Crown Retention System on Stress Concentration of Posterior Implant-Supported Crowns

Objectives: This study evaluated the mechanical behavior of implant-supported crowns with different retention systems, using a three-dimensional (3D) finite element analysis.
Methods: A 3D model of a Morse-taper implant was designed (4.1 x 10 mm) with 3 mm of exposed threads according to the ISO 14801. The model was replicated in four groups with similar clinical indications to replace a maxillary left first molar. Four conditions were considered according to the crown-abutment retention (Cement retained [CR], Screw retained [SR], Lateral screw retained [LSR] and Modified lateral-screw-retained [MLSR]). For designing the occlusal contact, a tripoidism concept with 600 N load was applied in an axial (Z-axis) and oblique (30°) directions. Von-Mises stress (MPa) was used as analysis criterium and results were plotted in stress maps for qualitative comparison. The stress peaks in each structure were used for quantitative comparison.
Results: Regardless of the loading and retention system, the stress concentration occurred mainly on the cervical region. Thereby, the stress did not concentrate in the bulk of the crowns. However, LSR and MLSR models presented the highest stress concentration at the crown margin. The oblique loading increased the stress magnitude for all models, decreasing the differences between the groups. LSR showed the highest stress between crown and abutment joint. CR and SR behaved similarly with reduced stress in the crown margin and higher stress in the prosthetic screw while LSR and MLSR impaired the crown margin stress level and dampened the stress at the screw. The retention system influenced the implant-supported crowns mechanical behavior.
Conclusions: Lateral screw-based retention systems reduces the stress at the vertical prosthetic screw and implant; however, it increases stress formation in the crown and in the lateral screw. Non-axial loading showed a condition more prone to failure at the cervical level for all evaluated models.