Caries Risk Assessment Diagnostic Accuracy of Clinical Methods

Objectives: Improved evidence is anticipated to emerge from integration of results of studies that present an appropriate analysis of diagnostic accuracy, which until now was performed in few studies. Additionally, study quality, i.e. whether studies generate unbiased results, was rarely evaluated. The objective was to integrate results on diagnostic accuracy of clinical methods used to identify individuals with increased risk of developing coronal dental caries and at the same time assess quality of included studies.
Methods: This systematic review adhered to the PRISMA Statement. Clinical methods analysed were: previous caries experience, tests using microbiota, buffering capacity, salivary flow rate, oral hygiene, dietary habits and sociodemographic variables. In this review, as one of the first in dentistry, risk of bias in 17 included studies was assessed with QUADAS-2, while simultaneously paying careful attention to other methodological issues developed for systematic reviews. Study characteristics and 2 x 2 tables to calculate sensitivity, specificity, predictive values, and likelihood ratios (LR) were tabulated.
Results: Risk of bias was high or unclear as reporting was below acceptable level for methodological aspects in all studies. There was heterogeneity between studies of previous caries experience and diagnostic accuracy depended on thresholds and age of sample. Tests with mutans streptococci or lactobacilli resulted in low sensitivity and high specificity. For children with primary teeth, pooled LR for a positive test was 3 for previous caries experience and 4 for salivary mutans streptococci (threshold ≥10⁵ CFU/ml). Evidence regarding other methods was lacking.
Conclusions: Based on this review, we cannot conclude that previous caries experience is superior to tests with salivary mutans streptococci. As there were methodological limitations across included studies, the results should be interpreted cautiously. There is a need for further validation studies and more widespread dissemination and implementation of STARD. Also, it’s important to develop novel approaches with new candidate biomarkers.