Can Helicobacter pylori serology still be applied as a surrogate marker to identify peptic ulcer disease in dyspepsia?

Can Helicobacter pylori serology still be applied as a surrogatemarker to identify peptic ulcer disease in dyspepsia?

H. H.-X. XIA 1 , J . S. KALANTAR 1 , H. M. MITCHELL 2 & N. J. TALLEY 1

1Department of Medicine, The University of Sydney, Nepean Hospital, and 2School of Microbiology and Immunology,

The University of New South Wales, Australia

Accepted for publication 16 November 1999

INTRODUCTION

Peptic ulcer continues to be a common disease that

causes a substantial socioeconomic burden and negat-

ively impacts on quality of life.1, 2 The mortality from

peptic ulcer disease due to perforation or bleeding is still

of concern; the annual rate is 50 per million (or 850

deaths) in Australia and 67 per million (or 4000

deaths) in the UK.3±5 Helicobacter pylori infection has

been reported to be present in up to 80±100% of

patients with duodenal ulcer and 60±95% of patients

with gastric ulcer.6, 7 There are convincing data that

eradication of H. pylori infection leads to cure of

H. pylori-associated peptic ulcer and prevents ulcer

complications.6±10 Thus, identi®cation of H. pylori-

associated peptic ulcer is now an important issue for

both general practitioners and gastroenterologists.

Based on UK data that noninvasive screening of

patients for H. pylori infection using a urea breath test

or serology is useful for identifying peptic ulcer disease

amongst patients with dyspepsia,11±14 the British Soci-

ety of Gastroenterology has recommended that nonin-

vasive H. pylori testing be employed in patients with

dyspepsia prior to referring for an upper endoscopy.15 It

has further been suggested that H. pylori-negative

patients who are younger than 45 years do not require

an upper endoscopy if there are no alarm features.15 On

the other hand, the European Study Group has

recommended that patients with dyspepsia under

SUMMARY

Background: Helicobacter pylori infection and associated

peptic ulcer disease (PUD) has become less common in

some countries.

Aim: To determine if H. pylori serology alone or com-

bined with a history of ingestion of non-steroidal anti-

in¯ammatory drugs (NSAIDs) and an age threshold can

be used as an indirect ulcer test.

Methods: Two hundred and ®fty-two consecutive Aus-

tralian patients (121 males, mean age 52 years)

referred for endoscopy were enrolled. Blood was tested

by a validated ELISA. At endoscopy, eight biopsies were

taken for CLO-testing, culture and histology. NSAID use

over the prior 3 months was recorded.

Results: One hundred and six (42%) patients were

seropositive for H. pylori, 48 (19%) patients had PUD

and 30 (12%) used NSAIDs. Serology alone had a

sensitivity of 52% and a speci®city of 60% for identifying

PUD; the sensitivity and speci®city were 60% and 55%,

respectively, when combined with a history of NSAID

use. Serology, regardless of NSAID use, would have

saved 23% in endoscopy workload but would have

missed 17% of PUD cases if an age threshold of

< 45 years was chosen for omitting endoscopy.

Conclusions: Serology was a poor ulcer test despite an

excellent performance for detecting H. pylori. A strategy

combining serology and an age threshold with a history

of NSAID use to reduce endoscopy workloads may not

always be appropriate.

Correspondence to: Professor N. J. Talley, Department of Medicine, The

University of Sydney, Nepean Hospital, PO Box 63, Penrith NSW 2751,

Australia.E-mail: [email protected]

Aliment Pharmacol Ther 2000; 14: 615±624.

Ó 2000 Blackwell Science Ltd 615

45 years of age should receive empirical anti-H. pylori

therapy if noninvasive H. pylori testing is positive and

there are no alarm symptoms.16 The American Gast-

roenterological Association and an Asian Paci®c Con-

sensus Party have supported a similar `test and treat'

approach.17, 18 It is assumed that this approach will

reduce endoscopy workloads and will be cost-effect-

ive.16, 19, 20

Recent randomized controlled trials have reported that

H. pylori eradication is not of symptomatic bene®t for

patients with nonulcer dyspepsia.21, 22 While not all

studies agree,23 empirical anti-H. pylori therapy appears

not to be justi®able, at present, for all H. pylori-positive

patients with nonulcer dyspepsia if symptom relief is the

primary goal. Moreover, over several decades the

prevalence of H. pylori infection appears to have been

declining in developed countries, presumably due to an

improvement in living standards such as reduced

crowding in families.24, 25 Correspondingly, the inci-

dence of H. pylori-associated diseases such as peptic

ulcer and gastric cancer have also decreased.3, 26, 27

The prevalence of H. pylori infection and the incidence

of associated diseases will probably fall further as

effective eradication therapy for H. pylori infection is

more widely applied.7, 16±18, 28 Indeed, there have been

increasing reports of H. pylori-negative gastric and

duodenal ulcer.29±34 This will impact on a management

strategy that relies on noninvasive H. pylori testing.

In this prospective study, we aimed to determine

whether serology alone or combined with a history of

use of non-steroidal anti-in¯ammatory drugs (NSAIDs)

is a useful tool for identifying peptic ulcer disease

amongst patients presenting with upper gastrointestinal

complaints.

