IJC

International Journal of Cancer

Global burden of gastric cancer attributable

to Helicobacter pylori
ro
^me Vignat, David Forman and Catherine de Martel
Martyn Plummer, Silvia Franceschi, Je
International Agency for Research on Cancer, Lyon, France

Chronic infection with the bacterium Helicobacter pylori (H.

pylori) causes non-cardia gastric carcinoma (NCGC) and low
grade B-cell MALT lymphoma.1 We previously estimated
that worldwide 660,000 cases of cancer in the year 2008 were
attributable to H. pylori, corresponding to 32.4% of the 2 million cancer cases attributable to infectious agents and 5.2% of
the 12.7 million total cancer cases that occurred worldwide.2
The vast majority of the cancers attributable to H. pylori
(650,000) were NCGC.
These estimates of the global burden are based on the
attributable fraction (AF) for H. pylori, which is the proportion of cancers that would not have occurred if H. pylori was
absent from the world population. Quantication of the AF
requires accurate estimation of both the relative risk and the
prevalence of infection in cases.2 This effort was hampered
by the low sensitivity of serological tests for H. pylori in
NCGC cases. Gastric atrophy is a precursor lesion of gastric
cancer that reduces bacterial load and hence leads to a

declining antibody response, reducing the sensitivity of serological tests.2 Prospective studies in which the blood sample
for H. pylori testing was taken well before gastric cancer diagnosis are considered to be the most reliable. We previously
used a pooled analysis of 11 such studies in which blood
draw took place at least 10 years before the development of
cancer.3 In that pooled analysis, the average prevalence of
H. pylori infection in cases was 90%, and the corresponding
relative risk was 5.9, yielding an AF of 74.7% for H. pylori in
NCGC.
We now consider our AF estimate for NCGC to be an
underestimate. In recent years, evidence has accumulated that
immunoblot (western blot) is more sensitive for detection of
anti-H. pylori antibodies than ELISA, the detection method
used in the above pooled analysis. The purpose of this short
report is to update the AF estimate for H. pylori after briey
reviewing the evidence, and to reassess the global burden of
cancer attributable to H. pylori.

Key words: Helicobacter pylori, gastric cancer, attributable fraction,

immunoblot
Abbreviations: AF: attributable fraction; EBV: Epstein-Barr virus;
H. pylori: Helicobacter pylori; NCGC: non-cardia gastric cancer
Grant sponsor: Fondation de France; Grant number: 00039621
DOI: 10.1002/ijc.28999
History: Received 15 Apr 2014; Accepted 26 May 2014; Online 29
May 2014
Correspondence to: Martyn Plummer, International Agency for
Research on Cancer, 150 cours Albert Thomas, 69372 Lyon cedex
08, France; Tel.: 133-(0)4-7273-8446; Fax: 133-(0)4-7273-8345;
E-mail: plummer@iarc.fr

Methods

C 2014 UICC
Int. J. Cancer: 136, 487490 (2015) V

A review of the literature was undertaken in Pubmed and Web

of Science, using the following keywords: Helicobacter,
immunoblot*, western-blot*, and gastric neoplasms. We
included prospective studies and casecontrol studies in which
the risk of developing NCGC in individuals with and without
cancer was compared, using both ELISA and multiple antigen
immunoblot for detection of H. pylori. Studies that had used
immunoblot only for detection of antibodies to the CagA protein, which is expressed by some H. pylori strains, were not
included. The results from prospective studies were combined
by xed effects meta-analysis to give pooled estimates of the

Short Report

We previously estimated that 660,000 cases of cancer in the year 2008 were attributable to the bacterium Helicobacter pylori
(H. pylori), corresponding to 5.2% of the 12.7 million total cancer cases that occurred worldwide. In recent years, evidence
has accumulated that immunoblot (western blot) is more sensitive for detection of anti-H. pylori antibodies than ELISA, the
detection method used in our previous analysis. The purpose of this short report is to update the attributable fraction (AF)
estimate for H. pylori after briefly reviewing new evidence, and to reassess the global burden of cancer attributable to
H. pylori. We therefore reviewed the literature for studies comparing the risk of developing non-cardia gastric cancer (NCGC)
in cases and controls, using both ELISA and multiple antigen immunoblot for detection of H. pylori. The results from prospective studies were combined, and the new pooled estimates were applied to the calculation of the AF for H. pylori in NCGC,
then to the burden of infection-related cancers worldwide. Using the immunoblot-based data, the worldwide AF for H. pylori in
NCGC increased from 74.7% to 89.0%. This implies approximately 120,000 additional cases of NCGC attributable to H. pylori
infection for a total of around 780,000 cases (6.2% instead of 5.2% of all cancers). These updated estimates reinforce the
role of H. pylori as a major cause of cancer.

