Eradication of Helicobacter pylori and Gastric Cancer: A Systematic Review and Meta-analysis of Cohort Studies

Abstract

Background:Helicobacter pylori ( H. pylori ) is associated with an increased risk of gastric adenocarcinoma and gastric mucosa associated lymphoid tissue (MALT) lymphoma and a decreased risk of esophageal adenocarcinoma. We aimed to assess how eradication therapy for H. pylori influences the risk of developing these cancers.

Methods: This was a systematic review and meta-analysis. We searched PubMed, Web of Science, Embase, and the Cochrane Library and selected articles that examined the risk of gastric cancer, MALT lymphoma, or esophageal cancer following eradication therapy, compared with a noneradicated control group.

Results: Among 3629 articles that were considered, nine met the inclusion criteria. Of these, eight cohort studies assessed gastric cancer while one randomized trial assessed esophageal cancer. Out of 12 899 successfully eradicated patients, 119 (0.9%) developed gastric cancer, compared with 208 (1.1%) out of 18 654 noneradicated patients. The pooled relative risk of gastric cancer in all eight studies was 0.46 (95% confidence interval [CI] = 0.32 to 0.66, I ² = 32.3%) favoring eradication therapy. The four studies adjusting for time of follow-up and confounders showed a relative risk of 0.46 (95% CI = 0.29 to 0.72, I ² = 44.4%).

Conclusions: This systematic review and meta-analysis indicates that eradication therapy for H. pylori prevents gastric cancer. There was insufficient literature for meta-analysis of MALT lymphoma or esophageal cancer.

Helicobacter pylori is a bacterium estimated to be present in the stomachs of half of the adult human population and is typically acquired during childhood ( 1 , 2 ). H. pylori is associated with chronic gastritis, peptic ulcer, and gastric cancer, particularly in the presence of its virulence factor CagA ( 3 ). The incidence of gastric cancer is declining, possibly because of the decrease in H. pylori prevalence ( 4 ), but is still the fifth most common cancer and the third most common cause of cancer death globally ( 5 ). The overall yearly incidence rates globally are 15.6 to 18.1 and 6.7 to 7.8 per 100 000 individuals in men and women, respectively ( 6 ). There are considerable variations in incidence geographically, and half of all cases occur in Eastern Asia ( 5 ). Gastric cancer is typically adenocarcinoma (95%), which is subclassified into intestinal or diffuse type, both associated with H. pylori ( 7–9 ), and mixed type. The intestinal type develops from a gastric mucosa with chronic gastritis, atrophy, intestinal metaplasia, and dysplasia to invasive adenocarcinoma ( 10 ). For the diffuse type, the carcinogenic pathway is less clear. The remaining adenocarcinomas are of the mixed type. Gastric mucosa–associated lymphoid tissue (MALT) lymphomas are also associated with H. pylori ( 11 ). Interestingly, H. pylori seemingly decreases the risk of esophageal adenocarcinoma ( 12 ). A possible explanation is that H. pylori –related atrophic gastritis reduces gastric acid secretion, which in turn counteracts gastro-esophageal reflux, a main risk factor for this cancer ( 12 ). The increasing incidence of esophageal adenocarcinoma might be because of the decreasing prevalence of H. pylori ( 13 ). CagA-positive H. pylori strains might also be associated with an increased risk of esophageal squamous cell carcinoma ( 14 ). In theory, H. pylori eradication should reduce the risk of gastric adenocarcinoma, MALT lymphoma, and esophageal squamous cell carcinoma and increase the risk of esophageal adenocarcinoma. However, the available literature is limited. Two previous meta-analyses have indicated a decreased risk of gastric adenocarcinoma while no meta-analysis has examined MALT lymphoma or esophageal cancer ( 15 , 16 ). We aimed to examine the role of H. pylori eradication therapy in the development of gastric adenocarcinoma, MALT lymphoma, and esophageal cancer in a systematic review and meta-analysis.

