Arsenic exposure levels were measured using maternal hair samples that were collected at the first visit to the hospital (N=497). Arsenic and the other trace elements were analyzed using ICP-MS after careful cleaning involving multiple washes with different media. Sonication was not mentioned. The LOD for As was 0.011 ug/g and concentrations below the LOD were replaced with the respective LOD divided by the square root of two. Only 30% of samples had detectable arsenic concentrations.

A potential source of measurement error or bias is that exposure assessment was not adequately standardized due to the large variation in the length of hair samples analyzed. The authors reported that 60.5% of women provided the 24cm hair samples which they hoped to measure, but the length of samples was as low as 4cm. They did not report whether detection rates or concentrations of any elements varied by length of hair, i.e., by the timeframe captured by the available sample. Given the low detection rate for arsenic, it is a concern that women with longer hair samples were more likely to have had detectable levels. The authors did not report whether similar proportions of cases and controls provided hair samples of varied length.

Ren, 2022, 10181214

Study evaluation