation. CG usually showed superficial oedema of themucosa. The leucocytes in ACG were usually focal and superficial,in and near the surface epithelium. In many cases they onlyinfiltrated the necks of occasional gastric glands. The superficialepithelium was often irregular, with reduced mucinogenesis and acobblestone surface.When there was no inflammation bacteria were rare. Bacteria

were often found in CG, but were rarely numerous. The curvedbacilli were almost always present in ACG, often in large numbersand often growing between the cells of the surface epithelium(figure). The constant morphology of these bacteria and theirintimate relationship with the mucosal architecture contrasted withthe heterogeneous bacteria often seen in the surface debris. Therewas normally a layer of mucous secretion on the surface of themucosa. When this layer was intact, the debris was spread over it,while the curved bacilli were on the epithelium beneath, closelyspread over the surface (figure).The curved bacilli and the associated histological changes may be

present in any part of the stomach, but they were seen mostconsistently in the gastric antrum. Inflammation, with no bacteria,occurred in mucosa near focal lesions such as carcinoma or pepticulcer. In such cases, the leucocytes were spread through the fullthickness of the nearby mucosa, in contrast- to the superficialinfiltration associated with the bacteria. Both the bacteria and thetypical histological changes were commonly found in mucosaunaffected by the focal lesion.The extraordinary features of these bacteria are that they are

almost unknown to clinicians and pathologists alike, that they areclosely associated with granulocyte infiltration, and that they arepresent in about half of our routine gastric biopsy specimens innumbers large enough to see on routine histology. The only otherorganism I have found actively growing in the stomach is Candida,sometimes seen in the floor of peptic ulcers. These bacteria were notmentioned in two major studies of gastrointestinal microbiology 1,2possibly because of their unusual atmospheric requirements andslow growth in culture (described by Dr B. Marshall in the

accompanying letter). They were mentioned in passing by Fung etal.3How the bacteria survive is uncertain. There is a pH gradient

from acid in the gastric lumen to near neutral in the mucosal vessels.The bacteria grow in close contact with the epithelium, presumablynear the neutral end of this gradient, and are protected by theoverlying mucus.The identification and clinical significance of this bacterium

remain uncertain. By light microscopy it resembles Campylobacterjejuni but cannot be classified by reference to Bergey’s Manual of

1. Gray JDA, Shiner M. Influence of gastric pH on gastric and jejunal flora Gut 1967; 8:574-81.

2. Drasar BS, Shiner M, McLeod GM. Studies on the intestinal flora II: The bacterial floraof the gastrointestinal tract in healthy and achlorhydric persons. Gastroenterology1969; 56: 71-79.

3. Fung WP, Papadimitriou JM, Matz LR. Endoscopic, histological and ultrastructuralcorrelations in chronic gastritis. Am J Gastroenterol 1979; 71: 269-79.

Curved bacilli on gastric epithelium.Section is cut at acute angle to show bacteria on surface, forming network

between epithelial cells. (Warthm-Starry silver stain; bar= 10 m.)

Determinative Bacteriology. The stomach must not be viewed as asterile organ with no permanent flora. Bacteria in numberssufficient to see by light microscopy are closely associated with anactive form of gastritis, a cause of considerable morbidity (dyspepticdisease). These organisms should be recognised and theirsignificance investigated.Department of Pathology,Royal Perth Hospital,Perth, Western Australia 6001 J. ROBIN WARREN

SIR,-The above description of S-shaped spiral bacteria in thegastric antrum, by my colleague Dr J. R. Warren, raises thefollowing questions: why have they not been seen before; are theypathogens or merely commensals in a damaged mucosa, and are theycampylobacters?

‘

In 1938 Doenges’ found "spirochaetes" in 43% of 242 stomachsat necropsy but drew no conclusions because autolysis had renderedmost of the specimens unsuitable for pathological diagnosis.Freedburg and Barron2 studied 35 partial gastrectomy specimensand found "spirochaetes" in 37%, after a long search. Theyconcluded that the bacteria colonised the tissue near benign ormalignant ulcers as non-pathogenic opportunists. When Palmer3 3examined 1140 gastric suction biopsy specimens he did not usesilver stains, so, not surprisingly, he found "no structure whichcould reasonably be considered to be of a spirochaetal nature". Heconcluded that the gastric "spirochaetes" were oral contaminantswhich multiplied only in post mortem specimens or close to ulcers.Since that time, the spiral bacteria have rarely been mentioned,except as curiosities,4 and the subject was not reopened with the

1. Doenges JL Spirochaetes in the gastric glands of Macacus rhesus and humans withoutdefinite history of related disease. Proc Soc Exp Med Biol 1938, 38: 536-38

2. Freedburg AS, Barron LE. The presence of spirochaetes in human gastric mucosa. AmJ Dig Dis 1940, 7: 443-45.

3 Palmer ED Investigation of the gastric spirochaetes of the human? Gastroenterology1954; 27: 218-20

4 Ito S. Anatomic structure of the gastric mucosa. In: Heidel US, Cody CF, eds.Handbook of physiology, section 6. Alimentary canal, vol II· Secretion

Washington, DC: American Physiological Society, 1967 705-41.


