Proof testing to improve the reliability and lifetime of ceramic dental prostheses

Ceramic dental prostheses exhibit increasing failure rates with service time. In particular, veneered crowns and bridges are susceptible to chipping and other fracture modes of failure. The purpose of this manuscript is to introduce a computational methodology and associated software that can predict the time-dependent probability of failure for ceramic prostheses and subsequently design proof test protocols to significantly enhance their reliabilities and lifetimes.

Transient reliability and corresponding proof testing theories are introduced. These theories are coded in the Ceramic Analysis and Reliability Evaluation of Structures (CARES/Life) code. This software will be used to demonstrate the predictive capability of the theory as well as its use in designing proof test protocols to significantly improve the reliability (survival probability) and lifetime for dental prostheses. A three-unit fixed dental prosthesis (FDP) with zirconia core (ZirCAD) and veneering ceramic (ZirPress) are used to compare the predicted probabilities of failure to general clinical results. In addition, the capability to use proof testing to significantly improve the performance (reliability and lifetime) for this restoration is demonstrated.

The probability of failure, P_f, after five years without proof testing is predicted to be 0.337. This compares to clinical studies showing the failure rate to be between 0.2 and 0.23 after 5 years. After 10 years, reference 18 found the clinical failure rate for similar bridges (but not the same) to be up to 0.28 compared to the predicted P_f of 0.38. The difference may be due to the analysis applying the load at an inclination of 75° which is more critical than vertical loading. In addition, clinical studies often report a simple survival rate instead of using Kaplan-Meier analysis to properly account for late enrollees. Therefore, true clinical failure rates may be higher than reported and may more closely match the predictions of this manuscript. The effectiveness of proof testing increases with selecting materials less susceptible to slow crack growth (higher SCG exponent, N). For example, proof testing the ZirPress glass-veneered bridges with N = 43.4 analyzed in this manuscript at 400 N bite force for 1 s which induces a failure rate during proof testing of 0.31, reduces the P_f of bridges not proof tested from 0.45 to an attenuated-proof-tested probability of failure P_fa of 0.21 after 20 years of usage. If another material is selected with improved resistance to SCG of N = 60 and the same loading conditions, the failure rate for the proof tested bridges after 20 years of service drops to 2 in 10,000 from 2.4 in 100 had they been not proof tested. The failure rate during proof testing for this material would be 5.1 in 100. Proof testing a material with absolutely no susceptibility to SCG at the same service load (in this case 285 N, not even the 400 N load used above) results in 0 % failure rate and is of course independent of time.

The transient reliability and proof test theory presented in this paper and associated computational software CARES/Life were successful in predicting the performance of ceramic dental restorations when compared to clinical data. Well-designed proof test protocols combined with proper material selection can significantly enhance the reliabilities and lifetimes of ceramic prostheses. This proof test capability can be a translational technology if properly applied to dental restorations.