Retention of graphene-reinforced polymethyl methacrylate crowns bonded to titanium-base abutments: an exploratory in vitro study

To evaluate the effects of resin cement type, chemical surface pretreatment, and their interaction on maximum pull-out force and residual-cement distribution in graphene-reinforced polymethyl methacrylate crowns bonded to titanium-base abutments. Twenty-four CAD-CAM crowns were distributed among four protocols ( n = 6): Multilink Hybrid Abutment or RelyX Unicem 2, each with or without chemical pretreatment. Pretreatment comprised Monobond Plus on titanium and SR Connect on the crown intaglio surface. After cementation, specimens were stored dry for 1 week at 20 °C and underwent axial pull-out testing at 1 mm/min. Maximum pull-out force was reported in N. A 2 × 2 factorial analysis assessed cement, pretreatment, and their interaction; heteroscedasticity-robust HC3 inference was used as a sensitivity analysis. Welch analysis of variance and Games-Howell comparisons were retained as exploratory protocol-level analyses. Residual-cement scores were analysed using Fisher’s exact test (α = 0.05). Mean maximum pull-out force was 452.9 ± 74.0 N for Multilink Hybrid Abutment with pretreatment (Group A), 688.7 ± 185.1 N for RelyX Unicem 2 with pretreatment (Group B), 400.7 ± 94.7 N for Multilink Hybrid Abutment without pretreatment (Group C), and 366.8 ± 82.4 N for RelyX Unicem 2 without pretreatment (Group D). The cement × pretreatment interaction was significant (F = 7.86; P = 0.011; partial η² = 0.282) and remained significant with HC3 robust inference (interaction coefficient, 269.7 N; 95% CI, 49.9 to 489.5 N; P = 0.019). Residual-cement distributions differed among protocols (Fisher exact P < 0.001). The significant cement × pretreatment interaction indicates that the effect of chemical pretreatment depended on the resin cement. The estimated increase associated with pretreatment was 52.2 N for Multilink Hybrid Abutment and 321.9 N for RelyX Unicem 2; these exploratory estimates should be interpreted in light of the small sample size and wide confidence intervals. Within the limitations of this exploratory in vitro study, chemical pretreatment showed a cement-dependent association with maximum pull-out force. The largest increase was observed with RelyX Unicem 2. Pretreatment was also associated with a shift in residual-cement distribution toward mixed failure patterns.