Rare Helicobacter pylori Infection May Explain Low Stomach Cancer Incidence : Ecological Observations in Bali , Indonesia

Helicobacter pylori (H. pylori) is regarded as a carcinogen for stomach cancer along with other contributory factors, such as smoking and intake of salt and salty foods (IARC, 1994; 2003; Graham and Graham, 2002; WCRF/AICR, 2007). It destroys gastric mucous membranes and induces chronic atrophic gastritis which is the precursor lesion for stomach cancer (Parsonnet et al., 1991; Graham and Graham, 2002; Oishi et al., 2006). Similar to Indonesian ethnic groups, Malay people are well-known to have a low stomach cancer incidence, compared with high-incidence populations like Japanese and Chinese (IARC, 2002; Ministry of Health, Malaysia, 2002; Soeripto et al., 2003; RGPCRJ, 2003). Since it has been suggested that the incidence of stomach cancer is low in Bali, Indonesia from a hospital-based cancer registry


Introduction
Helicobacter pylori (H.pylori) is regarded as a carcinogen for stomach cancer along with other contributory factors, such as smoking and intake of salt and salty foods (IARC, 1994;2003;Graham and Graham, 2002;WCRF/AICR, 2007).It destroys gastric mucous membranes and induces chronic atrophic gastritis which is the precursor lesion for stomach cancer (Parsonnet et al., 1991;Graham and Graham, 2002;Oishi et al., 2006).
Similar to Indonesian ethnic groups, Malay people are well-known to have a low stomach cancer incidence, compared with high-incidence populations like Japanese and Chinese (IARC, 2002;Ministry of Health, Malaysia, 2002;Soeripto et al., 2003;RGPCRJ, 2003).Since it has been suggested that the incidence of stomach cancer is low in Bali, Indonesia from a hospital-based cancer registry

Subjects
In February 2006, we (IKM, M, and TGO) recruited 291 healthy volunteers (136 men and 155 women) from the general population in Bali, Indonesia.

Questionnaire survey
At Sanglah Hospital in Bali, the subjects were requested to respond to an interview-based questionnaire on lifestyle factors, including smoking, alcohol consumption and dietary habits, recent use of antibiotics, and drinking well water along with ethnicity (Balinese or Javanese) and religion.

Blood sampling
Overnight-fasting venous blood was sampled by hospital technicians to detect serum IgG antibodies for H. pylori and HAV-Ab, and to assess serum concentrations of PG I and PG II.Serum H. pylori IgG antibodies were examined using enzyme immunoassay (EIA) kits (Eiken Chemicals, Tokyo, Japan), and EIA values of ≥ 10 U/ ml were defined as positive.For the subjects showing positive UBT testing, serum CagA IgG antibodies were detected by enzyme-linked immunosorbent assay (ELISA) kits (Genesis Diagnostics Ltd., Cambridgeshire, United Kingdom) and ELISA values of ≥6.25 U/ml were defined as positive (Parsonnet et al., 1997;Ohata et al., 2004).Serum PG I and PG II were measured by chemical luminescence immunoassay (Abbott Japan Co. Ltd., Tokyo, Japan) with cut-off points at PG I <= 70 ng/ml and PG I/PG II <= 3.0 (Miki et al., 2003).HAV-Ab were analyzed by EIA (Abbott Japan Co. Ltd., Tokyo, Japan) (Fiore, 2004).Using white blood cells, we identified single-nucleotide polymorphisms (SNPs) of interleukin-1-beta C-31T (IL-1β C-31T), tumor necrosis factor-alpha T-1031C (TNF-α T-1031C), and PTPN11 G/A at intron 3 by two pairs of polymerase chain reaction (PCR-CTPP) (Ishida et al., 2006;Migita et al., 2007;Hamajima et al., 2008).

Urine sampling
In order to assess intakes of salt, fruit and vegetables, second morning voiding urine (SMVU) was sampled at the hospital.We estimated 24-hour urinary sodium and potassium excretions using SMVU according to Kawasaki's formula (Kawasaki et al., 1993).

Statistical analysis
Age-adjustment with the World Population (IARC, 2003) was made for values of stomach cancer incidence rates, H. pylori-related markers, smoking rates, and urinary excretions of salt and potassium.The relations between positive UBT and positive HAV-Ab were evaluated by Chi-square test.The associations across the SNPs (IL-1β C-31T, TNF-α T-1031C, and PTPN11 G/A at intron 3) and positive UBT were examined by Chisquare test after confirming that the frequencies of the control groups were within Hardy-Weinberg equilibrium.All data were analyzed by SAS software (ver.8.0, SAS Institute) and Excel.

