We thank Professor Mjör for his thoughtful and insightful reading of our article, and for his plausible explanation in his letter. As correctly stated by Professor Mjör, when good penetration and wetting of the dentin by an adhesive into the branches of the tubules occurs, the bond strength between the dentin and the resin will be a good bond. Penetration of the bonding resin in our study into the gingival wall was impeded by wetness that was more prominent on the gingival versus the axial wall.
A follow-up fractographic scanning electron microscopy analysis, which measured the area of voids present on the gingival versus the axial wall under in vivo and in vitro conditions, found under in vivo conditions that 48.8 ± 29.2 percent of the area of the gingival wall compared with 13.6 ± 25.6 percent of the area of the axial wall contained voids, which were probably due to wetness. Under in vitro conditions 11.7 ± 17.6 percent of the gingival wall was found to have voids, compared with 0.0 ± 0.0 percent of the axial wall.
Clearly, penetration under in vivo conditions is more difficult than under in vitro conditions, and penetration at the gingival wall is more difficult than at the axial wall. This could be attributed to curing the material from the top of the matrix band, which is approximately 5.0 millimeters away from the gingival increment, resulting in an incomplete cure, although the cure at the axial wall at this distance did not seem to suffer. This also is seen clinically when practitioners remove failed composite resins in Class II restorations, when the gingival box increment seems to "fly" out so easily.
We would have to reduce the amount of voids at the gingival wall before we can conclude whether or not good adhesion can be attributed to the branches of the tubules, bonding to parallel or perpendicular dentin or for another unknown reason.
