Premium
Is antibiotic use a risk factor for breast cancer? A meta‐analysis
Author(s) -
Sergentanis Theodoros N.,
Zagouri Flora,
Zografos George C.
Publication year - 2010
Publication title -
pharmacoepidemiology and drug safety
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.023
H-Index - 96
eISSN - 1099-1557
pISSN - 1053-8569
DOI - 10.1002/pds.1986
Subject(s) - medicine , breast cancer , meta analysis , publication bias , confounding , antibiotics , odds ratio , medical prescription , cancer , oncology , gynecology , pharmacology , microbiology and biotechnology , biology
Abstract Purpose Breast cancer is a disease integrating hormonal and non‐hormonal factors. Given the widespread antibiotic use, attention has been recently drawn upon the association between antibiotic use and breast cancer; however, the published studies have yielded contradictory results. In addition, various types of quantification in antibiotic use have been adopted. This meta‐analysis aims to examine whether antibiotic use is associated with breast cancer risk presenting two analyses: one on antibiotic ever‐use and one on the number of antibiotic prescriptions. Methods Eligible studies were retrieved by a search in MEDLINE, Cochrane, and EMBASE databases till July 2009. Odds ratios (OR) pertaining to antibiotic ever‐use were appropriately calculated. The random effects model was used to estimate the pooled OR. Publication bias was assessed through Begg's and Egger's tests. Meta‐regression with the number of antibiotic prescriptions was performed. Results Five case–control studies were eligible at the ever‐use versus never‐use analysis (13 069 cases and 73 920 controls). Antibiotic ever‐use was associated with slightly elevated breast cancer risk (pooled OR = 1.175, 95%CI: 0.994–1.387). No publication bias became apparent. Meta‐regression showed a borderline dose–response effect implicating the number of antibiotic prescriptions. Conclusions Antibiotic use seems associated with slightly elevated breast cancer risk. The underlying nature of the association remains elusive, as it may be direct or due to secondary associations, that is, causal or confounding. At any case, this is a finding with potentially important public health implications, which should be further examined in the literature. Copyright © 2010 John Wiley & Sons, Ltd.