Estimation of measurement uncertainty for the determination of loss on drying of biologicals

Abstract

Scientific relevance. GOST ISO/IEC 17025-2019 requires testing laboratories to evaluate the measurement uncertainty of their results. Estimating the uncertainty of analytical methods intended for biologicals is a challenging task that requires time, effort, and a special approach. Measurement uncertainty estimation is of particular interest in the case of measuring loss on drying (LOD) for biologicals, since LOD testing procedures involve analysing measurements of a physical value, i.e. mass.Aim. This study aimed to estimate the measurement uncertainty of LOD determination in biological medicinal products.Materials and methods. The study examined a powdered active substance intended for a Bifidobacterium product (test sample). The authors conducted the LOD test in accordance with the State Pharmacopoeia of the Russian Federation (OFS.1.2.1.0010.15). Statistical processing of the results was performed using Microsoft Excel. To estimate the measurement uncertainty, the authors employed the bottom-up approach or used the standard deviation from testing results.Results. The authors identified the uncertainty components that affected the LOD determination results. When calculated using the bottom-up approach, the expanded uncertainty was 0.34% (coverage factor, k=2; approximate confidence level, 95%). In particular, the largest contributor to the expanded uncertainty was the uncertainty of measuring the mass of weighing bottles containing dried test samples (0.147%), whereas the smallest contributor was the uncertainty of weighing empty bottles (0.003%). When calculated using the standard deviation, the uncertainty of two parallel measurements amounted to 0.32%.Conclusions. Both approaches to calculating LOD measurement uncertainty yield comparable results. According to the uncertainty budget analysis, the uncertainty of measuring the mass of weighing bottles with dried test samples is the major contributor to the test result. For this reason, the conditions of sample preparation should be carefully controlled. The study results confirm that the LOD measurement uncertainty can be calculated using the standard deviation. Testing laboratory teams may benefit from the methods for identifying the factors influencing LOD test results and the methods for calculating the uncertainty of measurement described in this study.