Mesenchymal stem cells in maxillary sinus augmentation : a systematic review with meta-analysis

Aim: To investigate the effectiveness of mesenchymal stem cells (MSCs) in maxillary sinus augmentation (MSA), with various scaffold materials. Methods: MEDLINE, EMBASE and SCOPUS were searched using keywords such as sinus graft, MSA, maxillary sinus lift, sinus floor elevation, MSC and cell-based, in different combinations. The searches included full text articles written in English, published over a 10-year period (2004-2014). Inclusion criteria were clinical/radiographic and histologic/ histomorphometric studies in humans and animals, on the use of MSCs in MSA. Meta-analysis was performed only for experimental studies (randomized controlled trials and controlled trials) involving MSA, with an outcome measurement of histologic evaluation with histomorphometric analysis reported. Mean and standard deviation values of newly formed bone from each study were used, and weighted mean values were assessed to account for the difference in the number of subjects among the different studies. To compare the results between the test and the control groups, the differences of regenerated bone in mean and 95% confidence intervals were calculated. Results: Thirty-nine studies (18 animal studies and 21 human studies) published over a 10-year period (between 2004 and 2014) were considered to be eligible for inclusion in the present literature review. These studies demonstrated considerable variation with respect to study type, study design, follow-up, and results. Meta-analysis was performed on 9 studies (7 animal studies and 2 human studies). The weighted mean difference estimate from a random-effect model was 9.5% (95%CI: 3.6%-15.4%), suggesting a positive effect of stem cells on bone regeneration. Heterogeneity was measured by the I (2) index. The formal test confirmed the presence of substantial heterogeneity (I (2) = 83%, P < 0.0001). In attempt to explain the substantial heterogeneity observed, we considered a meta-regression model with publication year, support type (animal vs humans) and follow-up length (8 or 12 wk) as covariates. After adding publication year, support type and follow-up length to the meta-regression model, heterogeneity was no longer significant (I (2) = 33%, P = 0.25). Conclusion: Several studies have demonstrated the potential for cell-based approaches in MSA; further clinical trials are needed to confirm these results.