Retentive strength of two-piece CAD/CAM zirconia implant abutments

The purpose of this study is to evaluate the retention of two-piece computer-aided design (CAD)/computer aided manufacturing (CAM) zirconia abutments after artificial aging under simulated oral conditions using three different types of resin-based luting agents.

Twenty-one CAD/CAM-generated zirconia copings (CERCON Compartis, Degudent, Hanau, Germany) were bonded to a prefabricated secondary titanium implant insert (XiVE Ti-Base, Dentsply Friadent, Mannheim, Germany), using three different types of resin-based luting agents: group A: Panavia 21 (Kuraray Co, Kurashiki, Japan); group B: Multilink Implant (Ivoclar Vivadent, Schaan, Liechtenstein); and group C: SmartCem2 (Dentsply DeTrey, Konstanz, Germany). The bonding surfaces of the titanium inserts and the zirconia ceramic copings were air-abraded and cleaned in alcohol. All specimens were stored in distilled water for 60 days and subsequently thermal-cycled 15,000 times (5-55 °C). The dislodging force of the copings along the long axis of the implant/abutment complex was recorded using a universal testing machine with 2 mm/min crosshead speed. Data were analyzed descriptively and by performing the Kruskal-Wallis test.

The mean retention values were 924.93 ± 363.31 N for Panavia 21, 878.05 ± 208.33 N for Multilink Implant, and 650.77 ± 174.92 N for SmartCem2. The Kruskal-Wallis test indicated no significant difference between the retention values of the tested luting agents (p = 0.1314). The failure modes of all tested two-piece abutments were completely adhesive, leaving the detached zirconia coping and titanium insert undamaged.

The use of resin-based luting agents in combination with air abrasion of titanium inserts and zirconia copings led to a stable retention of two-piece CAD/CAM abutments. The bonding stability of the investigated luting agents exceeded the general limits of fracture resistance of two-piece zirconia abutments. A notable difference between the mean retention values of the tested bond materials was shown. However, the statistical analysis revealed that this difference was not significant.