BACKGROUND The value of tympanometry in detection of middle ear effusion (MEE) has been widely studied in otitis media with effusion. There has been no direct comparison of tympanometric and tympanocentesis (TAP) findings in acute otitis media (AOM). We compared otoscopic, tympanometric and TAP findings in AOM including cases of AOM without effusion. METHODS In a study of AOM treatment of 90 children, a tympanogram was obtained, and TAP was performed before and after 5 days of treatment. Subjects were followed with otoscopy and tympanometry every 2 weeks for 3 months or until the MEE cleared. RESULTS In 130 AOM ears, otoscopic, tympanometric and TAP findings were available; MEE was obtained from 110 ears. Of 20 ears with a dry tap, 18 had abnormal tympanogram and otoscopic findings. With TAP findings as the standard, sensitivity and positive predictive value of type B tympanogram were 97 and 87%, respectively. Of 18 AOM ears yielding dry tap, 2 yielded MEE 5 days later, and 8 continued with evidence of MEE 5 days-12 weeks later. Five subjects with 8 AOM ears yielding dry tap were lost to follow-up; all had short duration of clinical symptoms. CONCLUSIONS Although otoscopic and tympanometric findings suggested the presence of MEE in AOM, MEE was unobtainable by TAP in 14% of cases. Dry tap cases likely represent early AOM before accumulation of detectable MEE. However, technical difficulty in obtaining small amounts of or highly viscous MEE could not be excluded. Sensitivity and positive predictive value of abnormal tympanograms in detection of MEE in AOM cases are comparable with those in otitis media with effusion.