The frequent use and overuse of antibiotics in the general population is a potential threat to public health because of rising rates of infections that do not respond well to antibiotics. In young children, over-prescription of antibiotics also poses additional risks during a critical time in the development of bacteria in the gut that can influence immune health, brain health, and ability to process nutrients. We will study how early exposure to antibiotics might help cause various chronic diseases to develop in children, including asthma, allergies, type 1 and type 2 diabetes, juvenile arthritis, celiac disease, attention disorders, depression, and autism. Children under age 12 who received antibiotics before age 2 will be compared with children who did not receive antibiotics before age 2. We will use specific codes to study children diagnosed with the diseases of interest. We will study whether risks of chronic diseases relate to what kinds of antibiotics and how many courses were prescribed. To account for information that is not present in the data but could affect our results, we will also study whether children who received antibiotics before age 2 are more likely to develop chronic diseases than siblings who did not receive antibiotics before age 2.
We will investigate the relationship between exposure to antibiotics before age 2 and incident chronic diseases linked to disruption of human microbiota, including conditions affecting the immune system, central nervous system (CNS), and metabolism. Antibiotic-exposed and -unexposed cohorts of children will be generated from CPRD data. Children under age 12 will be considered in primary analysis, but additional groups under age 18 will be included in secondary analysis. The study population will include children in the linked mother-child CPRD dataset born after the up-to-standard date. Linkage of children to their mothers will allow us to adjust for prenatal antibiotic exposures, maternal medical conditions, and other confounding variables (e.g., mode of birth delivery). We will define antibiotic exposure as a prescription for antibiotics within the first 2 years of life, with further categorization by type, spectrum of activity, and number of courses prescribed. Outcomes of interest include allergic or atopic diseases (asthma, allergic rhinitis, atopic dermatitis, food allergy), autoimmune diseases (type 1 diabetes, thyroid disease, juvenile idiopathic arthritis, inflammatory bowel disease, psoriasis, celiac disease), metabolic diseases (type 2 diabetes, obesity), and neuropsychiatric diseases (attention deficit hyperactivity disorder (ADHD), depression, anxiety, autism spectrum disorder, and learning disability). With the exception of obesity, we will define outcomes primarily by Read Codes and secondarily by Read Codes combined with disease-specific treatments. Obesity will be defined by a body mass index (BMI) Z-score >2 and secondarily by a BMI Z-score >3, based on UK growth standards. Wherever possible, we will use validated outcome definitions. Subjects with a diagnosis of interest prior to age 2 will be excluded from the study of that respective outcome. We will use multivariable Cox regression to study the association between antibiotic exposure and outcomes of interest, adjusting for maternal, child, and environmental confounders, including indication for antibiotic exposure. In addition to conducting a traditional cohort study, we will also use a sibling-comparison design to compare disease incidence among siblings with discordant antibiotic exposures before age 2 using stratified Cox regression.
Allergic or atopic diseases: asthma, allergic rhinitis, atopic dermatitis, food allergy
Autoimmune diseases: type 1 diabetes, thyroid disease (excluding congenital), juvenile idiopathic arthritis, inflammatory bowel disease, psoriasis, celiac disease
Metabolic diseases: type 2 diabetes, obesity
Neuropsychiatric diseases: ADHD, depression, anxiety, autism spectrum disorder, and learning disability
Daniel Horton - Chief Investigator - Rutgers Robert Wood Johnson Medical School
Edward Nonnenmacher - Corresponding Applicant - Rutgers, The State University of New Jersey
Abner Nyandege - Collaborator - Rutgers, The State University of New Jersey
Brian Strom - Collaborator - Rutgers, The State University of New Jersey
Cecilia Huang - Collaborator - Rutgers, The State University of New Jersey
Dawn Koffman - Collaborator - Rutgers, The State University of New Jersey
Dinesh Mendhe - Collaborator - Rutgers, The State University of New Jersey
Jason Roy - Collaborator - Rutgers, The State University of New Jersey
Martin Blaser - Collaborator - Rutgers, The State University of New Jersey
Matthew Beier - Collaborator - Rutgers Robert Wood Johnson Medical School
Matthew Iozzio - Collaborator - Rutgers, The State University of New Jersey
Soko Setoguchi - Collaborator - Rutgers, The State University of New Jersey
Tobias Gerhard - Collaborator - Rutgers, The State University of New Jersey
Avinash Gabbeta - Collaborator - Rutgers, The State University of New Jersey
CPRD Mother-Baby Link;Practice Level Index of Multiple Deprivation