Methods

This was a systematic review and meta-analysis analysing the risk of gastric adenocarcinoma, MALT lymphoma, and esophageal cancer after eradication therapy for H. pylori . The study was performed according to an a priori established study protocol.

Exposure and Outcome

The study exposure was H. pylori eradication therapy. All regimens intended to eradicate H. pylori were considered eligible. H. pylori– positive individuals who did not receive eradication and individuals for whom eradication was unsuccessful were considered unexposed (noneradicated). The outcomes were gastric or esophageal malignancy, where the following histological subtypes were considered for inclusion: adenocarcinoma, squamous cell carcinoma, or MALT lymphoma ( 17 , 18 ).

Search Strategy and Study Selection

We conducted a systematic search of the medical literature using PubMed, Web of Science, Embase, and the Cochrane Library up until November 2015. There were no restrictions regarding article language or date of publication. Articles were considered eligible if the risk of cancer was evaluated in individuals receiving eradication therapy, compared with a control group that did not receive eradication or in which eradication was unsuccessful. Only studies describing a population representative of the general population or representative of individuals receiving H. pylori screening and treatment in clinical practice were considered eligible. The following search terms were used: ‘ Helicobacter pylori ,’ ‘ campylobacter ,’ ‘eradication,’ ‘chemoprevention,’ ‘esophageal neoplasms,’ ‘esophageal,’ ‘stomach neoplasms,’ ‘gastric,’ ‘cancer,’ ‘carcinoma,’ ‘tumour,’ ‘adenocarcinoma,’ ‘malignancy’ (taking into account both British and American spelling), and the names of different medications used for H. pylori eradication. A detailed search description can be found in the Supplementary Materials ( Supplementary Table 1 , available online). The results of the search were first evaluated based on article titles by one researcher (ED). The next step of the search was performed by two independent researchers (ED and NB), who evaluated the abstracts and full texts of the remaining articles. Any disagreement between the researchers was solved by consensus. Initial exclusion criteria were animal studies, studies without original data (including commentaries and editorials), meeting abstracts, case reports, and case-control studies. The latter were excluded to maintain comparability of measures of effect. Furthermore, we applied backwards and forward citation tracking (sources cited in included articles and identifying articles that cited the included articles) to all included articles to identify other possible relevant studies. For studies that reported cancer development after eradication therapy for H. pylori but did not report the number of cancer cases in the control group, we contacted the authors to provide these data to be able to include these articles. When we found multiple articles based on the same study population, we included only the most recent article (with the longest follow-up).

Data Extraction and Quality Assessment

Data extraction on cancer development was performed independently by two researchers (ED and NB), who extracted both unadjusted data (ie, absolute numbers) and adjusted data while taking into account the follow-up time (from Cox or Poisson models). The following data were extracted for each study: geographical location, method to detect H. pylori , H. pylori eradication regimen, success of eradication, age (mean and range), sex ratio, follow-up time, and the histological type of the cancer.

Two researchers (ED and NB) independently assessed the quality of the studies according to the Newcastle Ottawa scale ( 19 ). The quality of each study was assessed by the following items: representativeness of the exposed cohort, selection of the nonexposed cohort, ascertainment of the exposure, demonstration that the outcome of interest was not present at the onset of the study, comparability of study cohorts on the basis of the design or analysis, assessment of outcome, length and adequacy of the follow-up of the cohorts. Studies with a score of 3 or less were considered of low quality, 4 and 5 were considered of moderate quality, and a score of 6 or higher defined good quality.