488

Helicobacter pylori in gastric cancer

Whats new?
Some 5.2 percent of the estimated 12.7 million new cancer cases in 2008, including 75% of non-cardia gastric cancer cases
(NCGC), were attributed to the bacterium Helicobacter pylori. However, those percentages may be an underestimate, given the
low sensitivity of detection of anti-H. pylori antibodies. Here, using improved estimates from prospective studies based on
immunoblot, the fraction of all cancers and NCGC attributable to H. pylori was found to be 6.2 and 89% percent, respectively.
Our findings reinforce the significance of the bacterium as a major cause of cancer.

relative risk and the prevalence of H. pylori in cases. These

pooled estimates were applied to the calculation of the AF for
H. pylori in NCGC, then to the burden of infection-related
cancers worldwide as previously described.2,4

Short Report

Results
Four published studies of NCGC have directly compared
ELISA with immunoblot (Table 1), all using the commercial
test Helico Blot 2.1 (Genelabs Diagnostics, Singapore). The
three prospective studies showed consistently higher prevalence of H. pylori among cases and consistently higher relative risks when using immunoblot compared with ELISA,58
whereas the casecontrol study found no difference between
the results by ELISA and immunoblot.9 Our literature search
also found a cross-sectional study of 54 Chinese patients
undergoing upper GI endoscopy using the earlier Helico Blot
2.0 test.10 This study is not included in Table 1 as it did not
directly compare immunoblot with ELISA, but we estimated
the odds ratio for NCGC from this study to be 20.9 (95% CI:
2.58980) for immunoblot, i.e., consistent with odds ratios in
Table 1.
When the results from the three prospective studies are
pooled, the prevalence in cases is 82.6% by ELISA and 94.6%
by immunoblot, and the relative risks are 4.77 (95% CI:
3.566.39) and 17.0 (95% CI: 11.625.0), respectively. Using
the immunoblot-based data, the AF for H. pylori in NCGC is
89.0%. Compared with our previous AF estimate of 74.7%
this implies approximately 120,000 additional cases of NCGC
attributable to H. pylori infection for a total of around
780,000 cases (6.2 % instead of 5.2% of all cancers). With
this revised AF, H. pylori is responsible for 36.3% of the 2.2
million cancers attributable to infection worldwide in the
year 2008. This proportion is higher in men (46.7%) than in
women (26.2%) and higher in more developed (50.3%) than
less developed regions (32.7%). Figure 1 shows the distribution of cancers attributable to infection by infectious agent
within more developed and less developed regions. As previously noted,2 H. pylori is one of four infectious agents that
dominate the infectious causes of cancer. The others are
human papillomavirus (causing cervical cancer, other anogenital cancers, and oropharyngeal cancers); hepatitis B virus
(hepatocellular carcinoma) and hepatitis C virus (hepatocellular carcinoma and non-Hodgkin lymphoma). With the
revised AF, H. pylori remains the most important infectious

agent causing cancer, especially in more developed countries,

where it bears half the burden of infection-related cancers.

