Exploring the Association of Antimicrobial Use with Serratia marcescens Resistance Rates via Multiple Linear Regression

Fangyuan Xia,1 Xiaohua Zhang,1 Sucai Chen,2 Dali Zeng,1 Zhen Tao,3 Wensi Gan,4 Yu Pan1 1Department of Pharmacy, Wenzhou Hospital of Integrated Traditional Chinese and Western Medicine, Wenzhou, Zhejiang, People’s Republic of China; 2Clinical Laboratory Department, Wenzhou Hospital of Integrated Traditional Chinese and Western Medicine, Wenzhou, Zhejiang, People’s Republic of China; 3Intensive Care Unit, Wenzhou Hospital of Integrated Traditional Chinese and Western Medicine, Wenzhou, Zhejiang, People’s Republic of China; 4Department of Infection Control, Wenzhou Hospital of Integrated Traditional Chinese and Western Medicine, Wenzhou, Zhejiang, People’s Republic of ChinaCorrespondence: Yu Pan, Wenzhou Hospital of Integrated Traditional Chinese and Western Medicine, Tel +86 15267758678, Email [email protected]: To investigate the macro-level quantitative relationship between Serratia marcescens resistance and antimicrobial consumption one quarter in advance, aiming to curb resistance and optimize antimicrobial use.Patients and Methods: A retrospective analysis was conducted on S. marcescens resistance rates and antimicrobial consumption data from our hospital. Multiple linear regression models were employed to identify independent linear correlations between resistance rates and defined daily doses (DDDs) of specific antimicrobials.Results: Over the past four years, our hospital identified 522 S. marcescens strains (3.22% of all bacterial isolates), with 86.59% isolated from respiratory samples. The strains showed sensitivity to cefoperazone-sulbactam, cefepime, ertapenem, imipenem, meropenem, amikacin, trimethoprim-sulfamethoxazole, and tigecycline, with resistance rates < 10%. The study showed significant correlations between S. marcescens resistance and antibiotic usage. Resistance to cefoperazone-sulbactam and imipenem had independent negative linear relationships with gentamicin DDDs; resistance to cefoxitin correlated negatively with piperacillin-tazobactam DDDs; resistance to cefepime showed a negative association with cefuroxime DDDs. These four relationships were strongly supported by consistent results from Bayesian, Bootstrap, and Winsorized regression. Additionally, amoxicillin-clavulanic resistance positively correlated with meropenem DDDs, and levofloxacin resistance positively correlated with gentamicin DDDs. These positive trends were supported by triple robustness testing. These findings have substantial implications for clinical practice. The negative correlations indicate that the strategic use of specific antimicrobials can effectively suppress the resistance rates to target drugs, while the positive correlations reflect increased co-resistance risks. These findings underscore the necessity for antibiotic rotation and optimized management strategies.Conclusion: The significant associations between S. marcescens resistance rates and prior antimicrobial consumption patterns underscore the critical impact of antibiotic use on resistance development. This highlights the need for better antimicrobial stewardship to delay resistance and guide prescribing.Keywords: S. marcescens, multiple linear regression analysis, correlation