Data Synthesis and Statistical Analysis

For the pooled analyses, we used a random effect model to take heterogeneity within and between studies into account, therefore estimating the average treatment effect and its precision. H. pylori eradication was compared with no eradication or unsuccessful eradication and expressed as unadjusted risk ratios (RRs) with 95% confidence intervals (CIs) for the development of cancer. A second analysis was used to pool the adjusted risk ratios and 95% confidence intervals, taking into account follow-up time, in which the adjusted (full model) reported hazard or incidence ratios that were used. To assess heterogeneity, we used the I ² statistic, where an I ² of greater than 50% was used to define a substantial degree of heterogeneity ( 20 ). The statistical significance of heterogeneity was assessed with a Cochran’s Q test, with values smaller than 0.10 representing substantial heterogeneity. Subgroup analyses were performed for baseline gastric histology, and type of control group to investigate possible explanations for any heterogeneity. A funnel plot and Egger’s test were conducted, where a large P value (>.05) indicated no evidence of small study effects or publication bias. All statistical tests were two-sided. The statistical analyses were performed using the statistical software Stata (Stata Corporation, version 12.1).

Results

Study Selection

The systematic search identified 3629 unique articles ( Figure 1 ). After screening of the titles and abstracts, the full texts of 38 articles remained for further evaluation of eligibility. None of these studies assessed the risk of MALT lymphoma, and only one study assessed the risk of esophageal cancer ( 21 ). Four studies were excluded because of the design; one was a case-control study ( 22 ), and three others were comments without original data ( 23–25 ). Three studies were excluded because they were reviews or gave notice of an ongoing study and thus contained no new data ( 26–28 ). Other studies were excluded for various reasons as shown in Figure 1 ( 29–32 ). Nine studies were randomized controlled trials (RCTs) already described in the most recent meta-analysis ( 21 , 33–40 ), and no new RCTs were identified after that meta-analysis ( 15 ). After this evaluation of the literature, we restricted our study to cohort studies when examining gastric cancer. Seven other articles were excluded because no suitable control group was described ( 41–45 ) or the risk of cancer was not reported in the control group; we did not retrieve this information after contacting the authors ( 46 , 47 ). Four other articles described the cancer risk based on the same population ( 48–51 ), so only the most recent article with the longest follow-up was included ( 51 ). After this review, eight observational studies assessing the risk of gastric cancer were selected for final analysis ( 51–58 ). One study, a randomized clinical trial assessing the risk of esophageal cancer, was included ( 21 ).

Figure 1.

Flow diagram of study selection adapted from Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for a meta-analysis assessing gastric cancer risk after Helicobacter pylori eradication.

Study Characteristics

Patient characteristics for the eight studies examining gastric cancer are presented in Table 1 . Seven of the studies were conducted in Japan and one in Finland. Five studies compared a successful with an unsuccessful eradication attempt ( 51–55 ), and three studies compared eradicated individuals with untreated individuals with H. pylori ( 56–58 ). The age range was 17 to 83 years ( 51 ), with mean ages between 50 ( 51 ) and 63 years ( 57 ). Most studies included more men than women ( Table 1 ), particularly in one of the studies on factory workers (98% men) ( 56 ). Three studies described the baseline gastric histopathology of the patients while the other five did not. In one study, all patients (100%) had gastric atrophy ( 51 ), and in another study only 5% to 7% had atrophic gastritis ( 53 ). One study included patients with intestinal metaplasia and a large (undefined) proportion of participants who presented with atrophy at baseline ( 58 ). None of the studies included patients with dysplasia or early gastric cancer at baseline. Table 2 describes the eradication therapies used. Success of eradication was measured in five out of eight studies, and the success rate varied between 65% ( 55 ) and 87% ( 56 ). Gastric cancer development and follow-up of patients is described in Table 3 . The mean follow-up time ranged from 2.9 years ( 53 ) to 10.9 years ( 52 ). The majority (up to 89%) of the cancers in both treatment and control groups were intestinal-type gastric adenocarcinomas. In the treatment group in one study ( 51 ) and the control group of another study ( 55 ), there were similar proportions of diffuse and intestinal-type adenocarcinomas.

Table 1.

