Aim: Clinical antimicrobial resistance (AMR) is a significant threat to public health, which is often unclear due to representative data from human populations that are challenging to obtain. Study the associations between consumption of antibiotics and antimicrobial resistance bacteria or genes (ARB or ARGs) that will benefit from elucidating the AMR. Methods: The details of antibiotics usage were calculated based on the actual consumption in the target hospital, ARB was detected by culture method, and ARGs were evaluated by metagenomics. Results: Our study revealed that culture-based single-indicator-strain approaches only capture the AMR in 16.17% infectious samples. 1573 bacterial species and 885 types of ARGs were found in the hospital sewage. The consumption of antibiotics influences the resistance profiles that were significant in E.coli, but the strength varies among bacteria. In all ARGs group, ARGs of aminoglycosides was the most common, followed by sulfonamide, tetracycline, phenicol, macrolides, and quinolones, comprising to 82.6% of all ARGs. Five hundred nineteen pairs of ARGs and bacterial species showed a significant correlation (r > 0.8). The co-occurrence patterns of bacteria- ARGs mirrors the AMR of the clinic. Antibiotic usage will affect the abundance of ARG in sewage, with a hysteresis effect. Conclusion: The ARGs- bacteria co-occurrence patterns from wastewater could be a valuable bio-indicator to reflect the emergence of ARB in the future. Developing a predictive risk model of AMR on this basis will facilitate the rational use of antibiotics.
