Estimating antibiotic resistance following antibiotic treatment in outpatients: a retrospective study

Objectives Understanding the impact of antibiotic use on future resistant infections is essential for appropriate antibiotic therapy. Our objective was to assess the effect of exposure to commonly used antibiotics in community settings on resistance of future urine culture isolates. Methods A retrospective study including adult Health Maintenance Organization members who purchased cefazolin, cefuroxime, amoxicillin-clavulanate (AMC), or fluoroquinolones (FQ) between 1 January 2017 and 31 December 2019. The outcome was the antibiotic susceptibility profile of the first urine culture within 6 months from antibiotic purchase. We emulated the two following target trials: (1) patients treated with first-generation cephalosporins (1st CEP) would instead receive second-generation cephalosporins (2nd CEP), AMC, or FQ. (2) Patients treated with AMC or FQ would instead receive 2nd CEP. We used matching, standardisation, postmatching covariate adjustments and inverse probability of censoring weighting to correct for confounding and selection bias and estimate the risk difference (RD) for resistance. Results Included were 23 297 patients. In the first target trial, exposure to any of the non-1st CEP antibiotics led to the highest increase in resistance of the same antibiotic class, greatest when treatment and outcome were FQ (RD, 22.1%; 95% CI, 20.0–24.2) and AMC (RD, 8.2%; 95% CI, 6.1–10.3). In the second trial, the 2nd CEP led to higher increases in resistance to all cephalosporins compared with AMC. The 2nd CEP was also associated with higher 3rd CEP resistance compared with FQ (RD, −2.1%, 95% CI, −3.9, −0.4). Conclusions Our study provides novel estimates of the impact of exposure to oral antibiotics on the resistance of future urinary pathogens. These estimates are relevant for both antibiotic treatment guidelines and clinical decision-making.