The Application of Statins in the Regeneration of Bone Defects. Systematic Review and Meta-Analysis

This systematic review aims to analyze the effect of the local application of statins in the regeneration of non-periodontal bone defects. A systematic study was conducted with the Pubmed/Medline, Embase, Cochrane Library and Scielo databases for in vivo animal studies published up to and including February 2019. Fifteen articles were included in the analysis. The local application of the drug increased the percentage of new bone formation, bone density, bone healing, bone morphogenetic protein 2, vascular endothelial growth factor, progenitor endothelial cells and osteocalcin. Meta-analyses showed a statistically significant increase in the percentage of new bone formation when animals were treated with local statins, in contrast to the no introduction of filling material or the introduction of polylactic acid, both in an early (4–6 weeks) and in a late period (12 weeks) (mean difference 39.5%, 95% confidence interval: 22.2–56.9, p <0.001; and mean difference 43.3%, 95% confidence interval: 33.6–52.9, p < 0.001, respectively). Basing on the animal model, the local application of statins promotes the healing of critical bone size defects due to its apparent osteogenic and angiogenic effects. However, given the few studies and their heterogenicity, the results should be taken cautiously, and further pilot studies are necessary, with radiological and histological evaluations to translate these results to humans and establish statins’ effect.

Although this is their main application, because mevalonate, the product of HMG-CoA reductase's reaction, is the precursor of many other non-steroidal isoprenoid compounds other than cholesterol; their inhibiting role has pleiotropic effects [1,3,[5][6][7]. Those include the anti-inflammatory, immunomodulatory and antimicrobial properties [3,4]. Statins have also been shown to interfere in the process of bone turnover and regeneration due to their action on different types of cells, including osteoblasts, osteoclasts, endothelial cells and mesenchymal stem cells [2][3][4][5][6][7]. For this reason, they have been studied for the treatment of osteoporosis without finding clear conclusions [3,5,6]. The potential benefit of these drugs in bone regeneration has recently been studied [2,3,5,7]. full texts of the 23 remaining articles were read, discarding two in which osteoporosis was induced to animals; one as no bone defect was created; another in which statins were administered transdermally; two as the statins' matrix was bone graft; one in which the drug was introduced at the same time as tricalcium phosphate; and another one because filling material was injected near the defect, but not inside it.
Following the selection process, this review analyzed a total of 15 in vivo animal studies evaluating the effect of local application of statins in the regeneration of induced bone defects (Figures 1-3) ( Table 1). The total animal population was 546 (18 dogs, 36 rabbits and 492 rats), distributed between 287 as controls and 315 in intervention groups, 56 of them being in both groups at the same time with different bone defects treated with different filling material [15,16,22].
Materials 2019, 12, x; doi: FOR PEER REVIEW www.mdpi.com/journal/materials Of the 192 articles initially located in the electronic search, 14 were discarded due to duplication. After reading titles and abstracts, 155 were excluded, as they did not fulfill the selection criteria. The full texts of the 23 remaining articles were read, discarding two in which osteoporosis was induced to animals; one as no bone defect was created; another in which statins were administered transdermally; two as the statins' matrix was bone graft; one in which the drug was introduced at the same time as tricalcium phosphate; and another one because filling material was injected near the defect, but not inside it.
Following the selection process, this review analyzed a total of 15 in vivo animal studies evaluating the effect of local application of statins in the regeneration of induced bone defects (Figures 1-3) ( Table 1). The total animal population was 546 (18 dogs, 36 rabbits and 492 rats), distributed between 287 as controls and 315 in intervention groups, 56 of them being in both groups at the same time with different bone defects treated with different filling material [15,16,22].
The samples were followed for a variable period of time, between 28 and 112 days, with the mean of the studies being 58.9 days. During this time, between one and five analyzes were performed, with two analyzes being the most frequent [8,22].
In all the trials in which the percentage of new bone formation was analyzed, statistically significant results were obtained when comparing the groups in which the filling material contained statins with the control group [8,10,16,18]. Only in two of them did the necessary data appear in order to compare them quantitatively [8,16]. Three other studies measured the reossification area [9,19,21], with statistically significant results in two of them [19,21]. In the only work in which the ratio of new bone formation was analyzed, the results were statistically significant [20].
Three [10,13,18] of the four studies [9,10,13,18] that analyzed bone density in the defect area showed significant results two months after local application of statins. Bone healing was statistically better in the intervention groups in the five trials in which it was analyzed, considerably reducing the healing periods compared to controls [11][12][13][14][15].
In relation to BMP-2, all the studies that analyzed it found a higher expression or a greater number of cells that stained positive for BMP-2 after the application of statins [8,10,13,16].
The vascular endothelial growth factor (VEGF) was significantly increased after the local treatment in the two cases in which it was analyzed [10,13].
The two trials that evaluated progenitor endothelial cells (EPCs), either in the peripheral blood [8] or in the perilesional tissue [13], obtained a statistically significant increase in the values after the local application of statins.
In three studies, osteoblast levels were analyzed. In one, density was evaluated [13], obtaining a significant increase in the intervention group; in another, the percentage of surface covered with osteoblasts was measured [17], without finding significant differences between both groups. In the last one, the number of osteoblasts in a given area was counted, obtaining a significantly higher value in the intervention group [19].
As for osteocalcin, a significant increase in its expression was also observed after the use of statins in the two trials where it was studied [10,13].
The biomechanical parameters were analyzed twice, obtaining a significant difference of the maximum strength and work to fracture between the intervention group and the control group in one of the studies [14], and of the relative ultimate stress and the relative extrinsic stiffness in the other [13].
Separate meta-analyses were performed to analyze the mean differences of percentages of new bone formed between studies [8,16]. Those two trials evaluated the values of that parameter at different points in time (4-6 and 12 weeks) after the application of local simvastatin in the intervention group (n = 26) versus no introduction of filling material or introduction of polylactic acid in the control group (n = 26).
The forest plot ( Figure 4) shows a percentage of new bone mean difference of 39.5% with a p-value < 0.001 (95% CI: 22.2 to 56.9, heterogeneity I 2 = 98.4%, P < 0.001) at 4-6 weeks after the intervention.  The forest plot ( Figure 5) shows a percentage of new bone mean difference of 43.3% and a pvalue < 0.001 (95% CI: 33.6 to 52.9, heterogeneity I 2 = 97.7%, P < 0.001) at 12 weeks after the intervention. Even with those encouraging results, regrettably, much value cannot be given to these metaanalyzes, since they only include two studies that are heterogeneous.

