Premium
Industry sponsorship in trials on fluoride varnish or gels for caries prevention
Author(s) -
Reda Seif,
Elhennawy Karim,
MeyerLückel Hendrik,
Paris Sebastian,
Schwendicke Falk
Publication year - 2017
Publication title -
community dentistry and oral epidemiology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.061
H-Index - 101
eISSN - 1600-0528
pISSN - 0301-5661
DOI - 10.1111/cdoe.12287
Subject(s) - fluoride varnish , medicine , fluoride , confounding , dentistry , clinical trial , publication bias , test (biology) , meta analysis , data extraction , varnish , medline , environmental health , inorganic chemistry , paleontology , chemistry , organic chemistry , political science , law , biology , coating
Objectives Fluoride is effective for caries prevention, but trials on fluoride varnish or gels are often industry‐sponsored. We assessed trial design and findings in sponsored and nonsponsored trials on fluoride varnish and fluoride gels for caries prevention. Methods Data on trials included in the most recent Cochrane Reviews on fluoride varnish and fluoride gels were extracted. Sample sizes/age/dentition, year/country of publication, follow‐up, test and control, risk of bias and spin (claims of a beneficial effect that were not supported by reported data) were assessed. Studies were categorized as certainly, possibly and not sponsored, and statistically compared. Inverse‐generic meta‐analysis and multivariable weighted least‐squares meta‐regression were used to assess impact of sponsorship status on effect estimates. Results Based on 19 nonsponsored, 14 possibly sponsored and 11 certainly sponsored trials, sponsored studies were published significantly earlier, always had >1 test group, and had significantly lower risk of spin. Caries‐preventive effects were higher in earlier trials, without indication for sponsorship bias in trials published until 1990 (there were no sponsored trials afterwards). If assessing the overall body of evidence and accounting for confounders, the caries‐preventive effect was significantly associated with year of publication (β: −0.06, 95% CI : −0.10/−0.02), but not sponsorship status. Conclusions Industry‐sponsorship bias had limited impact on the overall evidence.