Changes of stomach microenvironment and gastric cancer development

• Main Author: Song, Huan
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: Inst för medicinsk epidemiologi och biostatistik / Dept of Medical Epidemiology and Biostatistics 2015
• Subjects: Northern Sweden
• Online Access: http://hdl.handle.net/10616/44734

While acknowledging the importance of microenvironment-related events, e.g. Helicobacter pylori (H. pylori) infection and atrophy, occurring during the gastric cancer (GC) progression, at present we have inadequate data to say to what extent these conditions could affect the further development of GC and what we should do for patients with these changes. In this thesis, with different focuses for each project, we aim to confirm the observed increasing trend of atrophic corpus gastritis (ACG) among young middle-age Swedes, to detect true carcinogenic H. pylori strains, and to quantify excess GC risks among endoscopy or appendectomized patients. With all the data obtained from our projects, a more accurate risk stratification for patients with potential probability of GC development would be a reality. In Study I, based on Northern Sweden MONICA quinquennial population-based cross- sectional surveys conducted between 1990-2009, we randomly drew 5284 serum samples from all participants aged 35-64. Serologically defined ACG was determined by testing serum level of pepsinogen I (PG-I). The results revealed a surprising monotonic and significant upward trend in prevalence of PG-I-defined ACG (from 22 to 64 per 1000) in the youngest investigated age bracket 35-44 years. While in the oldest investigated age band (54- 65 years), there was a clear and expected decline. Our further exploration, through the cross- sectional case-control analysis, indicated that CagA seropositivity was the strongest predictor of ACG presence (odds ratio [OR] =2.29). Further significant risk factors included diabetes, low education level and high body mass index (BMI). Interaction analyses indicated that the association between BMI and ACG was confined to the age category 35-44 years, where overweight and obesity, respectively, were linked to 2.8-fold and 4.7-fold increased odds of having ACG. In Study II, a population-based case-control study, we measured antibodies against 17 H. pylori proteins using multiplex serology in sera from 268 cases and 222 frequency-matched controls. Unconditional multivariate logistic regression model was used for data analysis. In total, 15 proteins were significantly associated with increased GC risks, with the top three being CagA (OR=9.2), GroEL (6.6), HyuA (3.6). The excess risks were confined to non- cardia GC, but did not differ significantly between intestinal and diffuse subtypes. Using principal component analysis, we identified two significant factors among individuals without advanced chronic atrophic gastritis (AG): a CagA-dominant factor (antibodies against CagA, VacA, Omp as prominent markers), and a non-CagA factor (antibodies against NapA and Catalase as prominent markers). Strong dose-response relationships with non- cardia GC risks emerged for both antigen constellations: adjusted ORs (highest vs. lowest quartile) were 16.2 for the CagA-dominant factor, and 5.3 for the non-CagA factor. In Study III, based on a nationwide cohort consisting of all patients (n=405,211) with gastric biopsies on non-malignant indications between 1979 and 2011 in Sweden, we quantified that crude annual incidence rates of GC were 20×10-5 for the normal group, 42×10-5 for those with minor changes, 59×10-5 for the gastritis group, 100×10-5 for the AG group, 129×10-5 for the intestinal metaplasia (IM) group, and 263×10-5 for those with dysplasia. Compared to matching general Swedish population, the corresponding standardized incidence ratio (SIR) for overall GC was 1.0, 1.5, 1.8, 2.8, 3.4, and 6.5, respectively. The elevated GC incidence among precursor patients was stable throughout the follow-up period, and analysis restricted to the subset of participants with more than one biopsy suggested that the observed changes during re-evaluation(s) indeed have prognostic implications. Similarly, in Study IV, we estimated the relative risks of GC among appendectomized patients in another nationwide register-based cohort study, using Swedish general population as reference. In total, 480,382 eligible patients received appendectomy were followed during the period of 1970-2009 for GC occurrence. The results indicated that no excess GC risk (SIR=1.00, 95% CI 0.93-1.07) was noted after appendectomy. In conclusion, this thesis emphasizes that there is emerging evidence showing that GC, a fatal disease placing heavy burden on human health, might continue to be prevalent in the future. Given the large proportion of population affected by various stages of precancerous lesions, more accurate risk stratification is needed for optimizing the clinical management of these patients. H. pylori multiplex serology might be a useful tool for improving the overall cancer prediction capability. For patients biopsied on clinical indications, cost- effective endoscopy surveillance program needs to be developed for those with proven high-risk mucosal lesions (AG/IM/dysplasia).