OBJECTIVE To assess the clinical outcome and risk of failure after oral vs. intravenous treatment in otitis media caused by penicillin-resistant pneumococci. To determine the possible correlations between pneumococcal minimal inhibitory concentration (MIC) to penicillin and clinical outcome. METHODS Retrospective study of 156 cases collected between 1993 and 1995. Mean follow-up: 5 months. Setting. Two tertiary academic medical centers in Paris, France. METHODS Pneumococcus was isolated from 191 of 570 ear samples obtained from children with otitis media and shown to be penicillin-resistant in 156. Medical history, antibiotic therapy during the previous 3 months and day-care center attendance were reviewed. For the current episode microbiologic characteristics of the isolated strains, type of treatment, therapy efficacy and clinical outcome were analyzed. Patients were predominantly young (76.3% were <1 year old) and bacteriologic samples were taken mainly because of previous treatment failure. RESULTS Among 156 children with pneumococcal penicillin-resistant otitis media, 72.2% attended day-care centers, 71.8% had been previously treated with aminopenicillin and 52.5% with cephalosporins. Failure of previous empirical oral therapy was noted in 84% (one-third of these had been receiving amoxicillin-clavulanate). Patients treated intravenously had had a more protracted otitis but no greater number of previous episodes of acute otitis media than those receiving oral therapy. Acute mastoiditis occurred in 4 infants resulting in mastoidectomy. Oral treatment (mainly with high dose amoxicillin,120 to 150 mg/kg/day) and intravenous therapy (cephalosporin or glycopeptide) had been used in 59 and 41%, respectively. Mean duration of therapy was 10.7 days. Three failures (1.9%) and 10 recurrences (6.4%, average 28 days) occurred. No statistical difference was found between intravenous and oral therapy with respect to risk of recurrence. A high penicillin MIC value was correlated with previous antibiotic treatment but not with clinical outcome. CONCLUSIONS Oral therapy appears to be as effective as intravenous therapy for the treatment of penicillin-resistant pneumococcal otitis media. Intravenous treatment should not necessarily be dictated by the penicillin susceptibility value but should be considered in cases of failure to thrive, persistent otitis or other complications.