Fatigue behavior of ultrafine tabletop ceramic restorations

The goal of this study was to investigate the fatigue life, failure modes, and stress distribution of partial ultrafine restorations for posterior teeth in different ceramics.

Sixty standard tabletop preparations in epoxy resin G10 received lithium-silicate-based zirconia-reinforced (ZLS) or hybrid ceramic (PIC) restorations in 0.5- or 1-mm thickness bonded with resin cement. The same cycling protocol was applied for all specimens, which consisted of 5000 cycles at 200N, followed by 450-N cycles until the specimens' fracture or the suspension of the test after 1.5×10⁶ cycles. Axial load was carried out with a 4Hz frequency in Biocycle V2 equipment (Biopdi, São Carlos, SP), with samples immersed in water. The presence of cracks and/or fractures was checked every 2.5×10⁵ cycles, and the survival analysis was performed with the number of cycles in which each specimen failed. All specimens were evaluated by stereomicroscopy and scanning electron microscopy (SEM). After data tabulation, Kaplan-Meier and Mantel-Cox (log-rank test) analyses were performed, followed by multiple pairwise comparison, all with a significance level of 5%, and Weibull analysis. Through three-dimensional finite element analysis, stress distribution and maximum principal stresses in the posterior occlusal veneers were evaluated by comparison of different types of substrate (G10, enamel/dentin, enamel), thicknesses, and ceramic materials.

Zirconium-reinforced lithium silicate restorations with 0.5-mm thickness (ZLS.5) showed lower fatigue strength compared with that of 1.0-mm hybrid ceramic restorations (PIC1), and both were similar to other restorations (PIC.5 and ZLS1) (log-rank test, χ²=11.2; df=3; p=0.0107<0.05). ZLS groups presented random defects that culminated in fracture, whereas PIC groups presented defects that increased with mechanical fatigue after some cycling time. Stereomicroscope images show radial cracks due to the translucency of the material. There was no damage caused by the applicator. MPS (maximum principal stress) distributions were similar for the different substrate types, but the highest modulus of elasticity showed slightly lower stress concentration.

PIC is more likely to be used in thinner thickness than indicated by the manufacturer, with fatigue strength similar to that of thicker ZLS restorations.