OBJECTIVES To quantify associations between antimicrobial use and acquired resistance in indicator Escherichia coli over a period of time which involved sector-wide antimicrobial use reductions in broilers and pigs (years 2004-14), veal calves (2007-14) and dairy cattle (2005-14). Prevalence estimates of resistance were predicted for a hypothetical further decrease in antimicrobial use. METHODS Data reported annually for the resistance surveillance programme in the Netherlands were retrieved. Two multivariate random-effects logistic models per animal sector were used to relate total and class-specific antimicrobial use (as defined daily dosages per animal per year, DDDA/Y) with the probability of E. coli resistance to a panel of 10 antimicrobial agents. RESULTS Positive dose-response relationships (ORs) were obtained from all models. Specific resistance phenotypes were more often associated with total antimicrobial use than with class-specific use. The most robust associations were found in pigs and veal calves. Resistance to historically widely used antimicrobials (e.g. penicillins, tetracyclines) was, in relative terms, less influenced by drug use changes over time than resistance to newer or less prescribed antimicrobials (e.g. third-/fourth-generation cephalosporins, fluoroquinolones). In pigs and veal calves, prevalence estimates for the most common resistance phenotypes were projected to decline ∼5%-25% during 2014-16 if total antimicrobial use reduction reached 80%; projections for poultry and dairy cows were more modest. CONCLUSIONS Epidemiological evidence indicated that drug use history and co-selection of resistance are key elements for perpetuation of resistance. Data suggest that recent Dutch policies aimed at reducing total use of antimicrobials have decreased E. coli resistance in the pig and veal calf production sectors while the impact on the dairy cattle and poultry sectors is less clear.