Differential Effects of Antibiotics On Colon and Rectal Cancer Risk After Baseline Colonoscopy Negative For Cancer: A Territory-Wide Study of 97,162 Patients

Abstract

Aims & Methods: We aimed to investigate the effects of different antibiotics on CRC development in older subjects. We conducted a retrospective cohort study based on a territory-wide healthcare database of 7.4 million population. All patients aged at least 60 years who had undergone colonoscopy between 2005 and 2013 with no CRC found were recruited. We excluded cases with inflammatory bowel disease, prior colectomy, prior CRC and CRC detected < 6 months of index colonoscopy. Exposure of interest was any antibiotic use within 5 years before colonoscopy (including penicillins, cephalosporins, macrolides, carbapenems, quinolones, tetracyclines, aminoglycosides, nitroimidazoles, glycopeptides, sulpha/ trimethoprim, and others). Outcomes of interest were PCCRC-all (all postcolonoscopy CRC diagnosed >6 months after index colonoscopy), PCCRC-3y (between 6 and 36 months after index colonoscopy), and PCCRC>3y (>36 months after index colonoscopy). Covariates included patient's factors, history of colonic polyps/polypectomy, concurrent drug usage (aspirin, NSAIDs, COX-2 inhibitors and statins) and endoscopy center's performance (polypectomy rate and colonoscopy volume). The adjusted hazard ratio (aHR) of PCCRC was calculated by multivariable Cox proportional hazards model. Stratified analysis was performed according to cancer subsites (proximal [cecum to splenic flexure], distal and rectal cancer) and nature of antibiotics (anti-anaerobic vs anti-aerobic, broad- vs narrowspectrum, and oral vs intravenous route). Results: Of 97,162 eligible subjects, 58,704 (60.4%) had antibiotic use before index colonoscopy. There were 1,026 (1.0%) PCCRC-all cases (694 [67.6%] PCCRC-3y and 332 [32.4%] PCCRC>3y cases). Antibiotic use was associated with lower rectal cancer (aHR:0.64; 95% CI:0.54-0.76) but higher proximal cancer risk (aHR:1.62; 95% CI:1.14-2.30). There was no statis-tically significant association between antibiotic use and distal cancer (aHR:0.99; 95% CI:0.76-1.30). Stratified analysis showed similar results for PCCRC-3y but non-significant results for PCCRC>3y. Penicillins were associated with lower PCCRC-all risk (aHR:0.83; 95% CI:0.72-0.95), while amino-glycosides were associated with higher risk (aHR:1.54; 95% CI:1.05-2.26). Anti-anaerobic (aHR:1.67; 95%CI:1.17-2.39), narrow-spectrum (aHR:2.06; 95% CI:1.01-4.21) and intravenous antibiotics (aHR:2.59; 95% CI:1.36-4.92) were associated with higher proximal cancer risk, while broad-spectrum (aHR:0.63; 95% CI:0.53-0.75) and oral antibiotics (aHR:0.65; 95% CI:0.54-0.78) were associated with lower rectal cancer risk. Conclusion: The effects of antibiotics on CRC in older patients varied according to cancer subsite, classes of antibiotics and route of administration. Further studies are necessary to elucidate the potential role of different antibiotics and gut microbiota on CRC development.