Clinical and Micromorphologic 29-year Results of Posterior Composite Restorations.

Prospective clinical studies of composite restorations revealed their safety and longevity; however, studies did not elucidate the dynamic mechanisms of deterioration caused by fractures and secondary caries. Therefore, the aims of this 29-y controlled study were 1) to follow up on the clinical behavior of posterior composite restorations annually and 2) to compare clinical outcomes with micromorphologic scanning electron microscopy features. After ethical approval, the single-arm study commenced in 1987 with 194 class I or II primary posterior composite restorations with glass ionomer cement providing pulp protection. Each restoration was evaluated annually for 15 y and then again at 29 y per the US Public Health Service-compatible Clinical, Photographic and Micromorphologic coding index, with clinical and photographic criteria for anatomic form, color matching, surface quality, wear, marginal integrity, secondary caries, and clinical acceptability. Parallel micromorphologic criteria were applied at baseline and after 1, 3, 5, 7, 10, 15, and 29 y to assess surface roughness, texture, marginal integrity, fractures, ledges, and marginal gaps with semiquantitative coding and with quantitative 3-dimensional scanning electron microscopy profilometric measurements of marginal grooves next to the enamel, grooves within the bonding zone, and ledges. Statistical analysis included the calculation of the annual failure rate and the use of Kaplan-Meier methodology and nonparametric tests. The cumulative survival rates were 91.7% (6 y), 81.6% (12 y), and 71.4% (29 y). The mean annual failure rate was 1.92%. Significant changes in the restoration-tooth interface from baseline to 5 y resulted in functional masticatory equilibrium. Clinical deterioration year by year, including micromorphologic microfractures and wear, reflected unique dynamic changes in long-term surviving restorations with very low secondary caries and fracture risks (German Network for Health Care Research VfD 29 99 003924).