Epidemiology of gastric and duodenal ulcers

Sarah Bowman

April 2008

What is an ulcer?

Acid breaks through protective substances on gut wall Duodenum (1st part small intestine) most common site Gastric ulcers in stomach

Pain, bloating, nausea, fullness, weight loss, tiredness

Complications: bleeding, perforation, obstruction

The duodenum is the first part of the small intestine, through which hydrochloric acid and pepsin flow to help digest food (1, 2). The duodenum makes chemicals (bicarbonate), mucus and prostaglandins which cover the inside lining and protect it from acid. Mucosal blood flow is also protective (3). An ulcer occurs where the acid breaks through the protection, not by over-production of acid (1). The duodenum is the most common site for a peptic ulcer - gastric ulcers are those occurring in the stomach (1).

Symptoms

Pain in the upper abdomen just below the sternum (breastbone) is the common symptom. It comes and goes and may occur mostly before meals. It may be eased on eating food, or taking antacid medicines (1)

Other symptoms may include: bloating, retching, nausea, particular 'fullness' after meals, food, weight loss and tiredness (2).

Complications develop in a small number of cases, but can be serious. These include:

Bleeding from the ulcer. Can range from a 'trickle' to a life-threatening bleed (1).

Perforation - the ulcer goes right through the wall of the duodenum. Food and acid in the duodenum leak into the abdominal cavity causing severe pain. This is a medical emergency. (1)

Narrowing and obstruction
Ulcers located at the end of the stomach where the duodenum is attached, can cause swelling and scarring which can narrow / close the intestinal opening. This obstruction can prevent food from leaving the stomach and entering the small intestine, resulting in vomiting (2)

Diagnosis

Endoscopy (>55yrs, first time). Capture all cases?Faecal / breath tests for H. pyloriGI series (rare)

False positive tests

Missed cases? risk of transmission / disease progression

Tests

Endoscopy (1). The NICE guidelines on Management of Dyspepsia in Adults states endoscopy is not required unless the patient is presenting for the first time above the age of 55 or unless there are other warning signs e.g. GI bleeding, dysphagia, unintentional weight loss, abdominal swelling and persistent vomiting" (3). There is a risk associated with endoscopy.

Faecal / breath tests for the H. pylori bacterium or from a biopsy sample taken during endoscopy (1)

A GI series (X-ray through swallowing barium) (2) (now v. rare (3)

There is a high frequency of false positive tests (14).

Cases may be missed if latent in the system for some time, and patient is asymptomatic / does not present. Particularly if faecal test does not detect bacteria but endoscopy not indicated (according to NICE)???

Why are they important?

HPA infectious disease but main effects are from chronic burdenPotential medical emergencyChronic symptoms health and economic costsH. pylori also linked to:Functional dypepsiaCancer (2-6x more likely, though still rare)Differential clinical outcome - interaction between bacterial properties (phenotypic variation), genetics and environmental / behavioural factors

Complications can be dangerous. Costs to the patient and the health service result.

Chronic symptoms costs to individuals health and wellbeing; economic costs in sickness absence

Other problems of having H. pylori:

Functional dyspepsia (4)
H. pylori is sometimes found in people with recurrent bouts of indigestion (dyspepsia) not caused by an ulcer or inflammation.

Cancer (4)
The risk of developing stomach cancer is thought to be slightly increased with long-term infection with H. pylori. However >25% of people in the UK become infected with H. pylori, and the vast majority do not get stomach cancer. Risk may be greater if person has H. pylori in addition to having a first degree relative diagnosed with stomach cancer.

Interventions may influence clinical outcome of infection. Some evidence exists for phenotypic variation in H. pylori bacteria, which may enable development of screening for bad bacteria in future, though this is not yet sufficiently understood (5).

Emergency admission for perforation

Rates per million resident population. Three-year moving averages


Implications for care of older people

(Bardhan et al. 2004, Digestive & Liver Disease 36(9), 577-588)

Rates are: higher for duodenal ulcer; higher with age.

Causes?

Ulcers are only found in white people, usually in long thin types given to worry and irritability (Robinson & Bruce 1940)

Causes & Risk Factors

Helicobacter pylori 90% duodenal ulcers70-75% gastric ulcersNSAIDsLifestyle factors increase risk smoking, physical stress, salt (GU)Genetic susceptibility / protection against H. pylori infection (twin studies, mouse models)RarelyZollinger-Ellison syndrome & others

Causes

Infection with H. pylori (1) in 90% of duodenal cases. The bacterium affects the duodenal lining by producing substances that weaken the protective mucus and produce more acid (2). H. pylori causes 70-75% of gastric ulcers (2, 3).

