Methodological quality and risk-of-bias assessments in systematic reviews of treatments for peri-implantitis.

Abstract

The purpose of the present study was to evaluate the methodological quality and risk of bias in systematic reviews (SRs) on the effectiveness of peri-implantitis treatments. We searched four electronic databases: MEDLINE, Web of Science, Cochrane Database of Systematic Reviews, and EMBASE. Previous SRs focusing on peri-implantitis treatment published between 2010 and 2017 were identified. After literature screening, eligible SRs were qualitatively assessed using two validated instruments: Assessing the Methodological Quality of Systematic Reviews (AMSTAR2) and Risk Of Bias In Systematic reviews (ROBIS). The characteristics and findings of SRs are also reported. A total of 23 SRs formed the basis of this study. Of the 23, six included randomized controlled trials (RCTs) only. Overall, the AMSTAR2 assessment revealed three studies with high and six studies with low methodological quality, and all the other SRs were judged as having critically low methodological quality. ROBIS revealed only one Cochrane review with a low risk of bias and the others with a high risk of bias. In particular, the assessment of non-randomized studies (NRSIs), appropriateness of ROB assessment, and meta-analysis did not satisfy the criteria in AMSTAR2 assessment. Furthermore, there were a few SRs that interpreted and discussed the results of risk of bias (ROB) and heterogeneity assessment, together with the impact of treatment. Due to the lack of head-to-head comparisons conducted in RCTs, review authors need to use other sources of evidence, such as clinical control trials (CCTs), cohort studies (CS), clinical research (CR), and animal studies. The end result is the presentation of low-quality evidence, with high ROB. Several SRs conducted network meta-analysis as an alternative to head-to-head conventional meta-analysis of RCTs. We suggest that the best methods to generate, access, and assess evidence in situations where RCT evidence is lacking should be discussed on an urgent basis.