Instructor Rutgers University New Brunswick, United States
Background: Helicobacter Pylori (HP) is a carcinogenic bacterium that causes peptic ulcers and gastric cancer (GC) but antibiotic treatment lowers GC risk in HP infected persons, starting 5-10 years after treatment. HP infection might increase colorectal cancer (CRC) risk. Thus, HP treatment may reduce CRC risk among HP infected persons
Objectives: To examine the association between HP treatment and CRC risk in the UK’s Clinical Practice Research Datalink Gold database with a nested case control study
Methods: Cases were ages 30-89 diagnosed with a first primary CRC (42491) between 2000-2020. Cancer-free controls (169720) were 4:1 individually matched to cases on age, sex, clinical practice, year of diagnosis and length of practice registration. UK HP treatment guidelines [e.g., PPI 2X/day + 1g amoxicillin + (.5g clarithromycin OR .4g metronidazole) for 7 days] were used to identify all HP treatment events occurring >5 years prior to case or control selection date. We used positive [GC N=5510] and negative [breast cancer N=71536/prostate cancer N=55088] nested case control studies to identify optimal lag periods/help interpret results. Control populations for each of these studies were selected identically to the CRC study. We estimated associations [odds ratios (ORs) and 95% confidence intervals (CIs)] between 1st observed HP treatment 5-10 (OR1), 10-15 (OR2), and >15 (OR3) years prior to selection and cancer compared to no treatment using conditional logistic regression. Exposures occurring < 5 years prior to selection were ignored to reduce reverse causality. Linear p-trend values were calculated.
Results: CRC study participants were 56% male and registered in their practice for a median of 19.5 years (IQR: 11,33) prior to selection. HP treatment was noted in 2.3% and 2.2% of cases and controls, respectively. We saw an inverse trend over time in the associations between first HP treatment and CRC risk [OR1=1.1, 0.9-1.2; OR2= 1.0, 0.8-1.2; OR3= 0.9, 0.8-1.2; p trend = 0.04]. The GC associations were [OR1=1.4, 1.0-.9; OR2= 1.0, 0.6-1.5; OR3= 0.6, 0.3-1.3; p trend < 0.001]. There was no association between HP treatment and breast or prostate cancers.
Conclusions: Similar patterns in the CRC and GC studies may reflect common underlying mechanisms (HP treatment reduces cancer risk through HP eradication). This preliminary study examined the effect of HP treatment on CRC risk in the full population, not just HP+ persons. HP prevalence in the UK is ~35% which is much larger than the proportion of treated individuals we saw. Future work should identify HP+, but untreated individuals, to estimate HP treatment effects on CRC in HP+ persons. Positive and negative control studies are useful in contextualizing novel studies that lack important information (e.g., HP status).
.jpg "Monica D'Arcy, PhD (she/her/hers) photo")