PATIENTS AND METHODS

Patients

Two hundred and ®fty-two consecutive patients referred

to the Nepean Hospital endoscopy unit in Australia

because of upper gastrointestinal symptoms were

enrolled. They were 121 males and 131 females, with a

mean age of 52.1 years (range 18±86 years). All patients

had no documented history of anti-H. pylori therapy in

the prior year. Use of NSAIDs over the prior 3 months

was recorded, based on an interview by a gastroenterol-

ogist (J.S.K.). Patients who had used aspirin or nonaspirin

NSAIDs for at least 3 days at any dosage during this

period of time were considered to be NSAID users.

Treatment with H2-receptor antagonists or proton pump

inhibitors over the prior 4 weeks was also recorded.

At endoscopy, a mucosal break with depth in the

stomach or duodenum was de®ned as an active ulcer,

whereas a healed ulcer was de®ned as present if there

was evidence of scarring or deformity. Re¯ux oesoph-

agitis was identi®ed by the presence of oesophageal

mucosal breaks.

All patients gave informed consent. The study was

approved by the Wentworth Area Health Service Ethics

Committee.

Serological assay

Blood samples (10 mL) were obtained from each

patient, and an enzyme-linked immunosorbent assay

kit, pylori DTect ELISA (Diagnostic Technology, Sydney,

Australia), which was validated in our laboratory, was

applied for the serological detection of H. pylori infec-

tion; of the patients included in the present study, 209

had been used previously to validate the kit.35

Con®rmation of H. pylori infection

At the time of endoscopic examination, eight biopsies

were obtained from each patient using a standard

mapping protocol.36 These included three biopsies taken

at the antrum within 2 cm of the pylorus along the

lesser curvature, two from midway between pylorus and

cardio±oesophageal junction at the greater curvature,

two from the deepest portion of the dome of the fundus

when retro¯exing the endoscope and one from the

angulus at the lesser curvature. One antral biopsy was

used for a rapid urease test (CLO-test; Delta West Pty

Ltd, Bentley, Western Australia). Three biopsies (one

each from the antrum, body and fundus) were cultured

on horse blood agar plates for 5±7 days under micro-

aerophilic conditions. Colonies suspected of being

H. pylori were Gram stained, and then urease activity

was tested. The remaining four biopsies were examined

for Helicobacter-like organisms by histology according to

the revised Sydney System after staining with haemat-

oxylin and eosin (H&E) as well as Giemsa.37 The

endoscopist and pathologist were unaware of the

serological status.

To establish the gold standard, H. pylori status was

only de®ned to be positive if a patient was shown to be

positive by one of the following: (i) culture (ii) the CLO-

616 H. H.-X. XIA et al.

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 615±624

test and histology on one or more biopsies, or

(iii) histology on at least two biopsies.38

Statistical analysis

The seroprevalence of H. pylori infection between

different groups of patients was assessed using the

Chi-squared test, with Yates's correction if required. All

P-values calculated were two-tailed; the alpha level of

signi®cance was set at P < 0.05.

RESULTS

Seroprevalence of H. pylori infection in peptic ulcer disease

and other endoscopic ®ndings

Overall, 106 (42.1%) of the 252 patients were seropos-

itive and 146 (57.9%) were seronegative. Thirteen of

the seropositive patients were negative, and four of the

seronegative patients were positive by the gold stan-

dard. Thus, the pylori DTect ELISA had a sensitivity and

speci®city of 96% and 92%, respectively, for the

detection of H. pylori infection. The prevalence of

H. pylori infection determined by the serology and

biopsy based tests is shown in Figure 1. The seropreva-

lence of H. pylori infection was signi®cantly higher in

males than in females (50% vs. 34%, P � 0.01), and

appeared to increase with age, with the rate being

signi®cantly higher in patients ³ 45 years of age than

those < 45 years of age (48% vs. 31%, P � 0.02)

(Table 1).

At endoscopy, 48 (19%) patients were diagnosed

as having chronic peptic ulcer disease (13 with duo-

denal ulcer, 29 with gastric ulcer and six with duo-

denal and gastric ulcers), 129 (51.2%) patients had

nonulcer dyspepsia (75 normal mucosa, 54 macro-

scopic gastroduodenitis), and two (0.8%) patients had

gastric cancer. The remaining 73 (29%) patients had

de®nite gastro-oesophageal re¯ux disease (seven

Barrett's oesophagus and 66 re¯ux oesophagitis).

The seroprevalence of H. pylori infection (52.1%) in

patients with peptic ulcer disease was not signi®cantly

different from the seroprevalence (44%) in patients

with a normal mucosa (Table 1). The lowest sero-

prevalence (30.1%) occurred in gastro-oesophageal

re¯ux disease; this was signi®cantly lower than in

peptic ulcer disease (v2 � 5.87, OR � 2.52, 95%CI

1.18±5.36, P < 0.02), but was not signi®cantly

different from the seroprevalence in nonulcer

dyspepsia (Table 1).

Overall, 69% of duodenal ulcer cases, 41% of gastric

ulcer cases and 67% of cases with both duodenal and

gastric ulcer were seropositive for H. pylori. The

seroprevalence was not signi®cantly different between

males and females with peptic ulcer disease (Table 1).

