Article Text
PDF
Abstract
Objective Gastric infection with Helicobacter pylori is a strong risk factor for non-cardia gastric adenocarcinoma. The aim of this study was to assess whether the risk of gastric adenocarcinoma and non-cardia gastric adenocarcinoma decreases after eradication treatment for H. pylori in a Western population.
Design This was a nationwide, population-based cohort study in Sweden in 2005–2012. Data from the Swedish Prescribed Drug Registry provided information on H. pylori eradication treatment, whereas information concerning newly developed gastric adenocarcinoma was retrieved from the Swedish Cancer Registry. The risk of gastric adenocarcinoma and non-cardia gastric adenocarcinoma in individuals who had received H. pylori eradication treatment was compared with the background population of the corresponding age, sex and calendar year distribution, yielding standardised incidence ratios (SIRs) with 95% CIs.
Results During the follow-up of 95 176 individuals who had received eradication treatment (351 018 person-years at risk), 75 (0.1%) developed gastric adenocarcinoma and 69 (0.1%) developed non-cardia gastric adenocarcinoma. The risk of gastric adenocarcinoma decreased over time after eradication treatment to levels below that of the corresponding background population. The SIRs were 8.65 (95% CI 6.37 to 11.46) for 1–3 years, 2.02 (95% CI 1.25 to 3.09) for 3–5 years and 0.31 (95% CI 0.11 to 0.67) for 5–7.5 years after eradication treatment. When restricted to non-cardia adenocarcinoma, the corresponding SIRs were 10.74 (95% CI 7.77 to 14.46), 2.67 (95% CI 1.63 to 4.13) and 0.43 (95% CI 0.16 to 0.93).
Conclusion Eradication treatment for H. pylori seems to counteract the development of gastric adenocarcinoma and non-cardia gastric adenocarcinoma in this Western population.
- helicobacter pylori
- gastric cancer
- antibiotic therapy
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
The strongest risk factor for non-cardia gastric adenocarcinoma is gastric infection with Helicobacter pylori.
Only a small percentage (3%) of those infected with H. pylori develop non-cardia gastric adenocarcinoma.
Eradication of H. pylori is associated with a reduced risk of gastric cancer in Asian populations.
What are the new findings?
This study assessed the role of H. pylori eradication treatment in the prevention of gastric adenocarcinoma in a large population-based study, using data from high-quality nationwide Swedish registers.
The risk of gastric adenocarcinoma (non-cardia adenocarcinoma in particular) sharply decreased from 5 years onwards after eradication treatment for H. pylori.
How might it impact on clinical practice in the foreseeable future?
The findings of this study may contribute to a better understanding of the risks and benefits of eradication therapy in regions with a low prevalence of gastric cancer.
Introduction
Globally, gastric cancer (95% adenocarcinoma) is the fifth most common cancer (incidence rate 16.1 per 100 000 individuals) and the second most common cause of cancer death (death rate 13.8 per 100 000 individuals).1 The highest incidence is found in Asian populations, and the lowest in Western populations.1 In Sweden (a Western population), the incidence rates are 10.4 and 6.4 per 100 000 individuals in men and women, respectively.2 Infection with the bacterium Helicobacter pylori is associated with an increased risk of non-cardia gastric adenocarcinoma, whereas the risk of cardia gastric adenocarcinoma is not increased.3 Worldwide, around half of the population is infected, with higher rates in Asia and Central and South America and lower rates in Western populations.4 The H. pylori prevalence in Sweden is around 15%, which is relatively low.5 Of those infected, it is estimated that the lifetime risk of developing gastric adenocarcinoma is around 3%.6
H. pylori eradication, used for treatment of functional dyspepsia and the prevention of recurrent peptic ulcer after an earlier ulcer episode, has been found to prevent gastric cancer in meta-analyses, showing a 50% risk reduction.7 8 However, these meta-analyses almost exclusively included studies from Asian populations. Thus, there is a need to estimate the role of H. pylori eradication treatment in the prevention of gastric cancer in large-scale studies from Western populations with specific analyses of the risk of non-cardia adenocarcinoma. Therefore, this study aimed to assess the risk of gastric adenocarcinoma and of non-cardia gastric adenocarcinoma after eradication treatment for H. pylori in a nationwide Swedish cohort study.
