Estimating the effect of increasing ambient temperature on antimicrobial resistance in China: A nationwide ecological study with the difference-in-differences approach

Antimicrobial resistance (AMR) and the possible consequences of rising ambient temperatures brought on by global warming have been extensively discussed. However, the epidemiological evidence on the effects of temperature on AMR is rare and little is known about the role of socioeconomic inequities. This ecological study obtained 31 provinces AMR data of Escherichia Coli (E. coli) from the China Antimicrobial Resistance Surveillance System (CARSS) over the period from 2014 to 2020, which were linked to the meteorological and socioeconomic data published in the China Statistical Yearbook. Modified difference-in-differences (DID) analyses were performed to estimate the effect of ambi-ent temperature on AMR of E. coli to third-generation cephalosporins (ceftriaxone and cefotaxime), carbapenems, and quinolones, adjusting for variations in meteorological and socioeconomic factors. We estimated that every 1 degrees C in-crease in average ambient temperature was associated with 2.71 % (95 % confidence interval [CI]: 1.20-4.24), 32.92 % (95 % CI: 15.62-52.81), and 1.81 % (95 % CI: 0.47-3.16) increase in the prevalence of E. coli resistance to third-generation cephalosporins (ceftriaxone and cefotaxime), carbapenems and quinolones, respectively. The link was more profound in the regions with lower temperature and a median level of average humidity, and the regions with lower income, lower expenditure (in economics), lower health resources, and lower hospital admissions. Neither the replacement of the temperature variable nor the alternative approaches for confounding adjustment changed the positive association between ambient temperature and AMR. In general, there exists a positive association between ambient temperature and AMR, although the strength of such an association varies by socioeconomic and health services factors. The association is possibly nonlinear, especially for E. coli resistance to third-generation cephalospo-rins. The findings suggest that AMR control programs should explicitly incorporate weather patterns to increase their effectiveness.