Material and methods
We used a food frequency questionnaire (FFQ) completed by a sample of 194 pregnant women to estimate their dietary exposure to chlordecone and compared it to blood levels. In a first approach, chlordecone daily intake was estimated as the product of daily eaten quantity of 214 foodstuffs, multiplied by their chlordecone content, and summed over all items. We then predicted individual blood chlordecone concentration with empirical weight regression models based on frequency of food consumption, and without contamination data.