The impact of accuracy and precision of analytical test methods on the determination of withdrawal periods.

Treatment of food-producing animals with veterinary medicinal products (VMPs) can result in residues in foodstuffs (e.g. eggs, meat, milk, or honey) representing a potential consumer health risk. To ensure consumer safety, worldwide regulatory concepts for setting safe limits for residues of VMPs e.g. as tolerances (US) or maximum residue limits (MRLs, EU) are used. Based on these limits so-called withdrawal periods (WP) are determined. A WP represents the minimum period of time required between the last administration of the VMP and the marketing of foodstuff. Usually, WPs are estimated using regression analysis based on residue studies. With high statistical confidence (usually 95% in the EU and 99% in the US) the residues in almost all treated animals (usually 95%) have to be below MRL when edible produce is harvested. Here, uncertainties from both sampling and biological variability are taken into account but uncertainties of measurement associated with the analytical test methods are not systematically considered. This paper describes a simulation experiment to investigate the extent to which relevant sources of measurement uncertainty (accuracy and precision) can impact the length of WPs. A set of real residue depletion data was artificially 'contaminated' with measurement uncertainty related to permitted ranges for accuracy and precision. The results show that both accuracy and precision had a noticeable effect on the overall WP. Due consideration of sources of measurement uncertainty may improve the robustness, quality and reliability of calculations upon which regulatory decisions on consumer safety of residues are based.