Materials and methods
Study design
This was a population-based, nationwide cohort study in Sweden, based on an a priori established study protocol. The exposure was eradication treatment for H. pylori, and the outcomes were gastric adenocarcinoma and non-cardia gastric adenocarcinoma. Eligible for inclusion were all individuals residing in Sweden aged 18 years and older who received eradication treatment for H. pylori during 1 July 2005 to 31 December 2012. The cohort has been described in more detail in a descriptive study on the use of eradication treatment for H. pylori in Sweden.9
Exposure
Individuals were considered exposed if they received at least one prescription and dispensing of a recommended eradication regimen for H. pylori. Eradication treatment for H. pylori in Sweden usually consists of a combination of a proton pump inhibitor together with the antibiotics amoxicillin and clarithromycin for 7 days.9 The following prescriptions are available in Sweden for recommended eradication treatment of H. pylori (specified by the Anatomical Therapeutic Chemical code):
A specific H. pylori eradication package that includes the proton pump inhibitor esomeprazole and the antibiotics amoxicillin and clarithromycin (A02BD06);
Separate prescriptions of a proton pump inhibitor (A02BC) together with two of the following antibiotics: amoxicillin (J01CA04) and clarithromycin (J01FA09) or metronidazole (J01XD01).
For the separate prescriptions, no other antibiotics were to be prescribed at the same time to be considered a recommended regimen. Additionally, the antibiotics had to be prescribed on the exact same date, whereas the proton pump inhibitor was allowed to be prescribed within a window of 60 days before or 5 days after the antibiotics prescription. This window was chosen to include individuals already using a proton pump inhibitor before the eradication treatment and to take temporary non-availability in the pharmacy into account. Recommended eradication regimens accounted for 95.4% of all eradication treatments prescribed in Sweden during the study period.9 The remaining 4.6% of the treatments included combinations of antibiotics that are not standard for H. pylori eradication and were therefore not included in this study.
Outcomes
The outcomes were first and primary gastric adenocarcinoma and non-cardia gastric adenocarcinoma diagnosed in the Swedish Cancer Registry during follow-up of the cohort after H. pylori eradication treatment, excluding tumours occurring within 1 year of the eradication treatment. Gastric cancer was defined by any of the following International Classification of Diseases 7 (ICD7) codes: 151 (gastric cancer), 151.0 (non-cardia gastric cancer), 151.1 (cardia cancer), 151.8 (multifocal gastric cancer) and 151.9 (gastric cancer, not further specified), whereas the code for cardia cancer (151.1) was excluded when analysing non-cardia cancer. Adenocarcinoma histology was determined using the code 096 from the C24 WHO classification of histology. Other histological types of gastric malignancies are not associated with H. pylori (or have a different level of association, eg, mucosa-associated lymphoid tissue lymphoma) and therefore were excluded.10 Unfortunately, no data were available to assess the specific subtypes of gastric adenocarcinoma (intestinal, diffuse or mixed).
Confounders
Adjustments were made for age (categorised into the age groups 18–59, 60–69 or ≥70 years), sex (men or women) and calendar period (2005–2006, 2007–2009 or 2010–2012) by standardising these variables. Place of residence (urban or rural), which is correlated with socioeconomic and lifestyle factors,11 was analysed for influence on the outcome.
Data sources
The Swedish Prescribed Drug Registry provided information on H. pylori eradication treatment. This register started on 1 July 2005 and contains information on all prescribed and dispensed medications for the whole Swedish population. Patient identifying data are missing in only <0.3% of all recorded prescriptions.12 Variables used were the age, sex, postal code (to assess place of residence),11 unique personal identity number of each individual and the substance, dates, amount and dosage of the dispensed drugs.
The Swedish Cancer Registry provided information on gastric and non-cardia gastric adenocarcinoma. This register has a completeness of 98% for recording of non-cardia adenocarcinoma.13 Data on the histological type and location of the cancer, and date and age at cancer diagnosis were used, and all these variables are 100% completely recorded in the register.