Patient characteristics of studies included in a meta-analysis of Helicobacter pylori eradication and risk of gastric cancer *

Study	Males receiving eradication, %	Male control subjects, %	Age range among all included individuals, y	Mean age for those receiving eradication, y	Mean age among control subjects, y	Baseline histology of the gastric mucosa
Uemura, 2001	57	NA	20-76	52	NA	53% moderate atrophy, 17% severe atrophy
						37% intestinal metaplasia
Kato, 2006	66	57	NA	55	53	Not reported
Takenaka, 2007	71	NA	NA	54	NA	5-7% atrophic gastritis
Ogura, 2008	57	55	NA	62	63	Not reported
Mabe, 2009	71	71	NA	52	58	Not reported
Yanaoka, 2009	98	NA	40-60	NA	NA	Not reported
Kosunen, 2011	44	44	NA	56	55	Not reported
Take, 2014	89	NA	17-83	50	NA	100% atrophy (mild to severe)

*NA = not available.

Table 2.

Helicobacter pylori eradication strategy in studies included in a meta-analysis of eradication and risk of gastric cancer *

Study	Eradication regimen	Method to detect H. pylori	Successful eradication, %	Control group
Uemura, 2001	Not available	Rapid urease test, histology, serology	Not reported	No eradication
Kato, 2006	PPI triple therapy (lansoprazole 60 mg, amoxicillin 1500 mg, and clarithromycin 400 mg)	Urea breath test or biopsy-based methods	Not reported	Unsuccessful eradication
	2-wk combination therapy with PPI and antibiotics
Takenaka, 2007	Triple or dual therapy: amoxicillin 750 mg, lansoprazole 30 mg, or omeprazole 20 mg and (clarithromycin 200/400 mg for triple therapy)	Urea breath test, rapid urease test, histology, culture, serology	85%	Unsuccessful eradication
Ogura, 2008	7 d: lansoprazole 30 mg, amoxicillin 750 or 1000 mg, and clarithromycin 400 mg or metronidazole 250 mg, twice daily	Urea breath test, rapid urease test, histology, culture, serology	74%	No eradication
Mabe, 2009	7 d: 30 mg lansoprazole or 20 mg omeprazole; 750 mg amoxicillin and 200 or 400 mg clarithromycin, twice daily	Rapid urease test	65% (ITT), 79% (PP)	Unsuccessful eradication
Yanaoka, 2009	2 wks: omeprazole 20 mg and amoxicillin 750 or 500 mg, twice daily	Serology	87%	No eradication
	1 wk: omeprazole 20 mg; amoxicillin 750 mg and clarithromycin 200 mg, twice daily
Kosunen, 2011	Not available	Serology	Not reported	Unsuccessful eradication
Take, 2014	2 wks: amoxicillin 750 mg and omeprazole 20 mg or lansoprazole 30 mg	Rapid urease test, culture	85%	Unsuccessful eradication
	1 wk: amoxicillin 750 mg; clarithromycin 200 mg or 400 mg and omeprazole 20 mg or lansoprazole 30 mg or rabeprazole 10 mg
	1 wk: metronidazole 500 mg; amoxicillin 750 mg or clarithromycin 200 mg and omeprazole 20 mg or lansoprazole 30 mg or rabeprazole 10 mg

*ITT = intention to treat; PP = per protocol; PPI = proton pump inhibitor.

Table 3.

Gastric cancer development following Helicobacter pylori eradication in a meta-analysis

