Acute otitis media (AOM) is the most frequent diagnosis in early childhood for which antibiotics are prescribed.1 Despite advances in antimicrobial therapy, the increasing incidence of resistance among strains of Streptococcus pneumoniae and Haemophilus influenzae has complicated empiric treatment of AOM. For an antibiotic to be effective in AOM, it should achieve concentrations in middle ear fluid (MEF) that exceed by at least 3- to 6-fold the MIC of the pathogen.2 For multidrug-resistant pneumococci many orally administered antibiotics do not achieve satisfactory MEF concentrations. Cefprozil is frequently prescribed for treatment of AOM and has been shown to be relatively effective for management of refractory AOM.3 The concentrations of cefprozil in MEF have been reported to be from 0.06 to 4.0 μg/ml at 0.4 to 6 h after 15-mg/kg doses.4 In the present study we determined the MEF concentrations of cefprozil in infants and children with AOM and compared these values with the MIC values of S. pneumoniae and H. influenzae. Methods. This was an open, noncomparative study. Children 6 months to 13 years of age (mean ± sd, 3.4 ± 2.5 years) with signs and symptoms of AOM and evidence of middle ear effusion by pneumatic otoscopy were eligible for study. The study was conducted at the Corporación para Investigaciones Biológicas in Medellín, Colombia, and the patients were referred for evaluation from private pediatricians' offices or local clinics to this research facility. Institutional review board approval was obtained at the Corporación para Investigaciones Biológicas. Written informed consent was obtained from a parent or legal guardian. To be included in the study patients had to have signs and symptoms of AOM by history such as fever, lethargy, irritability or otalgia associated with otoscopic findings of middle ear effusion by pneumatic otoscopy and evidence of erythema with or without bulging of the tympanic membrane. Patients were excluded if they had a perforated tympanic membrane; had received treatment with a systemic antimicrobial agent in the last 7 days or with a long-acting parenteral antibiotic (i.e. benzathine penicillin) within the previous 4 weeks before enrollment in the study; had known renal or hepatic disease, pregnancy, history of hypersensitivity to a penicillin or cephalosporin; previous participation in this study; or known severe immunodeficiency. On the day of enrollment a complete medical history and physical examination were performed and study drug was dispensed as a 250-mg/5 ml cefprozil suspension. After the first dose of 15 mg/kg patients were randomly assigned to have MEF, and serum samples were obtained at 1, 2, 4 or 6 h after administration of the dose. The goal was to have 10 patients at each of these time points. From 20 to 30 min before tympanocentesis the patients received midazolam (0.5 mg/kg po) for sedation. The MEF was aspirated through a 20-gauge spinal needle attached to a sterile tympanocentesis trap (Juhn Tym-Tap, Xomed-Treace). Plasma and MEF specimens were stored at -70°C. The concentrations of cefprozil in serum and MEF were determined by a standard disk diffusion microbioassay using Micrococcus luteus (ATCC 9341) as the test organism.5, 6 The lowest concentration of cefprozil detectable in MEF was 0.1 μg/ml. To correct for the presence of blood in MEF, the hemoglobin concentration in MEF was measured by colorimetric technique based on the catalytic action of hemoglobin on the oxidation of benzidine by hydrogen peroxide (Sigma Diagnostics, St. Louis, MO) and the cefprozil value in MEF was adjusted for the concentration contained in that amount of blood. Results. The study was conducted from August, 1997, to August, 1998, and 44 patients were enrolled. Four patients were excluded because no MEF was obtained on tympanocentesis. Thus 40 patients were evaluable for pharmacodynamic studies. Of the 40 MEF specimens 15 contained enough blood to result in a correction of >0.1 μg/ml. Individual and mean cefprozil concentrations in serum and MEF at 1, 2, 4 and 6 h after administration of the 15-mg/kg dose are presented in Figure 1. The mean MEF concentrations and range of values at the four time intervals were the following: at 1 h, 1.4 (0.2 to 3.5) μg/ml; at 2 h, 2.83 (0 to 7.4) μg/ml; at 4 h, 2.47 (0.1 to 5.8) μg/ml; at 6 h, 0.96 (0 to 1.9) μg/ml. The MEF concentrations were 30% of serum concentrations at 2 and 4 h.Fig. 1: Concentrations of cefprozil in serum and middle ear fluid of patients given a single 15-mg/kg dose. The concentrations are shown in relation to the MIC90 values for S. pneumoniae strains and H. influenzae.Of 40 MEF specimens 29 (72.5%) had concentrations greater than the MIC50 value (0.5 μg/ml) for cefprozil, whereas only 4 had values greater than the MIC90 (4 μg/ml) for intermediate resistant pneumococci.7 Four specimens had cefprozil concentrations greater than the MIC50 (4 μg/ml) and no specimens had values that exceeded the MIC90 (16 μg/ml) for resistant S. pneumoniae,7 or the MIC50 (8 μg/ml) and MIC90 (16 to 64 μg/ml) values for H. influenzae.8 Comment. With increasing rates of antimicrobial resistance worldwide, selection of an antibiotic for treatment of refractory or recurrent AOM is difficult because physicians rarely know the etiology of disease and do not appreciate the importance of the relation between the MIC for the pathogen and the concentration of drug in MEF (i.e. pharmacodynamics) of the drug.2 Concentrations of cefprozil in MEF of our patients exceeded 0.125 μg/ml in 36 patients (90%) and 4.0 μg/ml in 4 patients (10%), the cefprozil MIC90 values for penicillin-susceptible and intermediate-resistant pneumococcal strains, respectively. For a beta-lactam antibiotic to achieve bacteriologic eradication in 80 to 85% of cases of AOM, either MEF concentrations must exceed the MIC for the pathogen by 3- to 6-fold or serum values must exceed the MIC for 60 to 70% of the dosing interval.2 The MEF concentrations of cefprozil exceeded by at least 3-fold the MIC50 values for penicillin-intermediate resistant strains in 60% of our patients, but none exceeded by 3- to 6-fold the MIC90 values for these and for resistant strains of S. pneumoniae or the MIC50 and MIC90 values for H. influenzae. Additionally neither the serum nor the MEF concentrations of cefprozil exceeded the MIC90 values for penicillin-intermediate resistant pneumococci for >50% of the 12-h dosing interval. These findings do not imply that cefprozil is ineffective against all penicillin-nonsusceptible pneumococci and H. influenzae. Rather the data indicate that the physician must exercise caution in prescribing cefprozil routinely for patients who have failed previous therapy for AOM. This is because the agent would be expected to produce bacteriologic eradication in approximately one-third of patients with AOM caused by penicillin-intermediate resistant pneumococci and in very few, if any, with disease caused by penicillin-resistant pneumococci and by H. influenzae. Although clinical response to cefprozil therapy in these would be expected to be somewhat better than the bacteriologic response,9–11 it is our conclusion that cefprozil is not effective for treatment of AOM, especially in those patients with refractory or recurrent disease. Acknowledgment. The study was made possible by a grant from Bristol-Myers Squibb Co.