Discussion
In three prospective studies conducted in Europe and Australia, multiple antigen immunoblot substantially increases
the detection of H. pylori compared with ELISA in cases but
not in controls. This differential increase translates into
higher odds ratios for NCGC. In contrast, the similar results
for ELISA and immmunoblot in the casecontrol study suggest that immunoblot cannot necessarily overcome the
known difculty of detecting H. pylori by ELISA in people
with advanced precancerous lesions.
Evidence for the higher AF for H. pylori in NCGC comes
exclusively from low-risk populations. The age-standardized
incidence rate for gastric cancer in men is 4.9 per 100,000
person years in Sweden, 6.7 in Australia and ranges between
4.9 and 11.0 (median 7.5) in the 10 European countries contributing to the European Prospective Investigation into
Cancer and Nutrition.11 People with H. pylori infection are
therefore a high-risk sub-group even in low-risk populations.
Extrapolation of our AF estimate of 89.0% to the world population, including high-risk countries for gastric cancer, is
justied by the absence of other strong risk factors for gastric
cancer. As noted by Gonzalez et al.,8 the AF for H. pylori
may be underestimated by non-differential misclassication
of H. pylori status. The Helico Blot 2.1 test has a reported
sensitivity of 95.6% (95% CI: 91.599.6) and specicity of
92.6% (95% CI: 91.596.2).12 We incorporated these estimates into a misclassication model to estimate the AF
adjusted for false positive and negative results. This yielded a
95% condence interval of 89.8 to 100.0% for the AF.
Although the condence interval is consistent with an AF of
100% this statistical evidence is not sufcient to classify H.
pylori as a necessary cause of NCGC. Around 10% of gastric
carcinomas contain Epstein-Barr virus (EBV).1 The role of
EBV and its interaction with H. pylori is not fully understood; it is possible that EBV is an independent cause of a
minority of NCGC.
The studies6,8,9 in Table 1 also emphasise the role of
CagA-positive H. pylori strains, as the Helico Blot 2.1 test
includes a 116kDa (CagA) band for the detection of antiCagA antibodies. In one of the prospective studies6 all of the
cases that were positive for H. pylori were also CagA-positive
(65/65). In another,8 all but one of the H. pylori-positive
C 2014 UICC
Int. J. Cancer: 136, 487490 (2015) V

489

Plummer et al.

Table 1. Studies of non-cardia gastric cancer comparing H. pylori assessment by ELISA and by immunoblot
ELISA
Study

Country

N cases Follow-up (yrs) % Cases positive

Immunoblot

OR

95% CI

% Cases positive

OR

95% CI

Prospective
n et al., 19975
Sima

Sweden1 27

5.7

89%

11.1 2.4-71.8

n et al., 2007
Sima

Sweden1 67

9.2

97%

17.8 4.2-74.8

Mitchell et al., 20087

Australia

34

11.6

79%

2.3

0.9-5.8

94%

10.6 2.4-47.4

88

10.7

82%

6.81 3.0-15.1

93%

21.4 7.1-64.4

272

84%

0.66 0.37-1.17 76%

lez et al., 20128 Europe2

Gonza

Casecontrol
Peleteiro et al., 20109

Portugal

0.73 0.46-1.17

The same cohort with different follow-up.

Multicentric study, part of the European Prospective Investigation into Cancer and Nutrition (EPIC).
Abbreviations: OR 5 odds ratio; CI 5 confidence interval.

Figure 1. Number of new cancer cases in 2008 attributable to

infection, by infectious agent and development status (updating
Fig. 2 of de Martel et al. 2012).2

cases were CagA-positive (81/82). The third did not report

specic ndings for CagA.7 The casecontrol study9 also
found a strong association when the 116kDa band was interpreted as CagA-positivity, ignoring the other markers that
are required to dene H. pylori infection according to the

manufacturers instructions. The prevalence of CagA in cases

was 96.0% and the relative risk for CagA was 11.3 (95% CI:
5.6422.7), in contrast to the low prevalence and null relative
risk for H. pylori overall (Table 1). This reinforces the ndings of previous casecontrol studies that have supplemented
H. pylori detection by ELISA by CagA immunoblot.13,14
These previous studies were based on the hypothesis that
CagA antibodies are more persistent than antibodies to H.
pylori surface antigens, and are therefore a more sensitive
marker to past infection with a CagA-positive strain. It
should be noted, however, that this hypothesis is not
accepted by all authors.15 Further evidence for the role of
cagA-positive H. pylori strains come from studies of precancerous lesions, which allow detection and typing using PCR
on gastric biopsies. Such studies have shown strong crosssectional associations between cagA-positive strains and
severity of precancerous lesions16 and increased progression
among cagA-positive strains that also have the more virulent
vacA s1/m2 genotype.17
In conclusion, immunoblot is to be preferred to ELISA in
epidemiological studies of NCGC, although it cannot completely overcome the problems of the retrospective assessment of infection status in casecontrol studies. The
increased estimates of the burden of cancer attributable to H.
pylori obtained by taking these studies into account reinforce
its role as a major cause of cancer.

