Early Life Environmental Exposure to Cadmium, Lead, and Arsenic and Age at Menarche: A Longitudinal Mother-Child Cohort Study in Bangladesh

Abstract

Background: Several metals act as endocrine disruptors, but there are few large longitudinal studies about associations with puberty onset.

Objectives: We evaluated whether early life cadmium, lead, and arsenic exposure was associated with timing of menarche.

Methods: In a mother-child cohort in rural Bangladesh ($n=935$), the exposure was assessed by concentrations in maternal erythrocytes in early pregnancy and in girls' urine at 5 and 10 years of age using inductively coupled plasma mass spectrometry. The girls were interviewed twice, at average ages 13.3 [standard deviation $\left(\text{SD}\right)=0.43$] and 13.8 ($\text{SD}=0.43$) y, and the date of menarche, if present, was recorded. Associations were assessed using Kaplan-Meier analysis and multivariable-adjusted Cox regression.

Results: In total, 77% of the girls ($n=717$) had reached menarche by the second follow-up. The median age of menarche among all girls was 13.0 y (25th-75th percentiles: 12.4-13.7 y). At 10 years of age, median urinary cadmium was $0.25\mu g/L$ (5th-95th percentiles: $0.087-0.72\mu g/L$), lead $1.6\mu g/L$ ($0.70-4.2\mu g/L$), and arsenic $54\mu g/L$ ($19-395\mu g/L$). Given the same age, girls in the highest quartile of urinary cadmium at 5 and 10 years of age had a lower rate of menarche than girls in the lowest quartile, with an adjusted hazard ratio of (HR) 0.80 (95% CI: 0.62, 1.01) at 5 years of age, and 0.77 (95% CI: 0.60, 0.98) at 10 years of age. This implies that girls in the highest cadmium exposure quartile during childhood had a higher age at menarche. Comparing girls in the highest to the lowest quartile of urinary lead at 10 years of age, the former had a higher rate of menarche [adjusted $\text{HR}=$ 1.23 (95% CI: 0.97, 1.56)], implying lower age at menarche, whereas there was no association with urinary lead at 5 years of age. Girls born to mothers in the highest quartile of erythrocyte arsenic during pregnancy were less likely to have attained menarche than girls born to mothers in the lowest quartile [adjusted $\text{HR}=$ 0.79 (95% CI: 0.62, 0.99)]. No association was found with girls' urinary arsenic exposure.

Discussion: Long-term childhood cadmium exposure was associated with later menarche, whereas the associations with child lead exposure were inconclusive. Maternal exposure to arsenic, but not cadmium or lead, was associated with later menarche. https://doi.org/10.1289/EHP11121.