Several studies have indicated that either the surgical insertion of ventilation tubes (VTs) or long-term treatment with sulfonamide-based antibacterials is effective in the management of otitis media with effusion (OME; otherwise known as serous otitis media, secretory otitis media, and glue ear) when compared with a no-treatment control or placebo. This controlled trial is the first to compare directly the effectiveness of these two treatments for long-standing OME. Outcome variables are treatment success rates, hearing thresholds, recurrent acute otitis media episodes, and side effects of medication or complications of VT placement. One hundred twenty-five children (aged 2.5 to 7 years) who met the usual indications for surgery (long-standing [greater than 3 months] OME and conductive hearing loss) were randomly assigned to "medical" treatment (sulfisoxazole 75 mg/kg per day for 6 months) or "surgical" treatment (bilateral insertion of VTs). Subjects underwent pure-tone audiometry (500, 1000, 2000, 4000 Hz) and otomicroscopic examination at 2, 4, 6, 12, and 18 months. A significantly greater proportion of medical subjects (67%) than surgical subjects (48%) were treatment failures at 6, 12, or 18 months (P = .0208). Surgical subjects had significantly better hearing at 2 and 4 months (P values less than .01) but not at 6, 12, and 18 months (P values greater than .2). A significantly greater proportion of surgical subjects (50%) experienced complications of treatment than did medical subjects (9%) (P less than .001). Thirty-three percent of candidates for VT placement did not require surgery when treated with a 6-month course of sulfisoxazole.(ABSTRACT TRUNCATED AT 250 WORDS)