Study (ref)	Country	Participants, treatment group, No.	Included in analysis, treatment * , No.	Cases of gastric cancer, treatment, No., %	Type of adenocarcinoma (intestinal/diffuse), treatment, No.	Participants, control group, No.	Included in analysis, control, No.	Cases of gastric cancer, control, %	Type of adenocarcinoma (intestinal/diffuse), control, No.	Treatment mean follow-up (range), y	Control mean follow-up (range), y
Uemura, 2001 ( 58 )	Japan	253	253	0 (0)	0	993	993	36 (3.6)	23/13	4.8	8.5
Kato, 2006 ( 54 )	Japan	1788	1788	23 (1.3)	19/4	1233	1233	44 (3.6)	32/12	5.9	7.7
Takenaka, 2007 ( 53 )	Japan	2192	1519	6 (0.4)	4/2	373	288	5 (1.7)	4/1	3.25 (max. 9.8)	2.9
Ogura, 2008 ( 57 )	Japan	853	404	6 (1.5)	NA *	623	304	13 (4.3)	NA	3.2	3.1
Mabe, 2009 ( 55 )	Japan	3781	3781	47 (1.2)	35/10 †	352	352	9 (2.6)	5/4	5.6 (max. 8.0)	5.2 (max. 8.4)
Yanaoka, 2009 ( 56 )	Japan	852	474	5 (1.1)	4/1	4924	3664	55 (1.5)	36/19	Max. 10	Max. 10
Kosunen, 2011( 52 )	Finland	3650	3650	11 (0.3)	NA	11 628	11 628	37 (0.3)	NA	10.4 (max. 20)	9.3 (max. 20)
Take, 2014 ( 51 )	Japan	1135	1030	21 (2.0)	11/10	207	192	9 (4.7)	8/1	9.9 (max 17.4)	9.9 (max. 17.4)

*Number of patients included in the analysis after excluding those lost to follow-up. NA = not available.

†Two cancers of unknown type.

Quality Assessment of Included Studies

Of the eight studies, seven were considered of good quality ( 51–56 , 58 ) and one of moderate quality ( 57 ). All studies included a control group from the same community as the exposed group and had a good ascertainment of the outcome. All but one study had a good assessment of the exposure, yet in one study it was not clear if eradication was successful in all patients because of incomplete data ( 52 ). In two studies, the eradicated group was a selected group of individuals (ie, factory workers) ( 51 , 56 ). One study could not demonstrate that gastric cancer was not present at start ( 52 ). Two studies had a rather short follow-up time, with a mean around three years ( 53 , 57 ). One study had a limited completeness of follow-up (56%) and no description of the individuals lost to follow-up ( 57 ), and one study had no statement about follow-up of the cohort members ( 54 ). Four studies reported risk estimates adjusted for time for follow-up and confounding ( 51 , 53 , 55 , 57 ). The confounders adjusted for were not specified in one study ( 57 ) while the other three studies adjusted for age and sex. One study further adjusted for alcohol use, tobacco smoking, and gastric mucosal atrophy ( 51 ), another adjusted for the indication for H. pylori eradication ( 53 ), and the last one also adjusted for location of peptic ulcer, salt consumption, and tobacco smoking ( 55 ).

Helicobacter pylori Eradication Therapy and Gastric Cancer Risk

Among 12 899 patients who were successfully eradicated, 119 (0.9%) developed gastric cancer during follow-up while such cancer was found in 208 out of 18 654 (1.1%) in the noneradicated group. The pooled analysis of all eight included studies provided a risk ratio of 0.46 (95% CI = 0.32 to 0.66) in favor of eradication therapy ( Figure 2 ). The heterogeneity was low (I ² = 32.3%), and there was no evidence of small study effects or publication bias ( P = .33). When the Finnish study was excluded, because of the low incidence of gastric cancer in Finland compared with East-Asian countries, the seven Japanese studies provided a pooled risk ratio of 0.40 (95% CI = 0.29 to 0.54, I ² = 0.0%). Another sensitivity analysis excluding the study where a proportion of the participants had intestinal metaplasia at baseline yielded a risk ratio of 0.48 (95% CI = 0.34 to 0.66, I ² = 24.2%). The five studies comparing successful vs unsuccessful eradication showed a risk ratio of 0.47 (95% CI = 0.31 to 0.71, I ² = 40.9%), and the three studies comparing eradicated with noneradicated individuals showed a risk ratio of 0.39 (95% CI = 0.14-1.08, I ² = 49.9%) ( Figure 2 ). The five studies not reporting baseline gastric histopathology (atrophy, intestinal metaplasia, or dysplasia) showed a risk ratio of 0.52 (95% CI = 0.35 to 0.77, I ² = 35.7%), and the three studies including patients with aberrant baseline histology (gastric atrophy or intestinal metaplasia) showed a risk ratio of 0.28 (95% CI = 0.11 to 0.72, I ² = 40.9%) (forest plot not shown). Four studies reported risk ratios adjusted for follow-up time and confounding, of which three were analyzed using Cox regression and one using Poisson regression, resulting in a pooled adjusted risk ratio of 0.46 (95% CI = 0.29 to 0.72, I ² = 44.4%), 0.33 (95% CI = 0.19 to 0.59) when including only the three studies using Cox regression (I ² = 0%), indicating a low heterogeneity between these three studies ( Figure 3 ).

