No analysis was performed if there was a possible difference in outcomes between the different centers. One reason is that the number of patients was not equally divided between the centers, making exploration for significant differences hardly reasonable. Next to this, randomization was performed using a block randomization sequence to provide equal distribution of subjects treated with 6 mm or 11 mm implants at each center. Also at each center not only the same protocol for surgery and prosthetics was prescribed, but also for aftercare. Therefore, the authors think that a priori there is no reason to suspect that there will be any qualitative differences between the centers regarding any of the efficacy variables nor regarding the safety variables.

Implant survival is also a relevant variable in comparison of clinical performance of dental implants and could also have been used as primary outcome. The authors have chosen to use peri‐implant bone level changes as primary outcome, because it is a predictor for possible implant loss in the future. In this way, in an earlier stadium significant differences can be calculated and conclusions drawn.

The aim of the study was to compare short 6‐mm implants with conventional 11‐mm implants. However, the results from this study cannot be directly compared with most of the other studies on short implants, since most studies have been focusing on short implants restored with single crowns aiming at evaluating the more challenging clinical situations. Thus, we cannot, based on our study results, determine if short implants will perform equally, as standard‐length implants for single tooth restorations. Previously, it was recommended to splint restorations to distribute forces, especially with shorter implants (Guichet et al., 2002). However, in a more recent study, comparing splinted and non‐splinted restorations, no significant differences were found (Vigolo & Zaccaria, 2010). In the studies of Thoma et al. (2018) and Guljé et al. (2019) non‐splinted single tooth restorations were made, showing the same promising results, as in the present study. One proposed advantage of non‐splinted restorations is to provide a better approach to oral hygiene. Another limitation of the study design is that the 6‐mm implants were inserted in minimally resorbed edentulous spaces in the posterior maxilla and mandible, meaning that the results cannot be extrapolated to extremely resorbed posterior edentulous spaces. The present data did not find any correlations between crown‐to‐implant ratio and implant survival or marginal bone level changes (data not shown). However, it is important to note that since the implants were placed in minimally resorbed alveolar bone, the implants were restored with prostheses with relatively short clinical crown height spaces.

A strength of the present study is its design. Being a prospective, randomized multicenter study, the results are regarded to be relatively high clinical evidence, supporting the validity of the measured outcomes. Another strength is the large number of patients taking part in the study, generating data with high statistical power.