The marginal depth of the access-hole filling at T = 0, T = 1 M, T = 3 M, and T = 6 M could not be defined successfully because of the overfilling phenomenon which disrupted the measurement of marginal gap depth and angle (Fig. 3a–e). Only the marginal depth and angle at T = 12 M could be measured with the same digital microscope, and the mean value for each group was calculated. Each access-hole was divided into four areas including mesio-buccal, disto-buccal, mesio-palatal and disto-palatal surfaces (Fig. 4) to measure the distance of marginal discrepancy at resin/ceramic interface with a 1 μm accuracy (Fig. 5). With this value, a “marginal discrepancy pattern” could be extrapolated for each group (Fig. 7a, b).

For each access hole, an occlusal photograph was taken at the time of T = 0, T = 1 M, T = 3 M, T = 6 M, and T = 12 M, respectively. The marginal integrity was evaluated and recorded at the baseline and 12 months, according to an esthetical scale VAS (visual analogue scale) described by Weininger et al. [11]. Results were recorded by the authors and summarized in Table 2.