Antibiotic resistance genes (ARGs) discharged through wastewater treatment plants (WWTPs) has aroused growing public concern for its risk to human health and ecological safety. Membrane bioreactor (MBR) has been recognized as an effective approach to remove ARGs in full-scale WWTPs, but its advantage over traditional processes was not clearly quantified. To address this, we investigated four full-scale WWTPs containing parallel MBR and traditional processes (oxidation ditch or sequencing batch reactor) to compare the reduction of eight types of ARGs (blaTEM, ermB, tetW, tetO, sul1, sul2, addD, and qnrS) and int1. In general, MBRs reduced the ARGs (1.1-7.3 log removal) better than parallel processes (0.4-4.2 log removal). Notably, the dominant ARGs in the influent, such as ermB, sul1 and int1 (106.39-107.79 copies/mL), were more effectively reduced by MBRs (1.5-7.3 log removal) than traditional processes (0.8-3.4 log removal). Meanwhile, the distribution of those ARGs in activated sludge was not significantly different between aforementioned processes (p > 0.05). The separation coefficient (Ksw) was proposed to represent the contribution of solid separation on ARG removal, subsequent analysis revealed surprisingly strong correlation between Ksw values of dominant ARGs (ermB, sul1 and int1) and their log removal by MBR (R = 0.79-0.96, p < 0.05), while such correlation was much weaker in traditional process (R = 0.33-0.37), indicating solid separation was the major pathway for removal of dominant ARGs and int1. According to the canonical correlation analysis between process operation and ARG removal in MBR, sludge retention time (SRT) seemed to be the major factor affecting removal of dominant ARGs and int1. This comparative study can be helpful for further understanding and operating MBR process to reduce the ARGs in effluent.
Keywords: Antibiotic resistance genes; Canonical correlation analysis; Log removal value; Membrane bioreactor; Separation coefficient.
Copyright © 2018. Published by Elsevier B.V.