2300

Introduction: Experimental studies have suggested endocrine disruption by BP-3, an active ingredient in sunscreen, with widespread exposure around the world. Methods: We recruited girls ages 6 - 8 years who returned semi-annually for pubertal maturation staging (thelarche, pubarche), to provide blood for serum hormone assay [estradiol, estrone, testosterone, dehydroepiandrosterone-sulfate (DHEA-S)], and urine for BP-3 assay. Parents completed yearly questionnaires on participant’s past-year sunscreen use and attainment of menarche. Quantile regression determined the relationship between each log-transformed sex hormone with BP-3 exposure, measured by either biomarker (with half-life of 16 hours) or sunscreen questionnaire (use over the last year). Cox-proportional hazards models examined whether higher quartile of BP-3 exposure was associated with age-of-pubertal milestone. Results: The median value of baseline BP-3 measurements was 25.0 μg/g-creatinine (N=353), with detection in 98.9% of samples. The median number of days sunscreen was used in the past year was 48 days (N=302). There was no evidence of associations between any of the four hormones and the BP-3 biomarker. Testosterone, measured at the time of thelarche, was inversely associated with sunscreen use by questionnaire data (N=157, adjusted β= -0.0163, 97.5%CI: -0.0300, -0.0026). There was no evidence of association between sunscreen use by questionnaire and testosterone during other time windows (-6 months and +6 months) or other hormones at any time points. Age-of-Thelarche was later among participants in the 2nd quartile of BP-3 biomarker compared to the 1st quartile (N=282, adjusted HR=1.5900, 97.5%CI:1.0430-2.4230). Age-of-menarche and age-of-pubarche were not different among participants in different biomarker or questionnaire quartiles. Conclusions: Results suggest that higher report of sunscreen use was associated with lower testosterone levels during thelarche and a non-linear relationship between age-of-thelarche and BP-3 biomarker. Results using questionnaire and biomarker to represent exposure are not completely consistent with each other, thus conclusions should be interpreted with caution.