Need for Reassessment of the Quality Assurance in Food Products: Meta Analysis of Mass Spectrometry of Pesticide Residues in Food Products

While mass spectrometry (MS) excels at identifying pesticides due to its high resolution, its accuracy for quantifying pesticide concentrations is problematic. This meta-analysis employed a representative random sample of publications; consequently, alternative samples might generate divergent observations. The criteria for selecting publications for the meta-analysis are comparability of technologies with each group of GC-MS/MS and LC-MS/MS, simultaneous analysis of multiple pesticides, and easy access to the data. Nevertheless, contemporary studies follow analogous analytical principles. The meta-analysis findings comprise a full re-calculation of reported data from 28 analyzed publications (14 GC-MS/MS, 14 LC-MS/MS) on pesticide quantification. Outliers were reported by only one publication. Analysis (using PoPC) revealed a critical bias—negative calibration slopes were systematically discarded, indicating researchers likely repeated calibrations until obtaining positive slopes. This invalidates the statistical validity of the methods. The ensuing quantifications were therefore postulated to be unreliable. The calibration slopes were too shallow, leading to extremely high relative uncertainties (often �100% or more). This makes reliable quantification impossible. A matrix-effect problem was disclosed; the massive uncertainties mean the reported "abundant matrix effects" cannot be reliably distinguished from the measurement error itself. Claims about matrix effects were made with high precision but low accuracy. Due to the methodological bias and resulting high uncertainty, the actual concentration of pesticides in these studies remains scientifically undetermined. Potential bias may lead to uncertain conclusions, and issues with statistical methods and reproducibility were noted, warranting future attention. Preconcentrating pesticides could improve reliability, as the PoPC showed lower variability at higher concentrations.