Seventy-three commonly administered pharmaceuticals from twelve different therapeutic classes were investigated at a municipal wastewater treatment plant in northern Italy featuring a conventional activated sludge system (full-scale) and a polishing horizontal subsurface flow bed (pilot plant). Removal of these micro-pollutants by the two systems was assessed in order to evaluate their respective contributions. Mean concentrations and standard deviations were calculated and found to differ for the compounds detected, ranging from few ng/L to over 1,165 ng/L in the secondary effluent and from 11 to 533 in the polished effluent. Eighteen compounds were consistently below the detection limit and the remaining 55 compounds were found at a minimum of one sampling point. Average removal efficiencies of both treatment steps and in treatment train as a whole are evaluated and discussed, highlighting the difficulties in predicting the fate of pharmaceuticals in both an activated sludge system and a horizontal subsurface flow bed. Comparison between the observed average removal efficiencies and those reported in the literature was also carried out for the pharmaceuticals of interest, and the discrepancies that emerged are discussed. The investigated constructed wetland did show efficacy in removing some of these compounds, and it contributed to the overall removal efficiency of each therapeutic class. Indeed, evaluation of the specific mass loadings of each class of PhC detected in the raw wastewaters, secondary and polished effluent evidences that the investigated constructed wetland is able to further reduce the load of micropollutants, which could become a necessity, especially where the receiving water body is an effluent-dominant river and mitigation measures of the discharge impact are required to protect and safeguard the aquatic environment.