Objectives . Worldwide bond strength tests are used to evaluate bonding effectiveness of different adhesive techniques. The purpose of this study was to systematically collect these data to identify the primary parameters that affect the outcome of bond strength tests, and to attempt to disclose trends in adhesive performance of the different adhesive approaches today available.

Materials and Methods . 871 studies were identified by entering the search term ‘ dentin bond strength AND “published last 5 years”[Filter] ’ in PubMed, and subsequently inserted in a custom-made database. Only studies containing data on the adhesion of several commercial adhesives to dentin were retained. Besides the actual bond strength data (mean, SD, n and number of pre-testing failures), other relevant parameters were recorded with regard to the substrate bonded to, surface preparation, storage conditions, potential artificial ageing, alternative application techniques, type of composite used, and some specific test conditions.

Results . Up to now, 328 studies have been evaluated and from 110 studies relevant data were extracted, yielding 961 individual bond strength tests. A relatively simple statistical general linear model was generated to identify relevant parameters. The bond strength test used most often was the micro-tensile bond strength test (μTBS: 67%), followed by the macro-shear (MSB: 18%), micro-shear (μSB: 9%), macro-tensile (MTB: 4%) and push-out bond strength test (PoB: 2%). For further analysis, the push-out bond strengths and the adhesives that had been tested in less than 3 studies were excluded. From the remaining data, 56% involved any kind of experimental parameter as artificial aging, altered application techniques, etc. The actual bond strength obtained was significantly affected by the type of test employed (on average 29.2 and 18.1 MPa for μTBS and MSB respectively, p < 0.0001). Significant differences were also observed for the different adhesive classes (pooled data): three-step etch&rinse (mean ± 95% confidence interval: 36.9 ± 5.8, n = 44;), two-step etch&rinse (26.3 ± 1.7, n = 178), two-step self-etch (32.3 ± 2, n = 163) and one-step self-etch adhesives (23.6 ± 1.4, n = 253). Also between individual adhesives substantial differences were observed in the mean bond strengths, ranging from 49 MPa (OptiBond Fl, Kerr) to 10 MPa (Futurabond NR, Voco). About 36% of the groups involved any kind of artificial aging and this affected the outcome significantly (22.5 MPa versus 27 MPa for the control specimens, p < 0.001). Another remarkable parameter was the flexural modulus of the composite used; a lower flexural modulus was significantly correlated with lower bond strengths.

Conclusions . Although preliminary, this systematic analysis of the bond strength data in literature revealed some key parameters that affect the outcome of bond strength tests.