Meta-analysis for the effect of medical therapy vs. placebo on recovery of idiopathic sudden hearing loss

OBJECTIVES: To estimate the effect of recovery of idiopathic sudden hearing loss under placebo (first aim) and under medical therapy (second aim).

STUDY DESIGN: Systematic review and meta-analysis.

METHODS: A total of 1,674 studies published between January 1974 and April 2009 were found following suggestions in the Cochrane Handbook for Systematic Reviews. After filtering by criteria of Cochrane Collaboration, four trials remained for continuous and two for dichotomous data.

RESULTS: Using Review Manager, weighted mean difference as well as standardized mean effect of hearing recovery were calculated and pooled. The values for weighted mean difference of hearing gain in dB were 0.79, 95% confidence interval (CI) (-2.04-3.61) and for standardized mean effect 0.06, 95% CI (-0.13-0.24), respectively, which computationally favors active treatment, but statistically is not significantly different from no effect (0 dB). This was in accordance to the comparison of descriptive means between recovery under placebo with 14.3 dB and active treatment with 15.8 dB hearing gain. Treatment effect of dichotomous data (hearing gain vs. no hearing gain) suggested a statistically significant better outcome for active treatment; the odds ratio (OR) [fixed] is 2.18 (1.06-4.46).

CONCLUSIONS: In five different statistical analysis methods used, treatment effect of medical therapy was slightly better than recovery under placebo in which spontaneous recovery could be assumed, but no significant effect was detected. Against the background of recovery under placebo of 14.3 dB vs. 15.8 dB hearing gain of active treatment as averages of all measured frequencies, recovery under placebo seems not to have worse outcome than recovery under medical therapy.