Risk of gastric cancer after Helicobacter pylori eradication in diabetes mellitus patients: a territory-wide study with propensity score analysis

Abstract

Introduction: Whether diabetes mellitus (DM) increases gastric cancer (GC) risk remains controversial in prior studies due to inadequate adjustment for important risk factors including Helicobacter pylori(HP) infection, glycemic control, concomitant medication usage and cancer sites. Aims & Methods: We aimed to investigate whether type II diabetes mellitus (DM) increased GC risk in patients after HP treatment. This was a territory-wide cohort study of patients aged ≥45 years who had received clarithromycin-based triple therapy for HP between 2003 and 2012. Data were retrieved from the public electronic health database. Observation started from HP therapy to GC diagnosis, death or end of study (December 2015). Exclusion criteria included type I DM, GC diagnosed within first year of HP therapy, prior GC or gastrectomy, and failure of HP eradication. The-adjusted hazard ratio (aHR) of GCwith DM was calculated by Cox model with propensity score regression adjustmentfor 20 covariates (age, sex, smoking, alcoholism, past history of gastric and duodenal ulcers, other comorbidities [atrial fibrillation, ischemic heart disease, congestive heart failure, chronic renal failure, cirrhosis, stroke, hypertension and obesity] and medications [aspirin, non-steroidal anti-inflammatory drugs (NSAIDs), cyclooxygenase-2 (COX-2) inhibitors, and proton pump inhibitors (PPIs)]). Results: Of 46,460 eligible patients, 6,900 (14.9%) had DM. During a median follow-up of 7.1 years (IQR:4.8-9.8) with 337,313 person-years, 153 (0.33%) developed GC at a median age of 72.4 years (IQR:63.8- 82.6). There were 31 (20.3%) cardia cancers and 88 (57.5%) non-cardia cancers, while the remaining 34 (22.2%) cases did not have site specified. DM was associated with an increased GC risk (adjusted HR:1.67; 95% CI:1.08-2.58). This association was biased towards null if concomitant medication usage was not adjusted (adjusted HR:1.30; 95% CI:0.85-1.99), with the most influential effect from statins (Table). On the other hand, HR increased to 1.92 (95% CI:1.28-2.90) without adjusting for comorbidities. Stratified analysis shows the risk was increased for cardia cancer only (aHR:3.40, 95% CI:1.45-7.97), in those with suboptimal DM control (time-weighted average HbA1c ≥6.0%; aHR:1.68, 95% CI: 1.07-2.63) and metformin non-users (aHR 2.59, 95% CI 1.41-4.74). Conclusion: Type II DM was associated with an increased GC risk among HP-eradicated patients, in particular cardia GC and those with suboptimal DM control. Inadequate adjustment for concomitant medications and co-morbidities could potentially bias the results in previous studies.