The Swedish Causes of Death Registry was used to assess dates of death. This information is virtually 100% complete and was used for censoring of follow-up time.14
The Swedish Patient Registry was used for information on gastric histology and peptic ulcer diagnoses. Codes used were 540 (ICD7), 531 (ICD8,9) and K25 (ICD10) for gastric ulcer; 541 (ICD7), 532 (ICD8,9) and K26 (ICD10) for duodenal ulcer; 543,01 (ICD7), 535,03 (ICD8), 535C and 535F (ICD9) and K293 and K295 (ICD10) for chronic gastritis; 535B (ICD9) and K294 (ICD10) for atrophic gastritis. This information was collected from the Inpatient, Outpatient and Day Surgery Registries. The Swedish Inpatient Registry is nationwide complete since 1987 and was complemented with the Day Surgery Registry in 1997 and the Outpatient Registry in 2001.15 To the best of our knowledge, no validation for gastric lesions is available.
Statistical analysis
Standardised incidence ratios (SIRs) with 95% CIs were calculated as the ratio of the observed number of gastric and non-cardia gastric adenocarcinoma cases in the H. pylori eradication cohort compared with the expected number of such cases in the Swedish population of the same sex, age group and calendar period with categorisation as presented above (Confounders section). Clayton’s algorithm was used to calculate the follow-up time in person-years for each stratum, starting from the date of the first dispensed prescription of H. pylori eradication treatment.16 This index date was chosen as time zero to emulate a randomised clinical trial.17 Additionally, a ‘per-protocol’ analysis was performed which started from the last dispense date of eradication treatment. Individuals were followed up until the occurrence of any cancer, death or the end of the study period, whichever occurred first. Subgroup analyses were performed for the time after eradication treatment (grouped into 1–3, 3–5 and 5–7.5 years) and the number of received eradication treatments (1, 2 and ≥2). More than one eradication treatment indicated that H. pylori remained present after the previous treatment attempt, thus prolonging the duration of infection in the individual. The risk of cancer in individuals with peptic ulcer (gastric or duodenal) or gastritis (chronic or atrophic) was also assessed.
Poisson regression was used to analyse potential confounding by place of residence (urban or rural) and to assess any influence of the length of follow-up presented as incidence rate ratios (IRRs) and 95% CIs. Using the highest estimated incidence of gastric cancer in Sweden during the study period (17.3/100 000 individuals), we would need 121 621 person-years to verify a risk decrease of 55%8 with 80% power and an alpha level of 0.05. The statistical software STATA V.13.0 was used for all analyses.
Results
Participants
The study included 95 176 individuals receiving H. pylori eradication treatment during the study period (table 1). There were more men (53.7%) than women (46.3%) and the majority was younger than 60 years (60.1%). Most individuals resided in urban areas (74.6%). The mean length of follow-up was 3.7 years (maximum 7.5 years). During 351 018 person-years at risk, 75 (0.1%) individuals developed gastric adenocarcinoma and 69 (0.1%) developed non-cardia gastric adenocarcinoma. Out of 1235 individuals with duodenal ulcer, no one developed gastric adenocarcinoma, while 3 out of 732 (0.4%) with gastric ulcer developed gastric adenocarcinoma. Of the 1898 individuals with either chronic or atrophic gastritis, 2 developed gastric adenocarcinoma (0.1%).
