Objectives: A cross-sectional study was carried out to assess the effects of methylmercury exposure on neurodevelopment in Japanese children, in relation to the Madeiran cross-sectional study, and to estimate benchmark dose (BMD) levels using the data of two studies.
Methods: Mercury levels in hair samples obtained from 327 Japanese mothers and their 7-year-old children, and methylmercury levels in the umbilical cord, were determined. Neurodevelopmental examinations, including the brainstem auditory evoked potential (BAEP), were performed on the children.
Results: The medians of hair mercury were 1.63 (0.11-6.86) mug/g for mothers and 1.65 (0.35-6.32) mug/g for children, and a significant correlation was seen between the hair mercury levels in mothers and children. The maternal hair mercury was significantly correlated with the methylmercury in the umbilical cords obtained from 49 children. In 210 children whose mothers had not changed their dietary habits since pregnancy, most of the neurodevelopmental variables were not significantly related to hair mercury levels. The BAEP latencies were significantly shorter in the Japanese children than in the 113 Madeiran 7-year-old children, whose mothers had hair mercury of 1.12-54.5 (median 10.9) mug/g. Significant relationships between the maternal hair mercury level and BAEP latencies (peaks III and V, and interpeak I-III) were found only in the merged data of Japanese and Madeiran children. When the lower 95% confidence limit of BMD (BMDL) was calculated, the BMDLs of mercury exposure for BAEP latencies in the merged data were between 6.9 and 10.5 mug/g, and lower than those in the Madeiran children.
Conclusions: It is suggested that Japanese children may ingest similar doses per body weight of methylmercury to their mothers. If maternal hair mercury was used as a proxy for mercury exposure at birth, no significant dose-effect associations with the BAEP latencies were observed in Japanese children with exposure levels below 6.9 mug/g of hair mercury, but only when higher-level exposures from Madeiran children were included. The BMDL was lower for the merged data than for Madeiran children alone.