However, the seroprevalence in gastric ulcer and gastric

and duodenal ulcer appeared to be related to age; none

(0%) of the seven patients younger than 45 years was

seropositive for H. pylori, compared with a seropreva-

lence of 57.1% (16/28) in gastric ulcer patients

45 years or older (P � 0.01, Fisher's Exact Test, two-

tail) (Table 1).

Figure 1. Prevalence of H. pylori infec-

tion in 252 patients with dyspepsia and

re¯ux symptoms, according to age, as

determined by serology and biopsy-based

tests (BBT).

H. PYLORI SEROLOGY FOR IDENTIFYING PEPTIC ULCER DISEASE 617


One hundred and thirty-®ve patients had received

acid-suppressive treatment over the prior 4 weeks; 25

with proton pump inhibitors, 106 with H2-receptor

antagonists, and four with both. Nine patients (®ve with

duodenal ulcers and four with gastric ulcers) had

endoscopic evidence of healed ulcers; six of these

patients had received proton pump inhibitors or H2-

receptor antagonists. Thus, it is conceivable that ulcers

in some cases might have completely healed on acid-

suppressing agents and have been unidenti®able at the

time of endoscopy, which could have reduced the

association between the seroprevalence of H. pylori

infection and peptic ulcer disease. For this reason, we

separately re-analysed the data on patients who had not

received treatment with H2-receptor antagonists or

proton pump inhibitors (n � 117). Overall, 25 (21%)

of these patients had macroscopic evidence of peptic

ulcer disease; seven had duodenal ulcer, 15 gastric

ulcer, and three gastric and duodenal ulcer. Twelve

(48%) of these patients were seropositive for H. pylori

infection; the seroprevalence was 57.1% (4/7), 40% (6/

15) and 66.7% (2/3), respectively, in the three ulcer

groups. These results were very similar to those

described above where all cases were analysed.

Of the 39 patients with active peptic ulcer disease, 16

had gastric ulcers between 0.25 cm and 0.45 cm in the

diameter; all patients with duodenal ulcers had ulcers

ranging between 0.5 cm and 1.5 cm. To examine

whether H. pylori infection is more often associated

with active peptic ulcer disease, patients with ulcer sizes

of 0.5 cm or larger (n � 26) were re-analysed; the

seroprevalence was 75% (6/8), 44.4% (4/9) and 66.7%

(4/6), respectively, in patients with duodenal ulcer,

gastric ulcer and both duodenal and gastric ulcer.

These, again, were not signi®cantly different from the

rates obtained from all patients with peptic ulcer

disease.

NSAID use in peptic ulcer disease and other endoscopic

®ndings

Thirty patients were documented to have taken

NSAIDs; 20 aspirin, nine nonaspirin NSAIDs and one

both. Of these patients, 15 were seropositive and 15

seronegative (19 males and 11 females). NSAID use

was more common amongst patients ³ 45 years old

than in those < 45 years old (16.7% (28/168) vs.

2.4% (2/84), v2 � 10.9, P < 0.001). The prevalence of

NSAID use was signi®cantly higher in patients with

peptic ulcer disease than in those with nonulcer

dyspepsia (20.8% vs. 7.8%, v2 � 6.0, P < 0.05). How-

ever, there was no signi®cant difference in the

prevalence of NSAID use between seropositive and

seronegative patients with peptic ulcer disease (24% vs.

17%, P > 0.05) (Table 2).

Serology alone or combined with NSAID use and an age

threshold for detecting peptic ulcer disease

The overall sensitivity and speci®city of serology alone

for detecting peptic ulcer disease was 52.1% and 60.3%,

respectively, with a positive predictive value of 23.6%

and negative predictive value of 84.2%. When serology

was combined with a history of NSAID use, the

Table 1. Seroprevalence of H. pylori infection amongst peptic ulcer disease and other endoscopic ®ndings, according to age and gender

Percentage (%) of seropositive patients according to endoscopic ®ndings*

Peptic ulcer disease Other ®ndings

Duodenal

ulcer

Gastric

ulcer

Duodenal ulcer

+ gastric ulcer All PUD Normal

Gastro-

duodenitis

Gastric

cancer GERD Overall

Age group (years)

< 45 80 (4/5) 0 (0/6) 0 (0/1) 33 (4/12) 36 (10/28) 41 (7/17) NA (0/0) 19 (5/27) 31 (26/84)

³ 45 63 (5/8) 52 (12/23) 80 (4/5) 58 (21/36) 49 (23/47) 46 (17/37) 100 (2/2) 37 (17/46) 48 (80/168)

Gender

Male 67 (6/9) 47 (8/17) 75 (3/4) 57 (17/30) 61 (14/23) 56 (15/27) 100 (1/1) 35 (14/40) 50 (61/121)

Female 75 (3/4) 33 (4/12) 50 (1/2) 44 (8/18) 37 (19/52) 33 (9/27) 100 (1/1) 24 (8/33) 34 (45/131)

Total 69 (9/13) 41 (12/29) 67 (4/6) 52 (25/48) 44 (33/75) 44 (24/54) 100 (2/2) 30 (22/73)** 42 (106/252)

*Seropositive number/case number in parentheses; **Compared with peptic ulcer disease, P < 0.05.