Descriptive characteristics of the cohort of individuals receiving at least one eradication regimen for Helicobacter pylori in Sweden during 2005–2012
H. pylori eradication treatment and risk of all gastric adenocarcinoma
The SIRs decreased with longer time after eradication treatment. The SIRs were 8.65 (95% CI 6.37 to 11.46) for 1–3 years, 2.02 (95% CI 1.25 to 3.09) for 3–5 years and 0.31 (95% CI 0.11 to 0.67) for 5–7.5 years after eradication treatment (table 2). Poisson regression with 1–3 years after eradication treatment as the reference group showed IRRs of 0.05 (95% CI 0.03 to 0.09) 3–5 years and 0.00 (95% CI 0.00 to 0.01) 5–7.5 years after eradication treatment. Analysis from the date of last eradication treatment (per protocol) showed a similar trend, with decreased risks from ≥5 years after eradication treatment (data not shown). The risk of gastric adenocarcinoma increased with a higher number of eradication treatments, from SIR 1.88 (95% CI 1.44 to 2.41) in individuals with one eradication treatment to SIR 7.44 (95% CI 2.72 to 16.19) in those with more than two eradication treatments (table 2). Poisson regression showed no difference in gastric adenocarcinoma risk between individuals residing in rural or urban areas (IRR 0.98, 95% CI 0.59 to 1.61; P=0.93).
Risk of gastric and non-cardia gastric adenocarcinoma in individuals receiving Helicobacter pylori eradication treatment in Sweden from 2005 to 2012 compared with the Swedish standard population, expressed as SIR with 95% CI
Women had a higher risk of gastric adenocarcinoma 1–3 years after eradication treatment (SIR 11.69, 95% CI 7.49 to 17.40), compared with men (SIR 6.86, 95% CI 4.39 to 10.20), but the SIRs were similar after 3 years of follow-up (table 3). Younger age was associated with a higher risk of gastric adenocarcinoma, especially 1–3 years after eradication treatment (table 3).
Risk of gastric adenocarcinoma in individuals receiving Helicobacter pylori eradication treatment in Sweden from 2005 to 2012 compared with the Swedish standard population stratified by sex, age and calendar period, expressed as SIR with 95% CI
H. pylori eradication treatment and risk of non-cardia gastric adenocarcinoma
The SIRs of non-cardia gastric adenocarcinoma decreased with longer duration after eradication treatment, with SIRs of 10.74 (95% CI 7.77 to 14.46) 1–3 years, 2.67 (95% CI 1.63 to 4.13) 3–5 years and 0.43 (95% CI 0.16 to 0.93) 5–7.5 years after eradication treatment (table 2). Multiple eradication treatments increased the risk of non-cardia gastric adenocarcinoma, from SIR 2.38 (95% CI 1.80 to 3.10) in those with one eradication treatment to SIR 10.47 (95% CI 3.82 to 22.78) in those with more than two eradication treatments (table 2).
Eradication treatment for H. pylori did not influence the overall risk of cardia adenocarcinoma (SIR 0.60, 95% CI 0.22 to 1.30), but there were too few cases (six cases) to provide robust results regarding time latencies after eradication treatment of H. pylori and the risk of this tumour. Due to the low number of gastric adenocarcinoma cases in individuals with peptic ulcer or gastritis, further analyses were not deemed feasible.
Discussion
This study showed that the risk of gastric adenocarcinoma and non-cardia gastric adenocarcinoma declined over time after eradication treatment for H. pylori and was lower than that of the background population from 5 years after eradication treatment.