Approval from ethics committee
The protocol was approved by the Ethics Committee of the Nagoya City University Graduate School of Medical Sciences, and the Research and Development Unit/ Research Ethics Committee of the Medical Faculty of the University of Udayana/Sanglah Hospital, Denpasar, Bali in Indonesia.

Informed consent
All subjects voluntarily took part in this study.Written informed consent was obtained from all participants and from the parents of minors.

Results
Age-adjusted annual incidence rates of stomach cancer in Bali were estimated to be 0.6/105 for men and 0.5/105 for women, approximately 1/100th of those in Japan (RGPCRJ, 2003), and were also lower than those of Malay people in Malaysia (Ministry of Health, Malaysia, 2002) and Singapore (IARC, 2002) (Table 1).
In Bali, salt excretions were estimated to be 13.3±4.1 g/day for men and 11.1±3.1 g/day for women, and there was no significant difference between the excretions of the Balinese participants and the consumption in Japan (Ministry of Health, Labour and Welfare, Japan, 2008).In addition, we found no significant difference in the intake of salt and salty foods between H. pylori-positive and -negative Balinese subjects (data not shown).Potassium excretion figures were 3.2±0.7 g/day for men and 2.8±0.6 g/day for women in Bali, being significantly higher than those in Japan (p<0.05) (Ministry of Health, Labour and Welfare, Japan, 2008).
Age-adjusted smoking rates of those 20 years of age or over were significantly lower for both men and women in Bali than those in Japan (p<0.05) (Ministry of Health, Labour and Welfare, Japan, 2006).Alcohol consumption rates of those 20 years of age or over in the Balinese were 15.2% (5.8-24.6)for men and 1.1% (0.0-2.7) for women, which were also significantly lower than those in Japan (Ministry of Health, Labour and Welfare, Japan, 2008).Percentages of recent use of antibiotics were 17% for men and 13% for women.Rates of using well water were 18% for men and 21% for women.

(Table 4).
Positive rates of HAV-Ab were 78.9% in men and 74.2% in women in Bali, which were almost as high as those in Malaysia (Tan et al., 1986) (Table 5), but higher than the corresponding figures in Japan (Kiyohara et al., 2007).Positive rates of HAV-Ab of those over 40 years age were 98.6% for men and 93.4% for women.There were no significant relations between positive UBT and positive HAV antibodies in the Balinese, being compatible with the observations in Japan (Furuta et al., 1997).
The genetic polymorphisms in the control groups were within Hardy-Weinberg Equilibrium.No significant associations were noted across IL-1β C-31T, TNF-α T-1031C, or PTPN11 G/A at intron 3 and UBT positivity (Table 6).