PUD, peptic ulcer disease; GERD, gastro-oesophageal re¯ux disease; NA, not applicable.



sensitivity increased to 60.4% but the speci®city de-

creased to 54.9%, with a positive predictive value of

24% and negative predictive value of 85.5%. When an

age threshold of 45 years was chosen as an additional

criterion to serology alone or in combination with a

history of NSAID use, the sensitivity increased to 83.3%,

but the speci®city declined to 24% (Table 3). In other

words, serology alone would have saved 58% of endos-

copy workload but missed 48% of peptic ulcer cases. A

combination of serology with a history of NSAID use

would slightly increase the detection rate for peptic ulcer

disease, but the endoscopy workload saved would be

reduced accordingly. When the age threshold of 45 years

was added, 23% of endoscopies would have been saved,

but 17% of peptic ulcer cases would have been missed,

regardless of a history of NSAID use (Table 3).

Analysis of patients who had not received acid-

suppressing agents showed that serology alone had a

sensitivity of 48% (12/25) and a speci®city of 58.7%

(54/92) for predicting ulcers. The sensitivity and

speci®city were 52% (13/25) and 53.3% (49/92),

respectively, when serology was combined with a

history of NSAID use. Adding an age threshold of

45 years or older, the sensitivity and speci®city were

80% (20/25) and 21.7% (20/92), respectively. These

®gures were slightly lower than those obtained when all

patients were analysed.

When patients with re¯ux oesophagitis and Barrett's

oesophagus were excluded, the sensitivity, speci®city,

positive predictive value and negative predictive value

for detecting peptic ulcer disease were 52.1%, 55%,

29.8% and 75.8%, respectively, for serology alone, and

60.4%, 51.1%, 31.2% and 77.9%, respectively, for

serology plus a history of NSAID use. When the age

threshold of 45 years was added, the sensitivity, spec-

i®city, positive predictive value and negative predictive

value for detecting peptic ulcer disease were 83.3%,

34.4%, 31.7% and 84.5%, respectively (Table 3). These

results were similar to those obtained when all patients

were analysed, suggesting that exclusion of patients

with gastro-oesophageal re¯ux disease did not signi®-

cantly improve the performance of serology for detecting

peptic ulcer disease in dyspepsia.

DISCUSSION

Serology is an accurate test for diagnosis of H. pylori

infection. Our study applied the pylori DTect ELISA test

and achieved a sensitivity of 96% and speci®city of

92%. However, it appears that serology is unable to

Table 2. Distribution of patients with

a history of NSAID use in the prior

3 months

Percentage (%) of patients taking NSAIDs*

Seropositive

for H. pylori

Seronegative

for H. pylori Overall

Age

< 45 3.8 (1/26) 1.7 (1/58) 2.4 (2/84)

³ 45 18 (14/80) 16 (14/88) 17 (28/168)**

Gender

Male 15 (9/61) 17 (10/60) 16 (19/121)

Female 13 (6/45) 5.8 (5/86) 8.4 (11/131)

Endoscopic ®ndings

Non-ulcer dyspepsia 8.9 (5/56) 6.9 (5/72) 7.8 (10/129)

Normal 9.1 (3/33) 7.1 (3/42) 8.0 (6/75)

Gastroduodenitis 8.7 (2/23) 6.5 (2/31) 7.4 (4/54)

All peptic ulcer disease 24 (6/25) 17 (4/23) 21 (10/48)***

Duodenal ulcer 11 (1/9) 0 (0/4) 7.7 (1/13)

Gastric ulcer 42 (5/12) 18 (3/17) 28 (8/29)

Duodenal ulcer + gastric ulcer 0 (0/4) 50 (1/2) 17 (1/6)

GERD 18 (4/22) 12 (6/51) 14 (10/73)

Gastric cancer 0 (0/2) NA (0/0) 0 (0/2)

Total 14 (15/106) 10 (15/146) 12 (30/252)

*Number of patients with NSAID use/number of total cases in parentheses. **Compared with

patients < 45 years old, P < 0.001. ***Compared with patients with nonulcer dyspepsia,

P < 0.05.

NA, not applicable; GERD, gastro-oesophageal re¯ux disease.



Ta

ble

3.

Th

ev

alu

eo

fp

osi

tiv

eH

.py

lori

sero

log

ya

lon

eo

rw

ith

ah

isto

ryo

fN

SA

IDu

sea

ta

na

ge

thre

sho

ldo

f4

5y

ears

for

iden

tify

ing

pep

tic

ulc

erd

isea

sein

pa

tien

tsw

ith

up

per

ga

stro

inte

stin

al

sym

pto

ms

H.

pylo

ri

sero

log

yN

SA

IDu

seA

ge

³4

5y

ears

H.

pylo

rise

rolo

gy

/

NS

AID

use

H.

pylo

rise

rolo

gy

/

ag

e³

45

yea

rs

H.

pylo

rise

rolo

gy

/NS

AID

use

/ag

e³

45

yea

rs

+±

+±

+±

+±

+±

+±

All

pati

ents

*

Wit

hp

epti

cu

lcer

s(n

=4

8)

25

23

10

38

36

12

29

19

40

84

08

Wit

ho

ut

pep

tic

ulc

ers

(n=

20

4)

