Objective: To demonstrate the use of meta analysis for the comparison of products across a variety of in vitro laboratory models.
Methods: Two silica sodium fluoride dentifrices were studied using variations of published models: Enamel Fluoride Uptake (EFU), Enamel Solubility Reduction (ESR), Cycling Protection (CP) and Cycling Remineralisation/ Demineralisation (RD). The Fluoride content within each study was matched at 1100ppm or 1400ppm F. The CP and RD studies measured either surface microhardess (VHN) or lesion volume. Meta analysis was performed on the adjusted means and variance estimates calculated for each formulation within each study. This data was combined using the method described in Whitehead and Whitehead for normal data. A random effects model was used to allow for varying treatment effects from study to study. Where the meta analysis included different measures of the variable of interest, a standardised difference approach was used. The analysis involved individual analysis and a combined global analysis of EFU, ESR, CP and RD studies.
Results: Summary of results of Meta Analysis
Variable | N | Differencea | 95% CI for difference | P-Value | |
Enamel Fluoride Uptake (EFU)b | 14 | 0.911 | 0.407 | 1.414 | 0.0004 |
Enamel Solubility Reduction (ESR) (%) | 5 | 12.172 | 5.997 | 18.346 | 0.0001 |
Cycling Protection (CP)b | 1 | -0.233 | -0.741 | 0.275 | 0.3679 |
CP sensitivity analysisbc | 4 | 0.270 | -0.129 | 0.708 | 0.1752 |
Remin/Demin (RD) (VHN) | 4 | 4.503 | 1.561 | 7.446 | 0.0027 |
Overall Fluoride Efficacyb | 20 | 0.935 | 0.534 | 1.337 | <0.0001 |
Overall Fluoride Efficacy sensitivity analysisbc | 23 | 0.870 | 0.525 | 1.214 | <0.0001 |
N=number of studies. VHN=Vickers Hardness number a Difference is New minus current formulation. A positive difference favours new formulation b Standardised differences c Sensitivity analysis.
Conclusion: Meta analysis can be used to evaluate laboratory data across a variety of studies. There was an overall statistically significant difference between the fluoride performance of the two formulations compared.