The association between macrolide resistance mechanisms and clinical outcomes remains understudied. The present study, using an in vitro pharmacodynamic model, assessed clarithromycin (CLR) activity against mef(A)-positive and erm(B)-negative Streptococcus pneumoniae isolates by simulating free-drug concentrations in serum and both total (protein-bound and free) and free drug in epithelial lining fluid (ELF). Five mef(A)-positive and erm(B)-negative strains, one mef(A)-negative and erm(B)-positive strain, and a control [mef(A)-negative and erm(B)-negative] strain of S. pneumoniae were tested. CLR was modeled using a one-compartment model, simulating a dosage of 500 mg, per os, twice a day (in serum, free-drug C(p) maximum of 2 micro g/ml, t(1/2) of 6 h; in ELF, C(ELF(total)) maximum of 35 micro g/ml, t(1/2) of 6 h; C(ELF(free)) maximum of 14 micro g/ml, t(1/2) of 6 h). Starting inocula were 10(6) CFU/ml in Mueller-Hinton broth with 2% lysed horse blood. With sampling at 0, 4, 8, 12, 20, and 24 h, the extent of bacterial killing was assessed. Achieving CLR T/MIC values of > or =90% (AUC(0-24)/MIC ratio, > or =61) resulted in bacterial eradication, while T>MIC values of 40 to 56% (AUC(0-24)/MIC ratios of > or =30.5 to 38) resulted in a 1.2 to 2.0 log(10) CFU/ml decrease at 24 h compared to that for the initial inoculum. CLR T/MIC values of < or =8% (AUC(0-24)/MIC ratio, < or =17.3) resulted in a static effect or bacterial regrowth. The high drug concentrations in ELF that were obtained clinically with CLR may explain the lack of clinical failures with mef(A)-producing S. pneumoniae strains, with MICs up to 8 micro g/ml. However, mef(A) isolates for which MICs are > or =16 micro g/ml along with erm(B) may result in bacteriological failures.