Why were most doctors skeptical of Warren and Marshall’s hypothesis that bacteria caused ulcers?


Do bacteria cause ulcers?

Drs. Warren and Marshall hypothesize that bacteria may cause gastric ulcers.What are some alternative hypotheses to explain the presence of bacteria in the stomach.


A Pilot StudyNATIONAL CENTER FOR CASE STUDY TEACHING IN SCIENCE

Page !“Helicobacter pylori and the Bacterial " eory of Ulcers” by Debra Ann Meuler

Based on their data and the statistical analysis, Warren and Marshall concluded the following:a. " ere were no well-de# ned clinical symptoms associated with the presence of the bacteria except for burping

and abdominal pain.b. " ere was no evidence that other sources of infections such as animal contact, diet, and poor dental hygiene

were the source of the infection.c. Gastritis, which was histologically diagnosed, was closely associated with the presence of the bacteria.d. Bacteria were present almost exclusively in patients with chronic gastritis and were also common in those with

peptic ulceration of the stomach or duodenum.

In fact, 26 out of 30 patients with ulcers had the bacteria. " ose four patients with ulcers but negative for the bacteria had taken non-steroidal, anti-in$ ammatory drugs known to induce ulcers. Warren and Marshall now felt they had su% cient evidence linking Helicobacter pylori with gastritis and peptic ulcer disease. " e next step was to present their # ndings.

Presenting Their FindingsMarshall and Warren presented their # ndings to the College of Physicians at RPH. " eir work was not well received. Most were skeptical of their claims. Of the many objections, the main one was the link between gastritis and duodenal ulcers. Prevailing wisdom said that gastritis was associated with gastric ulcers, not duodenal ulcers. Marshall did further research and found a study done at the Mayo Clinic in the 1950s. In this study, stomachs from either motor vehicle fatalities or partial gastrectomy patients were examined for the presence of gastritis and ulcers. Researchers found a surprisingly high number of specimens that showed evidence of gastritis. Of those with duodenal ulcers, 100% (N = 250) also had gastritis. Here was evidence that supported Warren and Marshall’s hypothesis. Why had the ulcer research community skipped over the relationship between gastritis and duodenal ulcer? Marshall speculated it was because it didn’t # t with current beliefs and was therefore ignored.

Marshall continued reviewing the literature for any information pertaining to their research. He read that while acid-reducing drugs relieved the symptoms of the ulcers, they did not cure them. " e relapse rate was high. Cimetidine, a common acid-reducing agent, helped to heal the ulceration, but did not cure the disease. Marshall was further intrigued when he read that colloidal bismuth subcitrate (CBS; similar to the over-the-counter drug PeptoBismol) greatly reduced the relapse rate in those with gastric ulcers. Marshall hypothesized that if a drug lowered the relapse rate for gastric ulcers and if bacteria can cause gastric ulcers, then CBS should have anti-microbial properties. To test

Table II: Association of Bacteria with Endoscopic Diagnosis

Endoscopic Appearance* Total With Bacteria p

Gastric ulcer 22 18 (77%) 0.0086

Duodenal ulcer 13 13 (100%) 0.00044

All ulcers 31 27 (87%) 0.00005

Oesophagus abnormal 34 14 (41%) 0.996

GastritisT 42 23 (55%) 0.78

DuodenitisT 17 9 (53%) 0.77

Bile in stomach 12 7 (58%) 0.62

Normal 16 8 (50%) 0.84

Total 100 58 (58%)

*More than one description applies to several patients.T Refers to endoscopic appearance, not histological in! ammation.

Table III: Histological Grading of Gastritis and Bacteria

Bacterial Grade

Gastritis Nil 1+ 2+ 3+ total

Normal* 29 2 0 0 31

ChronicT 12 9 7 1 29

Active 2 5 15 18 40

Total 43 16 22 19 100

*Gastritis grades 0 and 1 are normal.T1 case showed bacteria on gram stained smear.

