The database PubMed and the US National Library of Medicine were screened from January 8, 2012, to January 6, 2016, for potential studies reporting on membrane perforations during MSA from 1980 till 2015. The search was conducted independently and in duplicate by two authors (MK and JC). The following search terms were used:

MeSH Terms:

Piezo-surgery

Ultrasound

Ultrasonic Osteotomy

Maxillary Schneiderian Membrane Perforation

Sinus Perforation

Maxillary Sinus Augmentation Complications

Lateral Sinus Osteotomy

Cross-references:

Piezo-surgery AND Sinus Floor Elevation

Piezo-surgery AND Sinus Lift

Piezo-surgery AND Maxillary Sinus Grafting

Piezo AND Maxillary Sinus Lift

Piezoelectric Bone Surgery AND Sinus

Piezo-surgery AND Maxillary Sinus Augmentation Complications

Piezo-surgery AND Schneiderian Membrane Perforation

Ultrasound AND Sinus Augmentation

Ultrasound AND Sinus Lift

Ultrasonic AND Sinus Lift

Ultrasonic Osteotomy AND Sinus

Ultrasonic AND Schneiderian Membrane Perforation

Sinus Elevation AND Conventional

Included were all studies reporting on the amount of membrane perforations during MSA by the lateral approach. Prospective and retrospective cohort studies and case series were also included. Both studies with split-mouth design and also studies without control group were also included. Excluded were studies describing any other procedure than the lateral approach for MSA, missing information on the occurrence of membrane perforation and in vitro studies. Titles and abstracts of the searches were initially screened for possible inclusion. After analysis of the abstracts, full-text evaluation was performed. Any disagreement was resolved by discussion between the authors (Fig. 1).

The following data were extracted:

Study characteristics: title, authors’ name, year of publication, study design, number of sinus floor elevations (SFE)

Interventions: the use of piezoelectric devices or rotary instruments for SFE

Outcomes: number and percentage of membrane perforation

For each study, the sample size was determined and the event rate (e.g. perforation or not) was noted in an Excel sheet. Then, a meta-analysis was performed using the Comprehensive Meta-Analysis (Version 3) (Biostat, Englewood, USA) applying the sample size of each study and the event rate (e.g. perforation of the membrane or not). The software calculates the suggested effect of the operation technique (piezo or conventional) on the specific event (perforation of the membrane). This way, the raw data were weighted on the sample size for significance analysis. Applying the weighted data forest plots were calculated, indicating the weight and the 95% confidence interval. Additionally, the random effect was calculated representing the average of all studies in the respective groups (piezo and conventional).