Discussion
We noted a very rare prevalence of H. pylori along with chronic atrophic gastritis and low incidence of stomach cancer in Bali, a result similar to those of our prior studies in other areas of Indonesia (Tokudome et al., 2005a;2005b).However, there were low smoking rates together with high intakes of fruit and vegetables, and no difference in the consumption of salt and salty foods between Bali and Japan.Thus, the low incidence of stomach cancer in Bali appeared attributable to a rare H. pylori infection and chronic atrophic gastritis along with low smoking rates and high consumption of fruit and vegetables.
We used UBT to detect the current H. pylori infection.UBT is regarded as the gold standard, which may be more sensitive than serological examinations, including H. pylori IgG (Cohen et al., 1999).Actually, although all our subjects with positive H. pylori IgG showed positive UBT, those with positive UBT were not always positive for H. pylori IgG.In other words, the specificity of the UBT test was distorted by bacteria having a urease function other than H. pylori.Nevertheless, we confirmed the low prevalence of H. pylori in Bali not only by UBT but also by H. pylori IgG antibodies.
H. pylori was also identified as a causative factor for chronic atrophic gastritis, a precursor lesion of stomach cancer, as was diagnosed using a PG test in a non-invasive examination (Oishi et al., 2006).The low positive rate of PG tests conducted in Bali was consistent with the fact that a history of gastric lesions rarely encountered by local clinicians (Moestikanigsih, personal communication), which may be due to the low positive rate of H. pylori.These results also appeared to be compatible with the low incidence of stomach cancer in Bali.
The cagA gene of H. pylori has been suspected of being a virulent factor for atrophic gastritis and stomach cancer (Parsonnet et al., 1997).We found that positive rates of CagA IgG detected in those with positive H. pylori IgG in Bali were as frequent as those in Japan.However, since the numbers of positive CagA IgG subjects in Bali appeared insufficient to establish the virulence of the cagA gene, we were unable to detect any statistically significant associations between that gene and stomach cancer.
The intake of salt and salty foods has been regarded as one of the risk factors for stomach cancer (WCRF/ AICR, 2007; Tsugane et al., 2004).It is conceivable that high intakes of salt and salty foods along with H. pylori infection synergistically affect the development of stomach cancer.However, the high consumption of salt and salty foods observed in both Bali and Japan does not in itself seem sufficiently responsible to explain the difference in the stomach cancer incidence between both areas.
We observed that the estimated excretion of potassium in Bali as a marker of the consumption of fruit and vegetables was significantly higher than the intake in Japan (Ministry of Health, Labour and Welfare, Japan,   (Kato et al., 1990;Inoue et al., 1994;Nishimoto et al., 2002;Goh et al., 2007).On the other hand, smoking has been well-known as a risk factor for stomach cancer (IARC, 2003).From our results, the low incidence of stomach cancer may be in part attributable to the higher consumption of fruit and vegetables and lower smoking rates in the Balinese.
It is suggested that H. pylori infection spreads via a fecal-oral, gastro-oral or oral-oral route (Graham and Graham, 2002), in contrast with HAV infection which is used as a marker of fecal-oral infection (Furuta et al., 1997).HAV may be transmitted by HAV contaminated foods and water (Serres et al., 1999).In agreement with a previous study (Furuta et al., 1997), the seropositivity of HAV did not synchronize with the seroprevalence of H. pylori in Bali.These results were compatible with the observations in Malay people in Malaysia (Goh and Parasakthi, 2001;Tan et al., 1986) (Table 5).One reason we were unable to verify the difference in infection routes/ modes between the two microorganisms may be that HAV is contracted directly by a single infection, whereas H. pylori infections appear to be established by multiple transmissions.
The inflammatory cytokine genes including IL-1β C-31T, TNF-α T-1031C, and PTPN11 G/A at intron 3 have been suggested to be associated with developing atrophic gastritis and gastric cancer (El-Omar et al., 2000;Goto et al., 2006;Sugimoto et al., 2007;Santos et al., 2012;Tahara et al., 2012;Zhao et al., 2013).However, the present study could not detect the association across these genetic polymorphisms and UBT positivity partly due to the small number of study subjects.Further investigation is warranted to identify an individual having a high risk of gastric cancer.
In terms of study limitations, an ecological study is admittedly generally recognized to be a rather unreliable method of generating evidence because it is not completely possible to eliminate ecological fallacies.The cancer incidence data were derived from a hospital-based cancer registry, not a population-based registration.Moreover, subjects were not all randomly recruited from the people of Bali, and the sample size in this study may be rather small to be representative with respect to the risk factors being studied.However, the very low prevalence of H. pylori observed in this study did not seem to be due to ecological errors or the small sample size.
In conclusion, we have confirmed that the exceptionally low prevalence of H. pylori infections may bring about not only a rare prevalence of chronic atrophic gastritis but also a low incidence of stomach cancer.Thus, H. pylori infection appeared to be a critical factor that determines gastric carcinogenesis, indicating that the prevention and control of stomach cancer may be successfully achieved by infection control and the eradication of H. pylori.

Table 2 . Demographic Characteristics, H. pylori-related Markers, Urinary Excretion of Salt and Potassium, and Smoking and Alcohiol Consumption Rates in Bali vs. Japan
(Watase et al., 2004)erent from Japan (p<0.01), †: Calculated using the 2006 data from the reference (Ministry of Health, Labor and Welfare, Japan, 2006) and age-adjusted by the World Population, NA: not available, #: Data from the reference(Watase et al., 2004), ##: 2005 data from the reference (Ministry of Health, Labor and Nutrition, Japan, 2008), All data in Bali were age-adjusted with the World Population

Table 6 . Associations Across Genetic Polymorphisms and UBT Positivity
Infection May Explain Low Stomach Cancer Incidence: Ecological Observations in Bali, Indonesia 2008).Case-control studies have suggested that such consumption reduces gastric carcinogenesis