The calculation of pharmacokinetic/pharmacodynamic surrogates from concentrations in serum has been shown to yield important information for the evaluation of antibiotic regimens. Calculations based on concentrations in serum, however, may not necessarily be appropriate for peripheral-compartment infections. The aim of the present study was to apply the microdialysis technique for the study of the peripheral-compartment pharmacokinetics of select antibiotics in humans. Microdialysis probes were inserted into the skeletal muscle and adipose tissue of healthy volunteers and into inflamed and noninflamed dermis of patients with cellulitis. Thereafter, volunteers received either cefodizime (2,000 mg as an intravenous bolus; n = 6), cefpirome (2,000 mg as an intravenous bolus; n = 6), fleroxacin (400 mg orally n = 6), or dirithromycin (250 mg orally; n = 4); the patients received phenoxymethylpenicillin (4.5 x 10(6) U orally; n = 3). Complete concentration-versus-time profiles for serum and tissues could be obtained for all compounds. Major pharmacokinetic parameters (elimination half-life, peak concentration in serum, time to peak concentration, area under the concentration-time curve [AUC], and AUC/MIC ratio) were calculated for tissues. For cefodizime and cefpirome, the AUCtissue/AUCserum ratios were 0.12 to 0.35 and 1.20 to 1.79, respectively. The AUCtissue/AUCserum ratios were 0.34 to 0.38 for fleroxacin and 0.42 to 0.49 for dirithromycin. There was no visible difference in the time course of phenoxymethylpenicillin in inflamed and noninflamed dermis. We demonstrated, by means of microdialysis, that the concept of pharmacokinetic/pharmacodynamic surrogate markers for evaluation of antibiotic regimens originally developed for serum pharmacokinetics can be extended to peripheral-tissue pharmacokinetics. This novel information may be useful for the rational development of dosage schedules and may improve predictions regarding therapeutic outcome.