To test if cavity dimensions and restorative protocol have potential to influence in-vitro adaptation of class II restoration after simulated thermo-occlusal stressing.

A total of 32 prepared teeth were randomly assigned to one of the 4 experimental groups depending on cavity size, composite system and filling technique; group 1: small cavity and multi-layered conventional restorative composite (Tetric Evo-ceram: TEC), group 2: small cavity and flowable bulk-filled composite (SDRFlow: SDR) + one single occlusal layer of conventional restorative composite (TEC), group 3: large cavity and multilayered conventional restorative composite (TEC) and group 4: large cavity and bulk-filled flowable composite (SDR) + one single occlusal layer conventional restorative composite (TEC). All specimens were submitted to 500'000 cycles of thermomechanical loading (50 N, 5 to 55 °C). The proximal tooth-restoration interface was analyzed quantitatively by SEM, prior and after thermomechanical loading.

Before loading, continuous enamel adaptation varied from 61.49% (Gr 1) to 68.39% (Gr 4) proximally and from 50.93% (Gr2) to 68.65% (Gr1) cervically, with no statistical difference among groups for both segments. After thermomechanical loading, continuous enamel adaptation varied from 36.6% (Gr2) to 46.6% (Gr1) proximally, without significant difference, and from 20.2% (Gr4) to 51.3% (Gr1) cervically; statistical differences in cervical enamel adaptation were found in-between groups 1 and 2 (p = 0.0479), 1 and 4 (P = 0.0116), 2 and 3 (p = 0.0028) and 3 and 4 (p = 0.001). Before loading, dentin continuous adaptation varied from 55.32% (Gr3) to 81.82% (Gr4) with statistical difference in-between those groups (p = 0.045); after loading, dentin continuous adaptation varied from 31.56% (gr2) to 51% (Gr4) with a statistical difference between those 2 groups (p = 0.019). The drop in adaptations values after loading was significant in all groups and segments.

The impact of the restorative technique and cavity size on marginal adaptation appeared essentially after simulated fatigue for enamel adaptation. Cervical enamel and dentin continuous adaption of small and large bulk-filled restorations dropped significantly while the change was lower in layered restorations made of conventional restorative composite.

Restorations made with bulk-filled flowable composite behaved differently from layered ones using conventional resin composite, according to cavity size and loading.