How much percentage of an isolate should be resistant in vitro for a certain antibiotic to be labeled as resistant so that the antimicrobial will no longer considered in a treatment guideline? For example the use of Azithromycin for gonococcus.
There are guidelines either clinical or epidemiological defining the cut off values to consider resistance to a given antibiotic and bacterial species. The ones most generally used were defined by CLSI (in USA) or EUCAST (in Europe). For particular cases there are also some defined by EFSA, the European Agency for Environmental and Food Safety. This values are based on the minimal inhibitory concentrations distributions for a given antibiotic and bacterium.
I understand the question in that way that you want to know a threshold up to which an antibiotic can safely be considered for empiric or calculated therapy (i.e. one can assume a high probability of cure based on the known or anticipated resistance rate of the likely pathogens of that infection). In this way, the question is not easy to answer since different models may apply.
For uncomplicated urinary tract infections, trimethoprim/sulfamethoxazole may be used empirically when resistance rates are
CLSI (in USA) or EUCAST (in Europe) will give you useful information. However in EUCAST you will get ECOFFs that give you an idea about the presence of genetic determinants of resistance even if phenotypes are susceptible.