Acknowledgement
Dr. Catherine de Martel was funded by the Fondation de France (grant number 00039621).

References
1.

2.

IARC. Monographs on the evaluation of

carcinogenic risks to humans volume 100b: a
review of human carcinogens: biological agents.
Lyon: International Agency for Research on
Cancer, 2012.
de Martel C, Ferlay J, Franceschi S, et al. Global
burden of cancers attributable to infections in

3.

C 2014 UICC
Int. J. Cancer: 136, 487490 (2015) V

2008: a review and synthetic analysis. Lancet

Oncol 2012;13:60715.
Helicobacter and Cancer Collaborative Group.
Gastric cancer and Helicobacter pylori: a
combined analysis of 12 case control studies
nested within prospective cohorts. Gut 2001;49:
34753.

4.

5.

Ferlay J, Shin HR, Bray F, et al. Estimates

of worldwide burden of cancer in 2008:
GLOBOCAN 2008. Int J Cancer 2010;127:
2893917.
Siman JH, Forsgren A, Berglund G, et al. Association between Helicobacter pylori and gastric carcinoma in the city of Malmo, Sweden. A

Short Report

490

6.

7.

8.

10.

11.

12.

13.

prevalence population. Cancer Causes Control

2010;21:125764.
Yang KC, Chu A, Liao CS, et al. Evaluation of
the role of H pylori infection in pathogenesis of
gastric cancer by immunoblot assay. World J Gastroenterol 2006;12:702932.
Ferlay J, Soerjomataram I, Ervik M, et al. GLOBOCAN 2012 v1.0, cancer incidence and mortality worlwide: IARC cancerbase no. 11. Lyon,
France: International Agency for Research on
Cancer, 2013. Available from: http://globocan.
iarc.fr.
Monteiro L, de Mascarel A, Sarrasqueta AM,
et al. Diagnosis of Helicobacter pylori infection:
noninvasive methods compared to invasive methods and evaluation of two new tests. Am J Gastroenterol 2001;96:3538.
Ekstrom AM, Held M, Hansson LE, et al. Helicobacter pylori in gastric cancer established by

14.

15.

16.

17.

CagA immunoblot as a marker of past infection.

Gastroenterology 2001;121:78491.
Brenner H, Arndt V, Stegmaier C, et al. Is Helicobacter pylori infection a necessary condition for
noncardia gastric cancer? Am J Epidemiol 2004;
159:2528.
Siman JH, Engstrand L, Berglund G, et al. Evaluation of western blot CagA seropositivity in Helicobacter pylori-seropositive and -seronegative
subjects. Clin Diagn Lab Immunol 2005;12:3049.
Plummer M, van Doorn LJ, Franceschi S, et al.
Helicobacter pylori cytotoxin-associated genotype
and gastric precancerous lesions. J Natl Cancer
Inst 2007;99:132834.
Gonzalez CA, Figueiredo C, Lic CB, et al. Helicobacter pylori cagA and vacA genotypes as predictors of progression of gastric preneoplastic
lesions: a long-term follow-up in a high-risk area
in Spain. Am J Gastroenterol 2011;106:86774.

Short Report

9.

prospective study. Scand J Gastroenterol 1997;32:

121521.
Siman JH, Engstrand L, Berglund G, et al. Helicobacter pylori and CagA seropositivity and its
association with gastric and oesophageal carcinoma. Scand J Gastroenterol 2007;42:93340.
Mitchell H, English DR, Elliott F, et al. Immunoblotting using multiple antigens is essential to
demonstrate the true risk of Helicobacter pylori
infection for gastric cancer. Aliment Pharmacol
Ther 2008;28:90310.
Gonzalez CA, Megraud F, Buissonniere A, et al.
Helicobacter pylori infection assessed by ELISA
and by immunoblot and noncardia gastric cancer
risk in a prospective study: the Eurgast-EPIC project. Ann Oncol 2012;23:13204.
Peleteiro B, Lunet N, Barros R, et al. Factors contributing to the underestimation of Helicobacter
pylori-associated gastric cancer risk in a high-

Helicobacter pylori in gastric cancer

C 2014 UICC
Int. J. Cancer: 136, 487490 (2015) V