Oral Antibiotics and Risk of New Colorectal Adenomas During Surveillance Follow-up

Antibiotic use may affect colorectal neoplasia risk by modifying the gut microbiome (1). A meta-analysis of 10 observational studies concluded oral antibiotic use was associated with 17% higher risk of colorectal cancer (2). This association may be stronger for broad-spectrum (2) and anti-aerobic (3) antibiotics, and for neoplasia of the proximal colon (3, 4). One study found early-life oral antibiotic use, but not recent use, was associated with higher risk of colorectal adenomas, the most common precursor lesions for colorectal cancer (4). It is unknown whether oral antibiotic use among those with an initial colorectal adenoma diagnosis is associated with subsequent risk of these polyps.

We aggregated individual-level data from the four Polyp Prevention Study trials (5–8). Conducted by a single consortium using similar protocols, each was a multicenter placebo-controlled randomized clinical trial among those recently diagnosed with colorectal adenomas. The Antioxidant trial investigated β-carotene and the combination of ascorbic acid and α-tocopherol among 864 participants enrolled between 1984 and 1988 (5). The Calcium trial investigated calcium among 930 participants enrolled between 1988 and 1992 (6). The Aspirin/Folate trial investigated aspirin at two doses and folic acid among 1,121 participants enrolled between 1994 and 1998 (7). The Vitamin D/Calcium trial investigated cholecalciferol and calcium among 2,259 participants enrolled between 2004 and 2008 (8). These studies were approved by Institutional Review Boards and all participants provided written informed consent.

In each trial, the primary endpoint was the occurrence of colorectal adenomas during the treatment period, planned to end at the time of a follow-up colonoscopy 3 to 5 years after randomization. A review of colorectal biopsies by a single study pathologist distinguished conventional adenomas from non-adenomatous polyps and adenocarcinomas. High-risk adenoma findings were defined as ≥1 advanced adenoma (tubulovillous/villous adenomas, ≥1 cm in diameter, with high-grade dysplasia, or adenocarcinoma) or ≥3 synchronous adenomas. Low-risk adenoma findings were defined as one or two non-advanced adenomas.

At baseline, the Antioxidant, Calcium, and Aspirin/Folate trials asked about current use of prescription drugs including antibiotics, whereas the Vitamin D/Calcium trial asked about use in the past year. Follow-up questionnaires administered every 6 months (4 months in the Aspirin/Folate trial) asked about use since the last questionnaire. Oral antibiotic types were classified by Anatomical Therapeutic Chemical code and grouped by class, narrow- or broad-spectrum, and anti-aerobic or anti-anaerobic.

Risk ratios (RR) with 95% confidence intervals (CI) adjusted for age, sex, clinical center, study, and treatment assignment (placebo vs. active) were estimated using log-linear Poisson regression for adenoma occurrence as a binary outcome with robust variance among those with ≥1 follow-up colonoscopy during the treatment period. Subgroup analyses were performed according to study, antibiotic class, and adenoma location. Adenomas diagnosed during interim colorectal exams before the scheduled end of the surveillance interval were included, unless <1 year after randomization. In sensitivity analyses, we excluded participants with interim exams. We defined antibiotic use as ≥2 separate reports of use at baseline or follow-up and evaluated total count of reports of antibiotic use as a continuous variable in a secondary analysis. Two-sided P < 0.05 was considered statistically significant.

Of 5,174 participants across the four trials, 763 (15%) participants used oral antibiotics at least twice (Table 1; Supplementary Table S1). Overall, 4,769 (92%) participants completed ≥1 follow-up colonoscopy. Oral antibiotic use was not associated with colorectal adenoma risk in any trial or the four trials combined (RR, 0.99; 95% CI, 0.90–1.10; Table 2; Supplementary Table S2). Results were similar by high- and low-risk status (Table 2), drug class (Supplementary Tables S3 and S4), anatomic location (Supplementary Table S5), and after excluding those with interim exams (Supplementary Table S6).

Oral antibiotic use during colonoscopic surveillance was not associated with colorectal adenoma risk in a secondary analysis of the four Polyp Prevention Study trials. Our results are inconsistent with a meta-analysis of 10 observational studies of invasive colorectal cancer (2), but consistent with an assessment of recent oral antibiotic use in a case–control study nested within the Nurses' Health Study of colorectal adenomas regardless of prior adenoma history (4).

Our study has limitations. The trials did not collect antibiotic dose or duration. History of antibiotic use before study enrollment was not ascertained, preventing us from assessing early-life use, previously found to be a risk factor for colorectal adenomas (4). Distinct reports of antibiotic use were separated by several months, and we could not distinguish continuous from discontinuous use. It was not possible to determine whether antibiotic use at baseline was related to symptoms leading to the initial adenoma diagnosis that qualified participants for these studies. Likewise, some participants may have been more likely to use antibiotics shortly before or after unscheduled follow-up exams that identified adenomas, but results were not meaningfully different when considering only those who completed the planned surveillance interval without any interim colorectal evaluations. In conclusion, these data suggest no association between oral antibiotic use, not otherwise specified, during surveillance and the risk of new colorectal adenomas.

M.N. Passarelli reports grants from NIH during the conduct of the study. E.L. Barry reports grants from NIH during the conduct of the study. J.R. Rees reports grants from NIH during the conduct of the study. J.A. Baron reports grants from NIH during the conduct of the study. No disclosures were reported by the other author.

M.N. Passarelli: Conceptualization, resources, data curation, formal analysis, funding acquisition, writing–original draft, writing–review and editing. L.A. Mott: Conceptualization, resources, data curation, software, formal analysis, project administration, writing–review and editing. E.L. Barry: Resources, data curation, investigation, writing–review and editing. J.R. Rees: Resources, data curation, investigation, writing–review and editing. J.A. Baron: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing.

On behalf of the Polyp Prevention Study Group, the authors express their appreciation to the study participants, investigators, and staff. The Polyp Prevention Study trials were funded by the NCI at the NIH (U01CA37287 and P30CA23108, to E.R. Greenberg; U01CA046927, R01CA059005, and R01CA098286, to J.A. Baron), and this analysis was funded in part by the National Institute of General Medical Sciences at the NIH (P20GM104416, to M.N. Passarelli).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.