81

12

32

01

84

13

27

29

21

12

15

45

01

55

49

Sen

siti

vit

y(%

)5

22

17

56

08

38

3

Sp

eci®

city

(%)

60

90

55

55

25

24

Lik

elih

oo

dra

tio

1.3

2.1

1.2

1.3

1.1

1.1

Po

siti

ve

pre

dic

tiv

ev

alu

e(%

)2

43

32

12

42

12

1

Neg

ati

ve

pre

dic

tiv

ev

alu

e(%

)8

48

38

68

58

68

6

En

do

sco

py

sav

ed(%

)5

88

83

34

82

32

3

Pat

ien

tsw

ith

dysp

epsi

aon

ly**

Wit

hp

epti

cu

lcer

s(n

=4

8)

25

23

10

38

36

12

29

19

40

84

08

Wit

ho

ut

pep

tic

ulc

ers

(n=

13

1)

59

72

10

12

18

64

56

46

78

64

58

64

5

Sen

siti

vit

y(%

)5

22

17

56

08

38

3

Sp

eci®

city

(%)

55

92

34

51

34

34

Lik

elih

oo

dra

tio

1.2

2.8

1.1

1.2

1.3

1.3

Po

siti

ve

pre

dic

tiv

ev

alu

e(%

)3

05

03

03

13

23

2

Neg

ati

ve

pre

dic

tiv

ev

alu

e(%

)7

67

67

97

88

58

5

En

do

sco

py

sav

ed(%

)5

38

93

64

83

03

0

*All

pa

tien

ts(n

=2

52

)w

ith

up

per

ga

stro

inte

stin

al

sym

pto

ms

are

incl

ud

ed;

**P

ati

ents

wit

hg

ast

ro-o

eso

ph

ag

eal

re¯

ux

dis

ease

(n=

73

)w

ere

excl

ud

ed.



accurately predict peptic ulcer disease; in the present

study, serology missed 48% of peptic ulcer cases. Even

when serology was combined with a history of NSAID

use, this approach missed 40% of peptic ulcer cases,

despite potentially saving 50% in endoscopy workloads

if such a strategy was applied. When an age threshold of

45 years was added, one-quarter of endoscopy work-

loads would be saved, but one-®fth of ulcer cases would

be missed.

Eradication of H. pylori infection cures H. pylori-

associated peptic ulcer disease, leading to a substantial

decrease in the number with peptic ulceration.7±10

Therefore, while the overall prevalence of peptic ulcer

cases, particularly duodenal ulcer cases, is decreasing,

the proportion of H. pylori negative peptic ulcer cases is

increasing, leading to a high percentage of patients with

peptic ulcer disease who are missed by serology. In the

present study, the low prevalence of duodenal ulcer

disease may be due to the fact that anti-H. pylori therapy

has been used to treat duodenal ulcer for more than a

decade in Australia.39

Many previous studies have recommended an age

threshold of 45 years or greater before referring patients

with upper gastrointestinal symptoms for endos-

copy.12±14, 40 These studies have shown that 93±

100% of peptic ulcer patients under 45 years old were

seropositive for H. pylori.13, 14, 40 However, the present

study observed that only 33% (4/12) of young ulcer

patients were seropositive for H. pylori, which means

that up to 70% of ulcer patients under 45 years old

would have been missed if serology had been used as

screening tool in this population (Table 1). It is also

notable that in this age group all seven patients with

gastric ulcer were missed by serology testing. Therefore,

factors other than H. pylori infection are likely to play a

role in peptic ulcer disease, particularly gastric ulcer, in

this young population.29±34

Whether serology can be applied as a screening tool for

peptic ulcer disease depends in part on the seropreva-

lence of H. pylori infection in patients with peptic ulcer

disease. Duodenal ulcers may occasionally be due to

NSAIDs.29±34 On the other hand, gastric ulcer disease is

more strongly linked to NSAIDs.30, 31 Thus, a higher

prevalence of gastric ulcer contributes to an increased

number of H. pylori-negative peptic ulcer cases.30, 31

The different results observed between the present study

and others may be due in part to the different

distribution (i.e. ratio of gastric ulcer cases over

duodenal ulcer cases, or gastric ulcer/duodenal ulcer Ta

ble

4.

Eff

ect

of

pre

va

len

cea

nd

ulc

erd

istr

ibu

tio

n(r

ati

oo

fg

ast

ric

ulc

erca

ses

ov

erd

uo

den

al

ulc

erca

ses)

on

the

pre

dic

tiv

ev

alu

eo

fse

rolo

gy

for

det

ecti

ng

pep

tic

ulc

erd

isea

se

Pre

va

len

ce(%

)o

fG

ast

ric

ulc

er/

Au

tho

rsC

ou

ntr

y

No

.o

f

case

s

Ga

stri

c

ulc

er

Du

od

ena

l

ulc

er

Ga

stri

cu

lcer

+

du

od

ena

lu

lcer

All

PU

D

du

od

ena

l

ulc

erra

tio

Ser

op

rev

ale

nce

(%)

of

H.

pylo

ri

Ser

op

rev

ale

nce

(%)

inP

UD

So

ba

laet

al.

19

91

11

UK

29

35

.88

.50

14

.30

.68

55

91

Co

llin

set

al.