Figure 2.

Forest plot of studies comparing eradication therapy to no treatment and studies comparing successful to unsuccessful treatment. A random effects model (DerSimonian Laird method) was used. The I ² was used to assess heterogeneity. The two-sided Cochran’s Q test was used to assess the significance of this heterogeneity. The diamond represents the observed effect size for the studies combined, the squares represent the weight given to the study, and the error bars represent the corresponding confidence intervals. CI = confidence interval; RR = relative risk.

Figure 3.

Forest plot of studies comparing eradicated and noteradicated groups that reported adjusted values, grouped by Cox or Poisson model. A random effects model (DerSimonian Laird method) was used. The I ² was used to assess heterogeneity. The two-sided Cochran’s Q test was used to assess the significance of this heterogeneity. The diamond represents the observed effect size for the studies combined, the squares represent the weight given to the study, and the error bars represent the corresponding confidence intervals. CI = confidence interval; NA = not available; RR = relative risk.

Helicobacter pylori Eradication Therapy and esophageal Cancer Risk

The only study assessing the risk of esophageal cancer was performed in China and compared eradication with placebo treatment ( 21 ). The mean age was 42 years (range = 35–65), and the proportion of men to women was similar in both groups. The eradication rate was 84%. Two out of 817 (0.2%) individuals who received eradication therapy developed esophageal cancer, compared with one out of 813 (0.1%) individuals who received placebo; all three cancers were squamous cell carcinomas ( 21 ).

Discussion

This meta-analysis indicates that eradication therapy for H. pylori prevents gastric adenocarcinoma. The literature is insufficient to allow analysis of the risk of MALT lymphoma or esophageal cancer following eradication in a meta-analysis.

The main strength of this meta-analysis is that by including only cohort studies we were able to conduct an objective analysis on a large number of participants over a long follow-up period. The eight included studies are based on cohorts close to clinical practice without obvious selection issues. The fact that the studies presenting both unadjusted and adjusted results showed similar effect sizes indicates lack of strong confounding by the factors adjusted for. Also, this study was based on an a priori established study protocol and a thorough systematic search of the literature.