Anti-inflammatory medicines (NSAIDs non-steroidal anti-inflammatory drugs) taken for joint pain, etc. e.g. aspirin and ibuprofen sometimes affect the duodenal lining (1). Patients taking NSAIDs have 2 to 4 times the risk of ulceration if H. pylori is positive. Testing patients for H pylori and eradication before starting NSAIDs reduces the incidence of ulcers (3).

Lifestyle factors: smoking, stress, and drinking heavily may increase risk. However, these are not the underlying cause of ulcers (1). Smoking increases the chances of getting an ulcer, slows the healing process of existing ulcers and contributes to ulcer recurrence (2). Caffeine seems to stimulate acid secretion in the stomach, which can aggravate the pain of an existing ulcer. However, the stimulation of stomach acid cannot be attributed solely to caffeine (2). Although no proven link has been found between alcohol consumption and peptic ulcers, ulcers are more common in people who have cirrhosis of the liver, a disease often linked to heavy alcohol consumption (2). Although emotional stress is no longer thought to be a cause of ulcers, people with ulcers often report that emotional stress increases ulcer pain. Physical stress, however, may increase the risk of developing ulcers, particularly in the stomach. For example, people with injuries such as severe burns and people undergoing major surgery often require rigorous treatment to prevent ulcers and ulcer complications (2). High salt is implicated in GU (7).

Genetic:

Traditional family and twin studies have shown that concordance for H. pylori infection and duodenal ulcer is more common in monozygotic (identical) twins than in dizygotic (non-identical) twins (26, 27 in 7]. One study calculated the inherited component to be greater than the environmental component (27 in 7).

The second line of evidence is factors known to be genetically determined have been shown to be associated with disease. For example individuals with blood group O have a higher risk of duodenal ulcer, while blood group A is more common in those with gastric cancer (7). Some specific disease associations with HLA type have also been identified (7).

The third line of evidence is extremely convincing and comes from mouse models. Researchers have found the whole spectrum of possible conditions (including various degrees of gastritis, ulcers, cancer and lymphoma) can be produced by infecting genetically different mice with the same strain of Helicobacter felis (7). Preliminary evidence suggests that a variety of genes involved in the inflammatory response affect the likelihood of Hpylori infection progressing to duodenal ulcer disease (5).

Rare causes e.g. Zollinger-Ellison syndrome where more acid than usual is made by the stomach (1). Other rare causes include Crohns disease, gastric secreting tumours, hypergastinaemia, idiopathic gastrinoma and hyperparathyroidism (5).

Treatment & Outcome (NICE)

10% cases fail treatment (HPA)1 course of combination therapy clears most cases (74% duodenal ulcers)Relapse greater for gastric ulcers (affected by lifestyle factors). At 3-12 months:Duodenal ulcers: 39% clear (acid suppression only); 91% (combination therapy)Gastric ulcers: 45% clear (acid suppression only); 77% (combination therapy)

Treatment

H. pylori is the cause

Combination therapy: two antibiotics and an anti-acid medicine to allow the antibiotics to function. One course of therapy clears infection in up to 90% cases. Further tests may be done after a course of combination therapy if: symptoms do not go, or if they return; there is a complication e.g. perforation. Rarely surgery may be considered in cases not responding to treatment or developing complications (2). Anti-acid treatments only common since 1970s and 1980s (8).

NSAIDs are the cause

Stop the anti-inflammatory medicine. Where this is not possible, take acid-suppressing medicine daily (1).

Lifestyle modification

No need for "ulcer diets" since the introduction of effective acid suppression but reduction of alcohol intake, coffee and spicy foods may be needed. Smoking cessation is advised. The new COX2 inhibitors are less likely to cause ulcers than NSAIDs but it is dangerous to assume they have no risk. This advantage is nullified if the patient also takes even low dose aspirin (3).

Effectiveness of interventions

Effectiveness is important because it has an impact on prevalence and disease burden. NICE guidelines give the following data (3):

In duodenal ulcer, acid suppression for 4 to 8 weeks produces healing of the ulcer in 69%. This rises by an extra 5.4% with eradication therapy. NNT=18.

In duodenal ulcer relapse at 3 to 12 months after treatment is 39% after short term acid suppression alone but eradication increases this by 52% to 91%. NNT=2.

In gastric ulcer, supplementation of acid suppression with eradication therapy does not improve healing rates but it does reduce relapse so that 3 to 12 months later 45% are free of ulcers after just acid suppression but eradication raises this by 32% to 77%. NNT=3.

