The included studies basically report on commercially available implants from two major manufacturers and involve comparisons of different implant surface technologies in regard with topography and/or chemistry modifications within each implant system. Various analyses were performed in the included studies; however, an attempt has been made to synthesize the various findings and discuss them herein irrespective of the specific implant systems, based on the assumption that basic biological mechanisms of peri-implant bone wound healing are largely implant system independent.

Of the 242 search identified studies, only four studies were finally included in the review, all focusing on the impact of implant surface on early human peri-implant bone healing (Figure 1; Table 3). Genetic analyses of total RNA isolated from either newly formed peri-implant bone harvested by trephination [34,35] or from cells adherent to implants retrieved by reverse threading [36,37] were performed. In total, 111 implants from 43 patients were analyzed. All four studies reported the use of commercially existing implant surfaces, i.e., either a chemically modified, hydrophilic, sand-blasted, acid-etched surface (SLActive®, Institute Straumann AG, Basel, Switzerland); or a hydrophilic (SLActive®) versus a hydrophobic unmodified SLA® (Institute Straumann AG, Basel, Switzerland) surface; or a micro-topographic titanium-oxide grit-blasted surface (TiOBlast®, AstraTech, Molndal, Sweden) versus a chemically modified nano-topographic grit-blasted surface (Osseospeed®, AstraTech, Molndal, Sweden). Implant retrieval times were at 3 or 4 days and 7 days in all studies and additionally at 14 days in two studies [34,35].

All studies used validated methods for gene expression analysis; genetic data was analyzed using microarray (three studies) or real-time PCR (RT-PCR) (one study) methods (Table 1). Total RNA was isolated from lysates of either trephined peri-implant tissues or implant-adherent cells, and subjected to microarray processing or RT-PCR. Although moderate-to-good agreement has been reported between the two methods, validation of DNA microarray results by the more sensitive PCR array is generally recommended [38]. None of the microarray studies identified have validated their results using RT-PCR. Genotyping data (gene lists) were imported and analyzed using computer software and further condensed into functionally and biologically relevant categories. Nevertheless, differential gene expression in relation to a particular cell type or region of tissue analyzed was not performed [35]. Gene ‘upregulation’ was reported when genes were expressed at a higher level on one implant surface in comparison to another; in context, differentiation between gene expression and over-expression may be difficult to define. Statistical methods were used to compare differences in gene expression between different time points and/or implant surfaces (P < 0.05 significance level), while correcting for possible errors, i.e., false gene discovery rate due to multiple sampling [39]. There was considerable heterogeneity across the included studies in terms of study design, population, implant surface technology, genotyping methods, and data analyses (Table 2). Therefore, no meta-analysis of association between gene expression and implant surface properties was relevant.