Limitations of meta-analyses in general are that the validity is dependent on the quality of the included studies, on heterogeneity between studies, and on possible publication bias. Also, by using cohort studies there is a higher risk of uncontrolled bias, mainly confounding. However, it is important to assess the effect from studies that closer resemble clinical practice, and the quality was considered good in seven out of eight studies and moderate in one; the statistical heterogeneity was low to moderate, and there was no evidence of small-study effects bias (publication bias). Moreover, the results were consistent in various sensitivity analyses. Analysis of studies comparing eradicated individuals with individuals not having undergone any attempt to treat the bacteria and studies comparing eradicated individuals with unsuccessfully eradicated individuals showed similar results. However, four out of eight studies did not adjust for confounders, and the four studies that did adjusted for different confounders. Because of this, confounding might still influence the results, limiting comparability. However, we assessed the quality of the included studies, and almost all were of good quality. Furthermore, a quality assessment tool is very useful to compare included studies in an objective way. However, this tool cannot take other items into account, for example, the extent of missing data in the individual studies. Missing data items in this study mostly concern the reporting of age. This was most serious in one study, where only the age range for all participants was reported and no mean ages for both treatment and control group were reported ( 56 ). The detection of H. pylori and consequent evaluation of success of eradication may have influenced the results, especially in the Finnish study, where success of eradication was less clearly defined (a large proportion of patients did not have information regarding cure) ( 52 ). Outcome was measured with endoscopy in seven studies and extracted from a cancer registry in one study ( 52 ). Of the seven studies assessing the outcome endoscopically, three reported doing this on an annual basis ( 51 , 53 , 55 ). There was no additional information in the other four studies. This makes it difficult to compare these studies. In three studies, patients presented with atrophy or intestinal metaplasia at baseline ( 51 , 53 , 58 ). There were two studies performed on a selected group, namely male factory workers, who are likely to be healthier than the general population (“healthy worker effect”) ( 59 ). A limitation was that almost all studies were conducted in Japan, an area with a high incidence of gastric cancer, which questions whether the results are generalizable to non-Asian populations.

The finding that H. pylori eradication therapy prevents gastric cancer is in line with the findings of two previous meta-analyses based on randomized clinical trials ( 15 , 16 ). The relative risk of gastric cancer after eradication therapy compared with placebo or no treatment was 0.66 (95% CI = 0.46 to 0.95) in the most recent meta-analysis, and thus our meta-analysis yielded a stronger preventive effect. The earlier meta-analysis was performed on the same randomized clinical trials but included data from one randomized clinical trial twice ( 16 , 60 ). Compared with the most recent meta-analysis, which investigated the risk of gastric cancer after H. pylori eradication in healthy asymptomatic individuals in six individual randomized clinical trials with follow-up ranging from maximum five to 14.7 years ( 15 ), the present meta-analysis included studies with a longer follow-up (up to 20 years) ( 52 ) and almost four times more treated individuals and more than five times more controls. Most of the randomized clinical trials included in the previous meta-analyses were conducted in China, leading to a limited generalizability of the results (similar to the present study). Also, the results for these studies are relevant mainly for otherwise healthy individuals whereas the result of our study would be more applicable to the general patient population. Furthermore, cohort studies better resemble clinical practice and reduce selection of participants compared with randomized clinical trials.

There is a debate whether H. pylori eradication prevents gastric cancer. There are suggestions that once the histology has reached the level of intestinal metaplasia in the gastric adenocarcinoma development pathway eradication therapy may no longer have cancer-preventive effects ( 61 ). However, it is likely that H. pylori eradication lowers the risk of gastric cancer compared with no treatment ( 62 ), which is supported by the findings of the present meta-analysis and the two previous meta-analyses. H. pylori is more strongly associated with the risk of intestinal type of gastric adenocarcinoma than with the diffuse type ( 63 ). In one study, it was suggested that H. pylori eradication would prevent only the intestinal type because the diffuse type developed only in individuals cured from H. pylori infection ( 48 ). In the present study, the results indicate a similarly strong preventive role of eradication for both histological types of gastric adenocarcinoma. Because the included studies did not provide information on the location of gastric cancer (cardia or noncardia), no conclusion can be drawn of the effect of eradication for different anatomical locations of gastric cancer.

To conclude, this systematic review and meta-analysis of eight cohort studies and 31 553 patients indicates that eradication therapy for H. pylori prevents gastric cancer. Research examining H. pylori eradication in relation to the risk of gastric MALT lymphoma and esophageal cancer is currently too limited to enable meta-analyses. More studies assessing the risk of esophageal cancer and MALT lymphoma after H. pylori eradication are needed. Because both types of cancers are rare compared with gastric cancer, RCTs may not have enough power. Thus, population-based studies may be preferred.

Funding

Funding was received from Swedish Research Council (SIMSAM), Swedish Cancer Society, and Strategic Research Area (SFO).

Note

The study sponsors did not have any role in any part of the study.