One independent oral and maxillofacial surgeon performed the clinical follow-up examinations according to a standardized protocol. A peri-implant probing including probing pocket depths and recessions on four sites of each implant was assessed as was bleeding on probing (BOP) to determine the status of oral hygiene objectively. Signs of gingivitis and pus suppuration were also recorded. The criteria of peri-implantitis were based on those published by Ong et al. [15]: peri-implant probing depth ≥ 5 mm and bleeding on probing and/or suppuration and radiographic bone loss ≥ 2.5 mm.

Based on panoramic radiographs, marginal bone levels were measured on the distal and mesial sites of each implant. Bone loss was calculated based on the known implant length and the radiographic magnification factors accordingly. Distances were measured to the nearest millimeter. Bone levels after sinus floor elevation were compared to bone levels in follow-up (Fig. 1).

The implant success rate was defined as the absence of patients’ complaints and objective signs of peri-implant inflammation (bleeding on probing, peri-implantitis, dehiscence defects, and implant stability).

Statistical data analyses were performed using GraphPad Prism version 6.0 (GraphPad Software, La Jolla, CA, USA). Descriptive statistics (mean value, standard deviation, and percentage distribution) were calculated, and the data were checked for Gaussian distribution applying the Shapiro-Wilk test. Comparisons between the groups with and without perforation were assessed with non-parametric statistic testing (Mann-Whitney-U-Wilcoxon). Fisher’s exact test was applied for combinations of factors, and implant survival was displayed in a Kaplan-Meier plot. Multiple comparisons according to Tukey were applied in cases of further subdivision of the datasets. If the probability of error was less than 5%, the result was presented as statistically significant.