To compare the success and survival rates of metal-ceramic crowns and direct composite resin restorations retained with glass fiber posts, and to identify causes of failure during extended follow-up. A prospective randomized controlled trial (NCT01461239) with parallel groups was conducted. Eighty-three teeth from 68 patients were initially randomized; however, six did not receive the assigned treatment. Thus, 77 teeth from 62 patients (42 composite resin, 35 metal-ceramic crown) with a glass fiber post and at least one remaining wall were analyzed. Standardized procedures were used, with annual follow-ups for up to 14 years. Failures were classified as repairable or non-repairable. Cox regression with shared frailty and Kaplan-Meier curves were used for statistical analysis. The median follow-up was 8.1 years for success and 10.58 years for survival. Overall, 10-year cumulative success and survival rates were 57.7% and 72.2%, respectively. Crowns had higher success (75.2%) and survival (79.4%) rates than composite resin (40.0% and 57.6%). Composite resin showed a significantly higher failure risk (HR = 3.53; p < 0.001), although survival did not differ statistically (p = 0.753). Most failures in composite resins were repairable (70%), while crown failures were mostly non-repairable (70%). At the final follow-up, 13 teeth were lost to follow-up in each group. Metal-ceramic crowns demonstrated superior long-term success and required fewer repairs than composite resin restorations. For endodontically treated teeth, material selection influences outcomes. Metal-ceramic crowns provide longevity, while composite resin restorations offer the advantage of repairability. • Data on long term outcomes of metal-ceramic crowns and composite resin restorations retained with glass fiber posts is scarce. • After up to 14 years of follow-up, crowns showed higher success and survival rates compared to composite resin restorations. • Most failures in composite resin restorations were repairable, whereas failures in crowns were predominantly non-repairable.
Galdino-Santos et al. (Sun,) studied this question.