19

92

12

*U

K8

7N

A3

4.5

NA

34

.5N

A6

79

3

Men

da

llet

al.

19

92

13

UK

36

72

.71

1.2

01

3.9

0.2

47

49

6

Th

am

etal

.1

99

41

4*

UK

52

NA

13

.5N

A1

3.5

NA

48

10

0

Pa

tel

etal

.1

99

44

0**

UK

11

93

.41

2.6

01

60

.27

50

90

Pa

tel

etal

.1

99

54

2*

UK

11

33

.52

6.5

03

00

.13

80

94

McC

oll

etal

.1

99

74

3**

*U

K3

18

4.1

20

.44

.72

9.2

0.3

75

79

3

Sh

euet

al.

19

97

39

Ta

iwa

n1

83

22

.44

3.2

06

5.6

0.5

18

69

1

Fra

ser

etal

.1

99

64

4N

ewZ

eala

nd

39

94

.09

.30

13

.30

.43

55

85

Pa

lli

etal

.1

99

74

5It

aly

31

87

3.7

14

.40

18

.60

.25

71

89

Wer

dm

ull

eret

al.

19

98

46

Net

her

lan

ds

12

94

2.8

3.9

06

.70

.72

48

74

Th

ep

rese

nt

stu

dy

Au

stra

lia

25

21

1.5

5.2

2.4

19

2.2

42

52

PU

D,

pep

tic

ulc

erd

isea

se;

NA

,n

ot

ap

pli

cab

le.

*Pa

tien

tsu

nd

er4

5y

ears

old

on

lyw

ere

incl

ud

ed;

**H

.py

lori

infe

ctio

nw

as

det

ecte

db

ya

sali

va

ryIg

Gen

zym

e-li

nk

edim

mu

no

sorb

ent

ass

ay

(EL

ISA

);**

*H.

pylo

riin

fect

ion

wa

sd

etec

ted

by

14

C-u

rea

bre

ath

test

.



ratio) of peptic ulcer disease in these countries

(Table 4).12±14, 41±46 In the UK, the gastric ulcer/

duodenal ulcer ratio ranges from 0 to 0.68. In Taiwan,

New Zealand, Italy and the Netherlands, the ratios were

0.51, 0.43, 0.25 and 0.75, respectively. However, in

the present study, the gastric ulcer/duodenal ulcer ratio

was 2.2 (Table 4). It appears that an increased gastric

ulcer/duodenal ulcer ratio is associated with a decreased

seroprevalence of H. pylori infection in peptic ulcer

disease.

Intake of NSAIDs is common and may contribute to a

substantial proportion of H. pylori negative peptic ulcer

cases. Studies carried out in Sydney several years ago by

Borody et al. found that 38% of gastric ulcer cases and

6% of duodenal ulcer cases were H. pylori negative,

and that 66% of H. pylori negative gastric ulcer cases

and 44% of H. pylori negative duodenal ulcer cases had

taken NSAIDs.29, 32 Nensey et al. from the USA showed

that 23% of their duodenal ulcer cases were H. pylori

negative, of whom 83% had taken NSAIDs.33 It has

been also reported, in the USA and in Finland, that 60%

and 81%, respectively, of H. pylori negative gastric ulcer

patients had a history of NSAID use.30, 31 The present

study was able to identify NSAID use in only 21% of

patients with peptic ulcer disease and the rate was only

17% in seronegative patients with peptic ulcer disease.

One major reason may be that we only recorded recent

use of NSAIDs (in the 3 months prior to the enrolment).

There is also a possibility that some patients were

unaware that they were taking NSAIDs. One study has

shown that 13% of patients who claimed to have not

used aspirin had objective evidence of current aspirin

intake, as determined by platelet cyclooxygenase activ-

ity.47 While NSAID use is a major risk factor for peptic

ulcer disease, the present study suggests that a recent

history of NSAID use only minimally improves the

sensitivity of serology for identifying peptic ulcer disease

in practice.

The present study evaluated patients with upper

gastrointestinal symptoms referred for endoscopy to a

teaching hospital in Australia. Because the prevalence

of peptic ulcer disease is likely to be lower in general

practice, the performance of the serology in the general

practice setting may be even poorer. For example, if the

prevalence of peptic ulcer disease in dyspepsia is 10% in

general practice rather than 19% as observed in the

present study population, then the positive predictive

value of serology for identifying peptic ulcer disease

would decrease from 24% to 13%, despite an increase of

the negative predictive value from 83% to 92% (with

the same sensitivity of 52.1% and speci®city of 60.3%).

On the other hand, in populations with a prevalence of

peptic ulcer disease of 30%, the positive predictive value

would increase to 36% although the negative predictive

value would decrease to 75%. Therefore, H. pylori

serology appears to be unable to adequately serve as an

indirect marker of peptic ulcer disease when, in a given

population, the seroprevalence of H. pylori is low in

peptic ulcer disease.

In conclusion, serology achieved a relatively low

detection rate for peptic ulcer disease in this population

despite a high sensitivity and speci®city for identifying

H. pylori infection. A strategy combining serology and

an age threshold with a history of NSAID use to reduce

endoscopy workloads may no longer be appropriate in

some parts of the world.

ACKNOWLEDGEMENT

The authors thank Mr Stuart Howell for statistical

analysis.