In patients taking NSAIDs, eradication did not improve ulcer healing rate but it did halve the number of endoscopically proven ulcers 6 months later from 18 to 9%.

The recurrence rate was significantly higher in gastric ulcer. Recurrence rates of patients who smoked, consumed alcohol, and used NSAIDs were significantly higher in those with gastric ulcer recurrence compared to duodenal ulcer. Relapsed ulcers tended to recur at the same or adjacent sites as the previous ulcers (3). Mass eradication of Hpylori is impractical because of the cost and the danger of generating antibiotic resistance, so we need to know how to target prophylaxis. (5)

Consequences

Primary care GP consultations, drug costs (increasing resistance)Secondary care complications, surgeryTertiary care rarely neededSocio-economic cost: Standardised average annual years of life lost (up to age 75) = 2.6 (per 10,000) (Females=1.8; Males=3.5) (1999 & 2001 pooled data, ONS)

High persistent morbidity. Burden is on primary care. Resistance to metronidazole occurs in approximately 50% of infected individuals in many European countries with significantly higher levels of up to 90% in developing countries. Resistance to clarithromycin is currently less common with levels below 10% in many European countries although there is evidence that rates may be increasing. Significantly, a recent meta-analysis has shown that pretreatment resistance to clarithromycin can reduce the effectiveness of therapy by 55%.

Socio-economic cost work days lost; discomfort and pain; quality of life / disability. HTA (2003) did not estimate cost per QALY. USA analysis (in NICE guidance) found significant cost-savings from a societal perspective on treatment more so for DU than GU.

Consequences

Mass eradication of Hpylori is impractical because ofgenerating antibiotic resistance, so we need to know how to target prophylaxis. (Calam & Baron 2001)Ulcers occurring in absence of H. pylori or NSAIDs / aspirin. Combination therapies less effective in absence of H. pylori data neededScreening? Cost-effective cost/LYS < 10,000 over 80yrs. But effects of eradication on morbidity / mortality?

Bardhan et al. (2004): Ulcers occurring in absence of H. pylori or NSAIDs / aspirin. Combination therapies less effective in absence of H. pylori data needed. Implications for services as ulcer surgery no longer practiced.

Screening

Roderick et al. (2003): HTA. Population screening H. pylori cost-effective: cost/LYS < 10,000 over 80yrs. Once-only screening at age 40

Major uncertainty re: effect of H.pylori eradication on gastric cancer and complicated ulceration risk; and antibiotic resistance. Favourable comparison with other screening programmes. Long follow up period therefore more evidence needed before can be recommended. The cost-effectiveness of screening would be reduced if extensive opportunistic screening of all dyspeptic patients presenting in primary care

Time, Person, Place

Barron & Sonnenberg (2002)

UK Incidence & Prevalence (Time)

Increases due to:

Increase in H.pylori? Different strain of H. pylori? Another concurrent gut infection?

But what about differing temporal changes of CU and DU and between men and women?

Graph shows admissions for gastric (GU) and duodenal (DU) ulcers, and dyspepsia in all 12 London hospitals associated with a medical school. Gastric ulcer, duodenal ulcer, and dyspepsia (including indigestion, gastralgia, and gastritis) are presented as mean rates per 10, 000 admission in each five year period. To further smooth out missing time periods and various lengths of data availability from individual hospitals, the data were plotted as the moving average of three consecutive time periods.

Peptic ulcer disease was comparatively rare until the mid-nineteenth century following which its incidence rose steadily, peaking 1940-1960. Since then there has been a slight decline, particularly in gastric ulcer (7).

These increases in peptic ulcer in the 19th century have been attributed to a postulated increase in an (cross) infectious disease (H pylori) in childhood, spread perhaps by the overcrowding and poor sanitation of early urbanisation. Or there may have been a new and more pathogenic strain of H pylori, a change in host susceptibility, or even a concurrent enteric infection modulating gastric inflammation and immune responses. However, these suggestions remain as yet untestable hypotheses, and do not explain the temporal changes in gastric versus duodenal ulcer and men versus women described (9).

UK Incidence & Prevalence (Time)

Evidence of cohort effect 1970-1986: (Primatesta et al. 1994)

Decreased hospitalized morbidity and mortalityRelated more to changes in risk factors (e.g. smoking) in different cohorts than new pharmacological treatments? - implications for public health!

