Assessing limits of detection in qualitative methods: A simple implementation of logistic regression in a web-based R Shiny application environment and its potential in toxicology and doping control.

The estimation of limits of detection (LOD) for solely qualitative methods in analytical chemistry may prove challenging because all the approaches with which chemists are familiar require some type of numeric data input. The best model to describe the binary response in these methods (detected/not detected) is a logistic model; however, these models are not easily handled by most of the laboratories and generally demand expensive statistical software packages. In this work, the advantages of applying this approach are discussed and its implementation using commercial spreadsheet software is demonstrated. A free online application based on the R environment using shinyapps was developed and its application was validated and discussed with a dataset of 57 different target compounds analyzed in urine according to the requirements of the World Anti-Doping Agency (WADA). This tool allows free, extremely quick, and easy determinations of LOD in qualitative analyses as well as the determination of the probabilities of detection in any given concentration.

Validation of an ultra-sensitive detection method for steroid esters in plasma for doping analysis using positive chemical ionization GC-MS/MS.

The standard approach to detect misuse with testosterone in sport is based on the determination and evaluation of the urinary steroid profile followed by the confirmation of atypical profiles using isotope ratio mass spectrometry. The detection capacity of these methods can be attenuated by confounding factors or testosterone preparations with endogenous isotopic fingerprints. An alternative detection method for misuse of an endogenous steroid in sports is the direct detection of the administered steroid ester present in most preparations. Thus unambiguous proof for doping misuse can be delivered. In this work, the sensitivity of gas chromatography coupled to a triple quadrupole with chemical ionization (GC-CI-MS/MS) is applied to detect trace levels of 10 testosterone and 2 nandrolone esters in plasma for in human doping analysis. The detection method was developed employing a liquid-liquid extraction and HPLC cleanup step before analysis on the GC-CI-MS/MS. The quantitative method was validated in a linear range of 100-2000 pg/ml and proved to be selective, reproducible and very sensitive with limits of detection as low as to 10 pg/ml. A clinical study with the administration of testosterone undecanoate in 3 volunteers was carried out and the compound was detectable up to 86 days after administration.