For volume assessment of abrasion, each 3D data set was segmented before and after CS in CTAn (CTAnalyzer V.1.15.4.0, Bruker microCT). Both data sets were overlapped, and the remaining volume of abrasion quantified in pixels and converted into cubic millimeters.

The maximum wear depth was determined by “blowing up” virtual bullets within the surface of abrasion. The diameter of the most massive bullet (at the spot of maximum wear depth) was measured in pixels and converted into millimeters.

The arrow in Fig. 3 shows the maximum wear depth after CS. Volume wear is demonstrated as a yellow surface. Descriptive statistical analysis was applied.

In addition to quantifying wear behavior, one specimen from the test group was randomly selected for analyzing qualitative wear behavior with a scanning electron microscope (SEM, REM, Phenom™ G2 pro, Phenom-World). Before SEM imaging, the crown was gold-coated (2 nm, Sputter Coater MSC1, Ingenieurbüro Peter Liebscher, Wetzlar, Germany) to prevent accumulation of electrostatic charge.

The crowns were removed from the implants using a universal testing machine (Z010, Zwick GmbH&Co.KG, Ulm, Germany). To do so, the embedded specimens (crown + implant) were placed in a specially built device and covered with a base metal alloy which was specially created as well. A preload of 1 N was applied vertically to the crown followed by traction of 0.75 mm/min. Load at breakage/removal was recorded. The bond strength from the specimens from both the control group without CS (n = 5) and the experimental group after CS (n = 20) was measured.

Luting agents on both the crown and the implant after CS were localized under a stereomicroscope (Stemi 2000-C, Zeiss, Karlsruhe, Germany). Representative pictures of each crown and implant were taken, and a percentage of luting agent on crown and implant was recorded descriptively.

The statistical analyses were performed using GNU Project (2015) (GNU PSPP (Version 0.8.5) [Computer Software]. Free Software Foundation. Boston, MA). The Kolmogorov-Smirnov test, visual inspection of the distribution of the data in histograms as well as in quantile-quantile plots, was applied to verify if the data were normally distributed. The ANOVA test was used to analyze the differences in the mean level of the four rounds of CS concerning bond strength, maximum vertical wear, and volume wear of the experimental groups. A t-test for independent samples was performed to find differences in bond strength between the experimental and the control group. Student’s t-test was applied assuming no different variants between control and experimental group as no empirical difference of the variances was observed (p = 0.755, Levene’s test). The exact confidence interval was calculated according to Clopper-Pearson.