Chronic Antibiotic Usage and Obesity in the Sister Study

Abstract

Chronic Antibiotic Usage and Obesity in the Sister StudyAbstract Number:2431 Melissa Furlong*, Sandra Halverson-Deming, and Dale Sandler Melissa Furlong* University of North Carolina, United States, E-mail Address: [email protected] Search for more papers by this author , Sandra Halverson-Deming Social and Scientific Systems, Inc, United States, E-mail Address: [email protected] Search for more papers by this author , and Dale Sandler National Institute of Environmental Health Sciences, Epidemiology Branch, United States, E-mail Address: [email protected] Search for more papers by this author AbstractChronic low-dosing of antibiotics promotes weight gain in livestock and laboratory animals. Antibiotics in early childhood have also been associated with obesity and weight gain. Here, we investigate the association between antibiotic use in adulthood and obesity in the Sister Study, a cohort of 35-74 year-old U.S. and Puerto Rican sisters of women with breast cancer. We first used log-binomial regression to estimate prevalence ratios (PRs) of the associations between self-reported ever-use of tetracycline or penicillin and obesity at study entry (n=48,819). Chronic antibiotic use was defined as ever used at least 3x a week for 3 months or more, and obesity was defined as examiner-measured body mass index (BMI) of >30. Risk ratios (RR) for obesity associated with initiating antibiotic use during a woman’s 30’s were also estimated using Poisson regression in a subset of the cohort that only included women who were not obese in their 30’s and had a BMI measurement in their 40’s or 50’s (n=27,865). Covariates were determined using backward selection and change-in-estimate, and final covariates included age at BMI measurement, race, exercise, and education. The PR for ever-use of penicillin was 1.15 (95% confidence interval [CI] 1.08, 1.22), while for tetracycline the PR was 0.85 (95% CI 0.79, 0.90). The RR of the association between initiation of antibiotic usage during a woman’s 30s and obesity during her 40s and 50s for penicillin was 2.04 (95% CI 1.49, 2.81), while for tetracycline it was 0.72 (95% CI 0.55, 0.94). For penicillin, risk estimates did not vary by indication for use. However, tetracycline risk estimates differed if use was for skin or non-skin related reasons (p=0.01). The RR associated with non-skin related tetracycline use was 1.29 (95% CI 0.73, 2.28]), while for skin-related use the RR was 0.69 (95% CI 0.50, 0.94). Chronic penicillin use may be associated with obesity while the association between tetracycline and obesity is less clear.