Methodological strengths of this study include the population-based design, large sample size and the high-quality data in the registers used. Additionally, studies on this topic have rarely been performed on a Western population, thus this study assesses a gap of knowledge. Also, non-cardia gastric adenocarcinoma was analysed separately, the gastric cancer type known to be associated with H. pylori, which was not always the case in earlier studies. The study also has limitations. The registers do not contain information on some potential confounders, for example, socioeconomic status. However, there was no effect of place of residence, which is correlated with socioeconomic status and lifestyle factors like smoking, diet and physical activity. Another limitation was the relatively short duration of follow-up. However, the maximum length of follow-up was still 7.5 years and the study had sufficient statistical power for robust analyses, including in the longest follow-up category. The fact that parts of the background population were infected with H. pylori would dilute the risk estimates, but not explain them, and this dilution should be limited because the population prevalence of H. pylori infection in Sweden is relatively low (15%).5 Furthermore, the strongly decreased risk for gastric cancer was found in the eradication treatment cohort, which is the group that should primarily be targeted (symptomatic individuals), considering the relatively low (3%) lifetime risk of developing gastric cancer in those with H. pylori. Additionally, the participants who received H. pylori eradication treatment were included in the background population, but because the study participants represented only a small proportion of the population (1.3%), this is unlikely to have had major influence on the results. If anything, it would again have diluted the associations and not explain them. Since our estimates are relative to the Swedish background population, they cannot be directly compared with estimates of studies comparing eradicated individuals with non-eradicated H. pylori-positive individuals. Another limitation is the lack of information about eradication treatments received before July 2005, making it unclear whether the first eradication date in the study was in fact the first eradication attempt. However, there is low antibiotic resistance in Sweden18 and only 9% of all cohort members received more than one eradication treatment during the study period, making this a minor issue. There was no confirmation of eradication success, meaning that H. pylori might still have been present after treatment. This should explain the higher risk among individuals receiving multiple eradication treatments compared with those with only one (more likely to be successful) eradication treatment. Detection bias could have been present if timing of gastric cancer diagnosis in eradicated individuals is different than in the background population. But since the majority of the background population is not infected with H. pylori and thus has a lower risk of gastric cancer, the decreased risk by time since eradication is more likely explained by the reasoning that once eradication treatment has prevented the gastric cancer development in the cohort, the background population has a higher risk of cancer, since there will be untreated H. pylori-infected individuals.
The clearly decreasing risk after eradication treatment over time is reassuring and in line with the results of meta-analyses from predominantly Asian populations.7 8Compared with previous cohort studies on this topic, where the largest studies included 3781 individuals receiving H. pylori eradication treatment,19 this study is considerably larger. A randomised clinical trial included 1130 individuals receiving eradication treatment with a maximum follow-up of 14.7 years, where an OR of 0.61 (95% CI 0.38 to 0.96) was found for gastric cancer compared with a placebo group (n=1128).20 However, that study did not analyse adenocarcinomas or non-cardia cancers separately. Of all previous randomised clinical trials or cohort studies, only one was conducted in a Western population, that is, in Finland with low prevalence of H. pylori and low risk of gastric cancer.21 The results of the present study are in line with the Finnish study, showing a strong decrease starting from the sixth year after eradication therapy. However, in this study, adenocarcinoma was not assessed separately.
An interesting finding in this study was that women had a higher SIR of gastric adenocarcinoma in the first 3 years after eradication compared with the SIR in men. Because these are relative risks, this finding might be explained by the higher risk of gastric adenocarcinoma in the male background population compared with the female background population.
This study provides evidence-based information about the preventive effect of H. pylori eradication on the risk of gastric adenocarcinoma, which might be especially relevant for healthcare providers in countries with a low incidence of gastric cancer to help decide whether to eradicate or not. Despite the favourable results of eradication for low prevalence settings, we believe there should be caution to implement widespread eradication, especially in asymptomatic individuals, since the long-term negative effects of eradication are insufficiently clear. However, more research is needed with regards to what would be the best point in time for eradication treatment, and if there is a so-called ‘point of no return’ in the developmental pathway to gastric cancer. Potential harmful effects of H. pylori eradication treatment need also to be clarified, and the inverse association between H. pylori infection and the risk of oesophageal adenocarcinoma should not be forgotten.
In conclusion, this population-based Swedish cohort study showed a sharp decrease in risk of gastric adenocarcinoma and non-cardia gastric adenocarcinoma after eradication treatment for H. pylori. The risk was below the level of the background population 5 years after the eradication treatment. Thus, eradication treatment for H. pylori seems to prevent the development of gastric and non-cardia gastric adenocarcinoma also in Western populations.
References
Footnotes
Contributors Design of the study: all authors; Data collection and preparation for analyses: NB and ED; Data analysis: ED with support from NB; Data interpretation: all authors; Writing of first draft: ED, revised and approved by all authors.
Funding Strategic Research Area (SFO), Karolinska Institutet internal funding for doctoral education (KID funding), Swedish Society of Medicine, Swedish Research Council and Swedish Cancer Society.
Competing interests None declared.
Ethics approval The study was approved by the Regional Ethical Review Board in Stockholm (2014/1291-31/4).
Provenance and peer review Not commissioned; externally peer reviewed.