We read with interest the article “Influence of orthodontic mini-implant penetration of the maxillary sinus in the infrazygomatic crest region” in the May 2018 issue of the AJO-DO , and we believe that this topic is relevant for current orthodontic practice. We want to point out some misunderstandings in the methodologic section and conclusions.

First, the authors demonstrated that a sample size of 60 mini-implants would be powerful enough to detect mini-implant penetration into the maxillary sinus. Although the initial sample included 60 mini-implants, the statistical calculus for membrane thickening was based on 25. We also observed a lack of important exclusion criteria, such as images suggesting sinus inflammation or pathology before mini-implant insertion.

In relation to the statistics, the purpose of interobserver agreement and why it was done with the paired t test were not clear. Means and standard deviations were presented, although the statistical tests used nonparametric tests. Nowadays, it is suggested that P values should be complemented by other statistical parameters such as effect size or confidence intervals, and this would be interesting to confirm the increase and relevance of Schneiderian membrane thickness. We believe that the term “incidence” repeated throughout the text could be replaced in some circumstances by “prevalence,” especially to characterize the status of the penetration of the mini-implant into the maxillary sinus involving the whole sample. “Incidence” needs a longitudinal design with a limited sample size.

Concerning the tomographic images, we must point out that CBCT is a limited method to evaluate the Schneiderian membrane thickness, because the image does not allow differentiation between liquids and soft tissues, and tomography assessment of the thickness can be 2.6 times higher than the histologic examination. Furthermore, there is no information about the voxel size, and that is important in studies involving tomographic images, because of the implications on image quality and resolution. The Schneiderian membrane is a thin structure and requires an appropriate image to better diagnose its thickness. Also, it was not clear to us how many tomographs were obtained for each patient. The article mentions that 1 inclusion criterion for subject selection was the “CBCT performed just before removal of the mini-implant anchorage (with or without CBCT scans immediately after mini-implant insertion).” In another part, the authors mentioned that “CBCT scans were carefully observed before mini-implant placement to select the preferred insertion site and direction in the infrazygomatic crest.” It appears that for some patients 3 scans were performed at different times related to mini-implants.

Finally, it should be noted that despite the high “incidence” (78.3%) of mini-implant penetration in the infrazygomatic crest, there was an overall success rate (96.7%) higher than in previous studies with mini-implants installed in different sites. Jia et al well known that clinical success depends on initial mechanical stability and proper loading. It was highlighted that no patient complained of clinical symptoms; there was no detected mobility, and all patients had stable anchoring performance until the end. Several studies have reported that the perforation of the maxillary sinus membrane during miniplate and dental implant placement is not significant in causing postsurgical complications. Therefore, the conclusions could be more cautious about the problem of perforation of the mini-implants more than 1 mm. What is the clinical relevance for the failure rate or the clinical symptoms, especially considering a temporary device?