Longitudinal association between common pediatric infections and exposure to PCB

PDS 75: Microbiome, infections and respiratory effects, Johan Friso Foyer, Floor 1, August 27, 2019, 1:30 PM - 3:00 PM Background: Exposure to polychlorinated biphenyls (PCBs) may affect the incidence of common pediatric infections. Methods: Pediatric infections were analyzed in subjects of a mother-child cohort launched in eastern Slovakia in 2002. We included 922, 697, 818, 477, and 386 children, whose median serum concentration of the sum of 15 PCB congeners was 325.0, 385.2, 324.1, 338.4, and 133.5 ng/g lipids at the age of 0, 6, 16, 45, and 72 months, respectively. Outcomes studied were diagnoses of upper and lower respiratory tract infections (common cold, tonsillitis, sinusitis, laryngitis, flu, bronchitis, pneumonia, bronchiolitis), acute otitis media, and indirect health indicators as hospitalization, antibiotic administration, and medical emergency calls. We recognized 5 exposure windows (EWs) (prenatal, 0-6, 7-16, 17-45, 46-72 months) and 4 observation periods (OPs) (0-6, 7-16, 17-45, 46-72 months) and examined relationships between individual exposures and outcomes by logistic regression adjusted for confounders. We estimated the odds ratios (OR) of individual outcomes in the second, third and fourth quartiles of PCB concentration against the reference first. For overview, we collected all significant ORs higher than 1 (= events) associated to exposure in EWs versus OPs combination and we created the sum of significant events. Results: The prenatal EW was associated with 1, 4, 5, and 1 events at 0-6, 7-16, 17-45, and 46-72 OPs, respectively. The 0-6 EW was associated with 17 and 8 events at 0-6 and 46-72 OPs, respectively. The 7-16 EW was associated with 6 and 3 events at 17-45 and 46-72 OPs, respectively. The 17-45 EW was associated with 9 events at 46-72 OP. Finally, the 46-72 EW was associated with 5 events at 46-72 OP. Conclusions: In our longitudinal analysis, we observed, however, inverse associations between PCB exposure and outcomes of interest as well, in agreement with cross-sectional literature data.