OR: Genetic factors may be more important (Malaty et al. 1994)

UK Incidence & Prevalence (Person)

H. pylori infection

Incidence: 1-3% of adults p.a. (HPA)Prevalence infection: 40% population (HPA: >50% of 50+yr olds)

Ulceration

Incidence: DU in 30-50yrs old; higher incidence in men GU in >60yr olds; higher incidence in womenLow prevalence in younger age groupsDuodenal ulcer: up to 10% of population

In the UK around 40% of the population are infected and the possible clinical outcomes of this infection are diverse. The majority of individuals will have no symptoms and will not develop any related problems throughout their lives (7). 1% of the 40% will have an endoscopy and 13% of those will have ulcer disease annually.

There are differential outcomes, which are incompletely understood:

- Ulcers are rare among teenagers and children (2)

- Duodenal ulcers usually first occur between the ages of 30 and 50 (2)

- Duodenal ulcers occur more frequently in men than women (2)

- Stomach ulcers are more likely to develop in people over age 60 (2)

- Stomach ulcers develop more often in women than men (2)

UK Incidence & Prevalence (Person)

Current trends:

Annual age-standardized period prevalence decreased 1994-1998, particularly deprived areas (men 3.3/1000 - 1.5/1000; women 1.8/1000 - 0.9/1000)Sex incidence evening out decreasing incidence in young men; increasing in older womenBut emergency admission rates for complications unchanged in last 30yrs Kang et al. (2006) increase in case fatality for DU. Due to concomitant comorbidity / increasing ulceration (NSAIDs) / H.pylori (i.e. changing natural history)?Future decrease in prevalence?

Frequency decreasing (H. pylori decreasing), particularly in deprived areas

Increase in older women could be due to increased use of NSAIDs (non-steroidal anti-inflammatory drugs). Reason for little change in emergency admissions could be increase in use of aspirin and anti-coagulants, especially in older people Scottish data showed many admissions are for bleeding (8 Kang et al. 2006; 11)

More research needed re: potential changing natural history of ulceration.

Future decreasing prevalence? Current older cohorts were children when H. pylori was more common and are also now at an age where they are exposed to more NSAIDs and anti-platelet drugs. Additive effect could mean higher prevalence of ulcers in that cohort and high rates of complications (bleeding) (15).

International prevalence (Place)

Rates infection higher in less developed countries. Quite large variation.

Place

Worldwide. Prevalence 100% in developing countriesPotential for re-spread in UK through travel?H. pylori - oral / faeco-oral transmission associated with poverty / overcrowding in childhoodIncreased prevalence in children with history of ulcer in the mothers due to common environmental factors?Variation between ethnic groups even within countries

Precise mode of transmission of H.pylori unknown

Clustering?: study by Brenner et al. (1998) (6)

Variation between ethnic groups (14)

Place Regional variation (NCHOD)

2004-06 pooled data from ONSSMRDirectly standardised age specific death rates (per 100,000 European standard population) - regional centres higher than national average High rates affected by lifestyle factors & e.g. aspirin in deprived areas (raised CVD risk)? Lowest = E. Midlands (89). Highest = London (112) Industrial areas = 114, London suburbs = 111, London cosmopolitain = 153

E&W

NE

All

Males

Females

99 (97-101)

100 (97-103)

99 (96-101)

111 (102-121)

124 (110-139)

100 (88-113)

NCHOD = National Centre for Health Outcomes Development

2004-06 pooled data all age groups: age specific death rates (per 100,000) = 6.2 (E&W) (females =6.5; males = 6.0). NE (GONE / SHA)= 7.0 (males 7.6; females 6.4). Centres with industry 6.4 (thriving London periphery 5.3); industrial hinterland 6.8; coastal & countryside 8.4; regional centres 6.7. High due to older popn? Age- specific data for age bands removed for anonymity due to small numbers.

Directly standardised all age groups: age specific death rates (per 100,000 European standard population) = 3.88 (CI 3.80 3.96) (E&W) (males=4.70, CI=4.56-4.83; females=3.18, CI=3.08-3.27). GONE / NE SHA= 4.44 (CI=4.05-4.84); males=5.81 (CI=5.11-6.50); females=3.35 (CI=2.90-3.79). Centres with industry 4.68 (thriving London periphery 3.68); industrial hinterland 4.33; coastal & countryside 3.69; regional centres 4.29. Industrial centres / hinterland associated. Fit with effect of lifestyle factors and use of e.g. aspirin due to raised risk CVD in deprived areas? London cosmopolitain=6.25! Highest. Why?

