Factors affecting outcome were analyzed from 3 antibiotic clinical trials that had identical case definition and outcome criteria. Overall 102 subjects with acute otitis media had an initial tympanocentesis, were enrolled in one of the clinical trials, were randomized to receive 10 days of oral treatment and had a posttherapy visit. The antibiotics used were cefixime (38), cefaclor (25), loracarbef (14), amoxicillin plus clavulanate (16) or amoxicillin (9). Fifty-five of the 102 (54%) study subjects were classified as cured or improved at the 21- to 28-day posttherapy visit. Factors analyzed in relation to outcome included antibiotic administered, isolation of a pathogen from the middle ear aspirate, study subject age and sex, history of recurrent otitis media, unilateral vs. bilateral involvement, season of enrollment and history of antibiotic administration in the month before enrollment. Univariate analysis identified the following four factors associated with higher posttherapy visit failure rates: a history of recurrent otitis media; enrollment during winter respiratory season (December through March); a history of being treated with an antibiotic during the month before enrollment; and administration of cefaclor compared with other antibiotics. However, only a history of recurrent otitis media and enrollment during the winter respiratory season met the 0.05 significance level for entry into a model derived from logistic regression to assess interactions among factors. Clinical guidelines for the management of otitis media should take into consideration that children with a prior history or recurrent otitis media and infection during the winter season more often fail to respond to antibiotic treatment and have a higher risk of developing a persistent middle ear effusion.