Effectiveness of dentist-prescribed, home-applied tooth whitening. A meta analysis.

INTRODUCTION Common clinical experience suggests that tooth whitening agents are 100% effective. This study uses meta-analysis of data from published randomized controlled clinical trials to determine the efficacy of tooth whitening agents. METHODS A MEDLINE search strategy was developed and implemented to systematically identify clinical trials on dentist-prescribed, home-applied tooth whitening agents, using 10% carbamide peroxide, published between 1989-1999. Inclusion criteria (e.g., in English, human clinical trials) and exclusion criteria (e.g., not placebo controlled) were established and clinical trials that met these criteria were critically appraised for validity and clinical applicability. Meta-analysis was then used to quantitatively integrate the findings. RESULTS Seven studies were identified that met the inclusion and validity criteria. These studies indicated that: Whitening results in a significant mean change of 6 4 shade guide units (p < 0.01), while the placebo control group exhibited little change (0.7 0.6, p > 0.05). 93% of the bleached patients exhibited 2 shade guide unit change, while 20% of the placebo control group exhibited this change. The brand of bleaching agent had a significant effect on tooth whitening, but the daily application time and duration of treatment did not. Whitening is maintained for 6 months for 1/2 of the people treated. Neither gingival indices nor plaque indices were adversely or favorably affected by bleaching. CLINICAL APPLICABILITY: The data from the reviewed studies indicate that rather than being 100% effective, on average: 73% (93% for bleached group minus 20% placebo group) of people who whiten their teeth will exhibit a whitening that is 2 shade guide units greater than the placebo. 20% of the people who use dentist-prescribed, home-applied bleaching will achieve a mean whitening effect of 5 shade guide units. Re-treatment for 50% of people may be necessary to maintain this effect longer than 6 months. The methods used here are Internet applicable for other clinical topics.