Erosion of dentin results in a complex multi-layered lesion. Several methods have been used to measure erosive substance loss of dentin, but were found to have only limited agreement, in parts because they assess different structural parameters. The present study compared the agreement of four different methods (transversal microradiography [TMR], Confocal Laser Scanning Microscopy [CLSM], Laser Profilometry [LPM] and modified Knoop Hardness measurement [KHM]) to measure erosive substance loss in vitro. Ninety-six dentin specimens were prepared from bovine roots, embedded, ground, polished and covered with nail-varnish except for an experimental window. Erosion was performed for 1 h using citric acid concentrations of 0.00% (control), 0.07%, 0.25% and 1.00% (n = 24/group). Adjacent surfaces served as sound reference. Two examiners independently determined the substance loss. After 1 h erosion with 1% citric acid solution, substance losses (mean±SD) of 12.0±1.3 µm (TMR), 2.9±1.3 µm (LPM), 3.9±1.3 µm (KHM) and 17.0±2.6 µm (CLSM) were detected. ROC curve analysis found all methods to have high accuracy for discriminating different degrees of erosive substance loss (AUC 0.83–1.00). Stepwise discriminatory analysis found TMR and CLSM to have the highest discriminatory power. All methods showed significant relative and proportional bias (p<0.001). The smallest albeit significant disagreement was found between LPM and KHM. No significant inter-rater bias was detected except for KHM. LPM is prone to underestimate erosive loss, possibly due to detection of the organic surface layer. KHM was not found suitable to measure erosive loss in dentin. TMR and CLSM detected the loss of mineralised tissue, showed high reliability, and had the highest discriminatory power. Different methods might be suitable to measure different structural parameters.
