AMOXICILLIN OTITIS MEDIA 'NEGATIVE' STUDY PUBLISHED IN JAMA; META-ANALYSIS OF OTITIS DRUG DATA UNDERWAY WITH AHCPR LOOKING AT GENERAL TREATMENT ISSUES

The publication of a four-year-old article concluding that amoxicillin is not efficacious in the treatment of otitis media with effusion should make that data part of a larger meta-analysis of otitis media treatments, Journal of the American Medical Association Deputy Editor Drummond Rennie, MD, wrote in the Dec. 18 JAMA. "A meta-analysis of trials of therapy [for otitis media] is now in progress," Rennie observed. "It is obvious that unbiased meta-analysis depends on access to the results of all trials." One ongoing large review of otitis media treatments is being conducted by the American Academy of Pediatrics for the outcomes research arm of the federal government, HHS' Agency for Health Care Policy Research ("The Pink Sheet" Dec. 9, T&G-12). AHCPR's interest in otitis media reaches beyond drug therapy. The Agency is also contributing to a National Institute of Child Health and Human Development-sponsored trial which will study the disease's natural history and the use of surgically-implanted tubes. The negative efficacy results, expressed in an article by Erdem Cantekin, PhD, in the Dec. 18 JAMA, have been discussed publicly for four years as part of a controversy over publication rights and alleged bias at the Pittsburgh Children's Hospital. Amoxicillin "is not effective for the treatment of persistent asymptomatic middle-ear effusions in infants and children," Cantekin's article concludes. The Cantekin report is a reanalysis of data from a Pittsburgh Children's Hospital-Otitis Media Research Center (OMRC) four-week efficacy study of amoxicillin. The study, published in the New England Journal of Medicine in 1987, supported efficacy of the drug treatment. Cantekin was an investigator in that study. His dissent from his colleagues led to charges of scientific misconduct and a National Institutes of Health investigation of Cantekin and the lead author of the NEJM article, Charles Bluestone, MD. Cantekin's charges against Bluestone were discussed during a 1988 hearing before Rep. Weiss' (D-N.Y.) House Government Operations Subcommittee as an example of a potential compromising of research results by pharmaceutical company dollars. Cantekin reported that his analysis of the Pittsburgh data on 514 children showed that resolution of effusion was 13.6% for amoxicillin and 11.3% for placebo, not a statistically significant difference. He further claimed that the amoxicillin group showed a two- to six-fold increase in recurrence of effusion. Cantekin's article argues that the original analysis was more favorable for the drug because the authors of the paper used a more subjective algorithm to determine a cure than the one in the trial protocol. Cantekin maintained that the previous paper overcounted "cures" even using its own algorithm. Under his analysis, only with the use of otoscopy can a significant difference in "cure" rates be seen, while all "objective" measures show no difference. Cantekin suggested further that the trial had an unusually low placebo cure rate, which he pins on a higher prevalence of poor prognostic indicators in the placebo group. A follow-up trial conducted by the OMRC between 1984 and 1987 sought to evaluate amoxicillin, Pediazole (Abbott's erythromycin suspension) and cefaclor (Lilly's Ceclor) versus placebo. "The OMRC discovered that the placebo group and the three antibiotic- treated groups had virtually the same cure rates at the four-week end point." Cantekin added that "recent publications indicate that antimicrobial treatment is not effective even for acute otitis media" as opposed to secretory otitis media. "Our re-examination of the study . . . is consistent with those recent findings."