Indirectly standardised: SMR (2004-2006 pooled). E&W=99 (97-101) males =100 (97-103), females=99 (96-101). NE=111 (102-121) (males=124 (110-139), females=100 (88-113)). Lowest = E. Midlands (89). Only 1 higher London +112. Industrial areas come out badly (114) but London suburbs 111 and London cosmopolitain=153!!

NCHOD data - critique

Based on original underlying cause of death (death certification)Numerator - mortality data 1993-2006 (ONS) with codes assigned using postcode of usual residenceChanges to coding causes of death in England & Wales. Data based on new coding Denominator data - latest revisions of ONS mid-year population estimates, current at Oct 2007 - quite accurateNCHOD regularly updated

-Death certificate problems particularly if older people with co-morbidities

- Comments on numerator data - Mortality data for years 1993-2006 were extracted by ONS in June 2007 with organisational codes assigned using the postcode of usual residence and the November 2006 edition of the National Statistics Postcode Directory

- In January 2001, the ONS implemented a change from ICD-9 to ICD-10 for coding causes of death in England & Wales. As part of an exercise to investigate the effects of this change, the ONS also re-coded all deaths registered in 1999. Deaths for years prior to 1999 and for year 2000 have not been re-coded.
For current analysis, therefore, all data are based on ICD-10.

- Denominator data based on latest revisions of ONS mid-year population estimates for respective years, current as at Oct 2007. So quite accurate

- NCHOD regularly updated

Data sources

Other potential sources: HES, primary care records, prescribing database

Data sourceValid?Reliable?Complete?Timely?Accessible?Relevant?Cochrane reviews (RCTs / meta-analyses) (RCTs / meta-analyses) Treatment focus PartialPubMed articles Some small studies Some small studies Epidemiol focus NEPHO ONS data ONS data Regional life years lost data London only 1999 & 2001

Data sources

Data sourceValid?Reliable?Complete?Timely?Accessible?Relevant?NCHOD ONS data. Updated though based on death certification ONS data. Updated though based on death certification Age-specific rates not split by sex (small numbers) Updated 2007 NEPHO website HPA Based on reporting from primary care Testing & trend monitoring. No routine publications

Data sources

Data sourceValid?Reliable?Complete?Timely?Accessible?Relevant?Patient UK NICE guidance & review articles. Written by clinicians. Reviewed 18monthly NICE guidance & review articles. Written by clinicians. Reviewed 18monthly Patient info. Based on recent guidance & articles Patient Plus Peer reviewed. Based on NICE guidance & review articles. Written by clinicians. Peer reviewed. Based on NICE guidance & review articles. Written by clinicians. Clinical focus diagnosis and management Based on recent guidance & articles Partial

Conclusion

Disease mechanism / transmission poorly understoodRisk factors multiple & interactingH. pylori is main cause but has changing natural historyEach generation has carried its own particular risk of bearing ulcers throughout adult life (Susser & Stein 1962)Current pattern = exposure to H. pylori + genetics + exposure to drugs + environmental / behavioural factorsDU / GU likely to continue causing significant chronic disease burden and personal / societal cost. Predicting future pattern difficult

Questions?

References

Bardhan, K. D., Williamson, M., Royston, C., Lyon, C. (2004) Admission rates for peptic ulcer in the Trent Region, UK, 1972-2000: Changing pattern, a changing disease? Digestive and Liver Disease, 36, 577-588

Barron & Sonnenberg, Gut (2002), 50(4), 568-570

Malaty, H., Engstrand, L., Pederson, N., Graham, D. (1994), "Helicobacter pylori infection genetic and environmental influences, a study of twins", Annals of Internal Medicine, Vol. 120 pp.982-6

Primatesta et al. (1994), Int. J. Epidemiol. 23(6), 1206-1217

Robinson, S.C., Bruce, R.M. (1940), "The body build of the ulcer patient", American Journal of Digestive Diseases, Vol. 7

Roderick P, Davies R, Raftery J, Crabbe D, Pearce R, Patel P. The cost-effectiveness of screening for Helicobacter pylori to reduce mortality and morbidity from gastric cancer and peptic ulcer disease: a discrete-event simulation model. Health Technol Assess, 2003;7(6). Available at: http://www.cinahl.com/cexpress/hta/summ/summ706.pdf

Bardhan et al. 2004
Elective surgery. Rates per million resident population. Three-year moving averages.

The number of elective anti-ulcer operations has declined, and more markedly so for DU. The greatest decline for both groups was in younger men, 3564 years, in whom such operations were most commonly performed (Fig. 4). The rate of decline, however, was no greater in the era of modern medical treatment (Table 3).