Fig. 3: Data Tables Published in The Lancet by Warren and Marshall in 1984

NATIONAL CENTER FOR CASE STUDY TEACHING IN SCIENCE

Page !“Helicobacter pylori and the Bacterial " eory of Ulcers” by Debra Ann Meuler

Based on their data and the statistical analysis, Warren and Marshall concluded the following:a. " ere were no well-de# ned clinical symptoms associated with the presence of the bacteria except for burping

and abdominal pain.b. " ere was no evidence that other sources of infections such as animal contact, diet, and poor dental hygiene

were the source of the infection.c. Gastritis, which was histologically diagnosed, was closely associated with the presence of the bacteria.d. Bacteria were present almost exclusively in patients with chronic gastritis and were also common in those with

peptic ulceration of the stomach or duodenum.

In fact, 26 out of 30 patients with ulcers had the bacteria. " ose four patients with ulcers but negative for the bacteria had taken non-steroidal, anti-in$ ammatory drugs known to induce ulcers. Warren and Marshall now felt they had su% cient evidence linking Helicobacter pylori with gastritis and peptic ulcer disease. " e next step was to present their # ndings.

Presenting Their FindingsMarshall and Warren presented their # ndings to the College of Physicians at RPH. " eir work was not well received. Most were skeptical of their claims. Of the many objections, the main one was the link between gastritis and duodenal ulcers. Prevailing wisdom said that gastritis was associated with gastric ulcers, not duodenal ulcers. Marshall did further research and found a study done at the Mayo Clinic in the 1950s. In this study, stomachs from either motor vehicle fatalities or partial gastrectomy patients were examined for the presence of gastritis and ulcers. Researchers found a surprisingly high number of specimens that showed evidence of gastritis. Of those with duodenal ulcers, 100% (N = 250) also had gastritis. Here was evidence that supported Warren and Marshall’s hypothesis. Why had the ulcer research community skipped over the relationship between gastritis and duodenal ulcer? Marshall speculated it was because it didn’t # t with current beliefs and was therefore ignored.

Marshall continued reviewing the literature for any information pertaining to their research. He read that while acid-reducing drugs relieved the symptoms of the ulcers, they did not cure them. " e relapse rate was high. Cimetidine, a common acid-reducing agent, helped to heal the ulceration, but did not cure the disease. Marshall was further intrigued when he read that colloidal bismuth subcitrate (CBS; similar to the over-the-counter drug PeptoBismol) greatly reduced the relapse rate in those with gastric ulcers. Marshall hypothesized that if a drug lowered the relapse rate for gastric ulcers and if bacteria can cause gastric ulcers, then CBS should have anti-microbial properties. To test

Table II: Association of Bacteria with Endoscopic Diagnosis

Endoscopic Appearance* Total With Bacteria p

Gastric ulcer 22 18 (77%) 0.0086

Duodenal ulcer 13 13 (100%) 0.00044

All ulcers 31 27 (87%) 0.00005

Oesophagus abnormal 34 14 (41%) 0.996

GastritisT 42 23 (55%) 0.78

DuodenitisT 17 9 (53%) 0.77

Bile in stomach 12 7 (58%) 0.62

Normal 16 8 (50%) 0.84

Total 100 58 (58%)

*More than one description applies to several patients.T Refers to endoscopic appearance, not histological in! ammation.

Table III: Histological Grading of Gastritis and Bacteria

Bacterial Grade

Gastritis Nil 1+ 2+ 3+ total

Normal* 29 2 0 0 31

ChronicT 12 9 7 1 29

Active 2 5 15 18 40

Total 43 16 22 19 100

*Gastritis grades 0 and 1 are normal.T1 case showed bacteria on gram stained smear.

Fig. 3: Data Tables Published in The Lancet by Warren and Marshall in 1984


Design an Experiment

In small groups design an experiment to test the “bacterial” hypothesis.Be specific about how you would treat your experimental groups and what you would measure.


PredictionPredict what the results of the trial would look like after 3 months if the excess acid hypothesis was correct.

No. o

f pat

ient

s wi

th u

lcer

s

ExperimentalGroup

ControlGroup



Ex C Ex C

Ex C Ex C

A) B)

C) D)


Prediction

Predict what the results of the trial would look like after 3 months if the bacterial hypothesis was correct.

No. o

f pat

ient

s wi

th u

lcer

s

ExperimentalGroup

ControlGroup



Ex C Ex C

Ex C Ex C

A) B)

C) D)


Epilogue

Medical doctors worldwide were not easily convinced. Warren and Marshall needed more evidence.Marshall began treating patients with antibiotics, which appeared to work, but this was not a controlled experiment.Marshall attempted to infect pigs with bacterium but was unsuccessful.


Epilogue

In desperation Marshall drank 30 ml of the bacterial culture. Within 8 days he developed stomach pain, vomited and developed ulcers. Her cured himself with antibiotics.

In 2005, Warren and Marshall were awarded the Nobel prize in Medicine and Physiology.


modiﬁed from “an end to ulcers” by k. hannam and r...

Documents