Use of human data for risk assessment of pesticides: A review including an evaluation of trichlorfon as case study

Pesticides are suspected to cause health effects in humans, but human-based data on their toxicity are often insufficient to establish associations and quantify risks. We reviewed literature on trichlorfon and investigated methodological aspects of risk assessment for pesticides based on human data. We provided an overview of epidemiology of pesticide toxicity, with focus on methodological features of the available studies, combined with a systematic review of the health effects of the pesticide, trichlorfon in humans, focusing on studies published after 1990. Studies on dichlorvos, the metabolite of trichlorfon, as well as metrifonate, a medication with the identical formula as trichlorfon, were included. A total of 60 publications were identified on health effects of trichlorfon, dichlorvos and metrifonate. Studies on acute effects (N = 23 publications) comprised mainly case-reports related to accidents and suicidal attempts, and were connoted by cholinergic syndrome, gastrointestinal and general symptoms. Evidence on chronic effects derived from analyses of the Agricultural Health Study (28 publications), as well as case-control and cross-sectional studies (9 publications). Evidence of possible associations between trichlorfon or dichlorvos exposure and various outcomes was heterogeneous and insufficient to establish causality. Critical features of epidemiology studies used for pesticide risk assessment include study design, exposure misclassification, lack of quantitative exposure data, and lack of consideration to potential confounders. Few high-quality epidemiology studies are available on potential health effects of trichlorfon. Future studies conducted according to established guidelines and supported by artificial intelligence might help to fill the gap on human health risks from pesticides.