REFERENCES

1 Jensen DM. Economic and health aspects of peptic ulcer

disease and H2-receptor antagonists. Am J Med 1986; 81:

42±8.

2 Wilhelmsen I, Berstad A. Quality of life and relapse of duo-

denal ulcer before and after eradication of Helicobacter pylori.

Scand J Gastroenterol 1994; 29: 874±9.

3 Sonnenberg A. Temporal trends and geographical variations

of peptic ulcer disease. Aliment Pharmacol Ther 1995;

9(Suppl. 2): 3±12.

4 Hugh TB, Kelly MD. Why are Australian peptic ulcer death

rates rising? Med J Aust 1991; 155: 136.

5 Axon A, Forman D. Helicobacter gastroduodenitis: a serious

infectious disease. Antibiotic treatment may prevent deaths in

the decades ahead. Br Med J 1997; 314: 1430±1.

6 Korman MG, Tytgat GNJ. Helicobacter pylori and peptic ulcer.

Scand J Gastroenterol 1995; 30(Suppl. 210): 92±6.

7 Consensus Development Panel on Helicobacter pylori in Peptic

Ulcer Disease. Helicobacter pylori in peptic ulcer disease. J Am

Med Assoc 1994; 272: 65±9.

8 Coghlan JG, Gilligan D, Humphries H, et al. Campylobacter

pylori and recurrence of duodenal ulcerÐa 12-month follow-

up study. Lancet 1987; ii: 1109±11.

9 Rauws EA, Tytgat GNJ. Cure of duodenal ulcer associated

with eradication of Helicobacter pylori. Lancet 1990; 335:

1233±5.

10 Graham DY, Hepps KS, Ramirez FC, Lew GM, Saeed ZA.

Treatment of Helicobacter pylori reduces the rebleeding in

peptic ulcer disease. Scand J Gastroenterol 1993; 28:

939±42.



11 Sobala GM, Crabtree JE, Pentith JA, et al. Screening dyspepsia

by serology to Helicobacter pylori. Lancet 1991; 338: 94±6.

12 Collins JSA, Bamford KB, Sloan JM, Collins BJ, Moorehead RJ,

Gary Love AH. Screening for Helicobacter pylori antibody could

reduce endoscopy workload in young dyspeptic patients. Eur J

Gastroenterol Hepatol 1992; 4: 991±3.

13 Mendall MA, Goggin PM, Marrero JM, et al. Role of Helicob-

acter pylori serology in screening prior to endoscopy. Eur J

Gastroenterol Hepatol 1992; 4: 713±7.

14 Tham TCK, McLaughlin N, Hughes DF, et al. Possible role of

Helicobacter pylori serology in reducing endoscopy workload.

Postgrad Med J 1994; 70: 809±12.

15 Axon ATR, Bell GD, Jones RH, Quine MA, McCloy RF.

Guidelines on appropriate indications for upper gastrointesti-

nal endoscopy. Br Med J 1995; 310: 853±6.

16 The European Helicobacter pylori Study Group. Current Euro-

pean concepts in the management of Helicobacter pylori in-

fection. The Maastricht Consensus Report. Gut 1997; 41:

8±13.

17 Anonymous. American Gastroenterological Association med-

ical position statement: evaluation of dyspepsia. Gastroente-

rology 1998; 114: 579±81.

18 Lam SK, Talley NJ. Helicobacter pylori Consensus. Report of the

1997 Asia Paci®c Consensus Conference on the management

of Helicobacter pylori infection. J Gastroenterol Hepatol 1998;

13: 1±12.

19 Sonnenberg A, Townsend WF. Costs of duodenal ulcer ther-

apy with antibiotics. Arch Intern Med 1995; 155: 922±8.

20 Briggs AH, Sculpher MJ, Logan RPH, Aldous J, Ramsay ME,

Baron JH. Cost effectiveness of screening for and eradication of

Helicobacter pylori in management of dyspeptic patients under

45 years of age. Br Med J 1996; 312: 1321±5.

21 Blum AL, Talley NJ, O'Morain C, et al. Lack of effect of treating

Helicobacter pylori infection in patients with nonulcer dys-

pepsia. N Engl J Med 1998; 339: 1875±81.

22 Talley NJ, Janssens J, Lauritsen K, Racz I, Bolling-Sternevald E.

Eradication of Helicobacter pylori in functional dyspepsia:

randomised double blind placebo controlled trial with 12

months' follow up. Br Med J 1999; 318: 833±7.

23 McColl K, Murray L, El-Omar E, et al. Symptomatic bene®ts

from eradicating Helicobacter pylori infection in patients with

nonulcer dyspepsia. N Engl J Med 1998; 339: 1869±74.

24 Haruma K, Okamoto S, Kawaguchi H, et al. Reduced incidence

of Helicobacter pylori infection in young Japanese persons be-

tween the 1970s and the 1990s. J Clin Gastroenterol 1997;

25: 583±6.

25 Sipponen P, Helske T, Jarvinen P, Hyvarinen H, Seppala K,

Siurala M. Fall in the prevalence of chronic gastritis over 15

years: analysis of outpatient series in Finland from 1977,

1985, and 1992. Gut 1994; 35: 1167±71.

