Systematic sampling and analysis of wastewater samples are increasingly adopted for estimating drug consumption in communities. An understanding of the in-sewer transportation and transformation of illicit drug biomarkers is critical for reducing the uncertainty of this evidence-based estimation method. In this study, biomarkers stability was investigated in lab-scale sewer reactors with typical sewer conditions. Kinetic models using the Bayesian statistics method were developed to simulate biomarkers transformation in reactors. Furthermore, a field-scale study was conducted in a real pressure sewer pipe with the systematical spiking and sampling of biomarkers and flow tracers. In-sewer degradation was observed for some spiked biomarkers over typical hydraulic retention time (i.e., a few hours). Results indicated that sewer biofilms prominently influenced biomarker stability with the retention time in wastewater. The fits between the measured and the simulated biomarkers transformation demonstrated that the lab-based model could be extended to estimate the changes of biomarkers in real sewers. Results also suggested that the variabilities of biotransformation and analytical accuracy are the two major contributors to the overall estimation uncertainty. Built upon many previous lab-scale studies, this study is one critical step forward in realizing wastewater-based epidemiology by extending biomarker stability investigations from laboratory reactors to real sewers.