Infections and antibiotic use in early life, and obesity in early childhood: a mediation analysis.

BACKGROUND/OBJECTIVE: Literature shows a positive association between antibiotics and obesity in childhood, but fails to account for confounding by indication. We evaluate the direct effect of infection on obesity and the indirect effect mediated by antibiotics by performing a mediation analysis of the infection-obesity association. METHODS: A Medicaid cohort of children age 2-14 years old between 2015-2019 (n = 61,330) is used to perform mediation analysis of infections and antibiotic use in the first year of life, and obesity in childhood in Missouri, U.S.A. RESULTS: An additional infection increases the risk of obesity in childhood (aIRR = 1.050, p < 0.001); however, mediation by antibiotic use is clinically and statistically insignificant. If the number of infections is not considered in the analysis, then antibiotic use as a risk factor for obesity is overstated (aIRR = 1.037 vs. 1.013 p < 0.001). CONCLUSIONS: The number of infections exhibits a significant relationship with obesity and is a stronger risk factor for obesity than antibiotic use. In particular, a greater number of bronchitis, otitis media, and upper respiratory infections in the first year of life are associated with a significant increased risk of obesity in childhood. We find only weak evidence that an additional antibiotics claim increases the risk of obesity in childhood and this risk may not be clinically meaningful. Further research is needed to explore the association between early childhood infections, especially in the first 6 months of life, and obesity including the biological mechanism and environmental factor of early life infections associated with obesity.

Coronavirus Disease 2019 (COVID-19) Disease Severity: Pregnant vs Nonpregnant Women at 82 Facilities.

BACKGROUND: Pregnancy has been reported to be a risk factor for severe COVID-19. We evaluated the impact of pregnancy on severe COVID-19 and mortality in an electronic medical record (EMR) database that enabled exclusion of labor and delivery (L&D) encounters. METHODS: In this retrospective cohort study, EMRs from 82 healthcare facilities in the Cerner COVID-19 Datamart were analyzed. The study comprised 38 106 individuals aged 18-45 years old with COVID-19 who had emergency department, urgent care, or inpatient encounters from December 2019 to September 2020. Subgroups were balanced through propensity score weights for age, race, smoking status, and number of comorbidities. The primary outcome was COVID-19-related mortality; secondary outcomes were markers of severe COVID-19: intubations, mechanical ventilation, use of vasopressors, diagnosis of sepsis, and diagnosis of acute respiratory distress syndrome. RESULTS: In comparing pregnant and nonpregnant women, no statistical differences were found for markers of severe COVID-19, after adjusting for age, smoking, race, and comorbidities. The adjusted odds of an inpatient encounter were higher for pregnant vs nonpregnant women (adjusted odds ratio [aOR], 13.2; 95% confidence interval [CI], 11.6-15.3; P < .001), but notably lower after excluding L&D encounters (aOR, 2.3; 95% CI, 1.89-2.88; P < .001). In comparison to women without L&D encounters, hospitalization was significantly more likely for men. CONCLUSIONS: We did not find an increased risk of severe COVID-19 or mortality in pregnancy. Hospitalization does not necessarily indicate severe COVID-19 in pregnancy, as half of pregnant patients with COVID-19 were admitted for L&D encounters in this study.