26 Tytgat GNJ. No Helicobacter pylori, no Helicobacter pylori-

associated peptic ulcer disease. Aliment Pharmacol Ther

1995; 9(Suppl. 1): 39±42.

27 Sipponen P, Kimura K. Intestinal metaplasia, atrophic gastritis

and stomach cancer: trends over time. Eur J Gastroenterol

Hepatol 1994; 6(Suppl. 1): S79±83.

28 Xia HH-X, Talley NJ. Prospects for improved therapy for

Helicobacter pylori infection. Exp Opin Invest Drugs 1996; 5:

959±76.

29 Borody TJ, Brandl S, Andrews P, Jankiewicz E, Ostapowicz N.

Helicobacter pylori-negative gastric ulcer. Am J Gastroenterol

1992; 87: 1403±6.

30 Laine L, Marin-Sorensen M, Weinstein WM. Non-steroidal

anti-in¯ammatory drug-associated gastric ulcers do not re-

quire Helicobacter pylori for their development. Am J Gastro-

enterol 1992; 87: 1398±402.

31 Hyvarinen H, Salmenkyla S, Sipponen P. Helicobacter pylori-

negative duodenal and pyloric ulcer: role of NSAIDs. Digestion

1996; 57: 305±9.

32 Borody TJ, George LL, Brandl S, et al. Helicobacter pylori-

negative duodenal ulcer. Am J Gastroenterol 1991; 86:

1154±7.

33 Nensey YM, Schubert TT, Bologna SD, Ma CK. Helicobacter

pylori-negative duodenal ulcer. Am J Med 1991; 91: 15±8.

34 Talley NJ. Is Helicobacter pylori negative duodenal ulcer a

separate disease? In: Malfertheiner P, Ditschuneit H, eds.

Helicobacter Pylori, Gastritis, and Peptic Ulcer. Berlin: Springer-

Verlag, 1990: 340±4.

35 Xia HH-X, Kalantar JS, Ma Wyatt J, et al. High sensitivity and

speci®city of a laboratory-based serological test, pylori DTect

ELISA, for detection of Helicobacter pylori infection. Diag

Microbiol Infect Dis 2000; 36: 69±74.

36 Xia HH-X, Kalantar J, Talley NJ. Metronidazole- and clari-

thromycin-resistant Helicobacter pylori in dyspeptic patients in

Western Sydney as determined by testing multiple isolates

from different gastric sites. J Gastroenterol Hepatol 1998; 13:

1044±9.

37 Dixon MF, Genta RM, Yardley JH, Correa P and the partici-

pants in the International Workshop on the Histopathology of

Gastritis, Houston 1994. Classi®cation and grading of gastri-

tis. The Updated Sydney System. Am J Surg Pathol 1996; 20:

1161±81.

38 Thijs JC, van Zwet AA, Thijs WJ, et al. Diagnostic tests for

Helicobacter pylori: a prospective evaluation of their accuracy,

without selecting a single test as the gold standard. Am J

Gastroenterol 1996; 91: 2125±9.

39 Sheu B-S, Shiesh S-C, Yang H-B, Chen C-Y, Lin X-Z. Im-

plications Helicobacter pylori serological titer for the histo-

logical severity of antral gastritis. Endoscopy 1997; 29:

27±30.

40 Patel P, Mendall MA, Khulusi S, et al. Salivary antibodies to

Helicobacter pylori: screening dyspeptic patients before endos-

copy. Lancet 1994; 344: 511±2.

41 Marshall BJ, Goodwin CS, Warren JR, et al. Prospective dou-

ble-blind trial of duodenal ulcer relapse after eradication of

Campylobacter pylori. Lancet 1988; ii: 1437±42.

42 Patel P, Khulusi S, Mendall MA, et al. Prospective screening of

dyspeptic patients by Helicobacter pylori serology. Lancet

1995; 346: 1315±8.

43 McColl KEL, El-Nujumi A, Murray L, et al. The Helicobacter

pylori breath test: a surrogate marker for peptic ulcer disease

in dyspeptic patients. Gut 1997; 40: 302±6.



44 Fraser AG, Ali MR, McCullough S, Yeates NJ, Haystead A.

Diagnostic tests for Helicobacter pyloriÐcan they help select

patients for endoscopy? NZ Med J 1996; 109: 95±8.

45 Palli D, Vaira D, Menegatti M, Saieva C on behalf of the Italian

Helicobacter pylori Study Group. A serological survey of Heli-

cobacter pylori infection in 3281 Italian patients endoscoped

for upper gastrointestinal systems. Aliment Pharmacol Ther

1997; 11: 719±28.

46 Werdmuller BFM, V/Der Putten ABMM, Veenedaal RA,

Lamers CBHW, Loffeld RJLF. Can screening for IgG antibodies

against Helicobacter pylori be used in clinical practice? Omit

endoscopy in seropositive or negative patients? Dig Dis Sci

1998; 43: 2296±300.

47 Lanas A, Serano P, Bajador E, Esteva F, Benito R, Sainz R.

Evidence of aspirin use in both upper and lower gastrointes-

tinal perforation. Gastroenterology 1997; 112: 683±9.



Documents

Can Helicobacter pylori serology still be applied as a surrogate marker to identify peptic ulcer disease in dyspepsia?