Discussion
In one of the three studies in which the area of reossification was measured [9], no significant results were obtained after the intervention, although in all those in which the percentage of new bone formed was measured [8,10,16,18], a very similar parameter, all the findings were statistically significant. The reason for no significance could be due to the concentrations of simvastatin used, since it was the only work in which it was said that some of the animals in the intervention group suffered neurological sequelae, so that the dose administered was perhaps too high.
In the same study [9], bone density was analyzed, and no significant increase in density was observed at 2 months, unlike in the rest of the studies [10,13,18]. However, this increase in the density did occur at 30 days, but then the values remained below the values of the control group over time. That could also be correlated with the concentration of statins. Too high of a concentration may not The forest plot ( Figure 5) shows a percentage of new bone mean difference of 43.3% and a p-value < 0.001 (95% CI: 33.6 to 52.9, heterogeneity I 2 = 97.7%, P < 0.001) at 12 weeks after the intervention. The forest plot ( Figure 4) shows a percentage of new bone mean difference of 39.5% with a pvalue < 0.001 (95% CI: 22.2 to 56.9, heterogeneity I 2 = 98.4%, P < 0.001) at 4-6 weeks after the intervention. The forest plot ( Figure 5) shows a percentage of new bone mean difference of 43.3% and a pvalue < 0.001 (95% CI: 33.6 to 52.9, heterogeneity I 2 = 97.7%, P < 0.001) at 12 weeks after the intervention. Even with those encouraging results, regrettably, much value cannot be given to these metaanalyzes, since they only include two studies that are heterogeneous.

Discussion
In one of the three studies in which the area of reossification was measured [9], no significant results were obtained after the intervention, although in all those in which the percentage of new bone formed was measured [8,10,16,18], a very similar parameter, all the findings were statistically significant. The reason for no significance could be due to the concentrations of simvastatin used, since it was the only work in which it was said that some of the animals in the intervention group suffered neurological sequelae, so that the dose administered was perhaps too high.
In the same study [9], bone density was analyzed, and no significant increase in density was observed at 2 months, unlike in the rest of the studies [10,13,18]. However, this increase in the density did occur at 30 days, but then the values remained below the values of the control group over time. That could also be correlated with the concentration of statins. Too high of a concentration may not Even with those encouraging results, regrettably, much value cannot be given to these meta-analyzes, since they only include two studies that are heterogeneous.

Discussion
In one of the three studies in which the area of reossification was measured [9], no significant results were obtained after the intervention, although in all those in which the percentage of new bone formed was measured [8,10,16,18], a very similar parameter, all the findings were statistically significant. The reason for no significance could be due to the concentrations of simvastatin used, since it was the only work in which it was said that some of the animals in the intervention group suffered neurological sequelae, so that the dose administered was perhaps too high.
In the same study [9], bone density was analyzed, and no significant increase in density was observed at 2 months, unlike in the rest of the studies [10,13,18]. However, this increase in the density did occur at 30 days, but then the values remained below the values of the control group over time. That could also be correlated with the concentration of statins. Too high of a concentration may not have beneficial effects on bone regeneration; on the contrary, it could harm the normal regeneration process and have adverse effects, such as necrosis, the formation of granulation tissue and inflammatory infiltration, and muscular degeneration [11]. Three papers analyzed the osteoblasts: one in which the density was measured and the results were statistically significant [13]; another in which the numbers of osteoblasts per area were counted also, with statistically significant results [19]; and the last one in which the percentage of trabecular surface covered by osteoblasts was evaluated, wherein the values were higher than in the control group but not enough to be meaningful [17]. That lack of significance could be due to the scarcity of the samples. Larger samples could give us results of greater value.
In all the studies that analyzed bone healing [11][12][13][14][15], BMP-2 [8,10,13,16], VEGF [10,13], EPCs [8,13] and osteocalcin [10][11][12][13], statistically significant results were obtained when the intervention group was compared with the control group. These studies showed differences in the type of simvastatin used, doses and other substances that were part of the filling material. This seems to indicate that a concentration of statins which is not too high to avoid the appearance of adverse effects can be very useful in the regeneration of bone defects, independently of the other components of the implanted substance.
It seems that the local application of statins could statistically improve some of the key parameters in bone formation, such as BMSCs [8], osteoprotegerin [10], alkaline phosphatase enzyme [10], RANKL [10] and osteoclasts [19], although more studies are needed to confirm the role of statins in osteogenesis, since these values were analyzed only once.
In one of the papers, it was concluded that the improvement in angiogenesis induced by statins favors the fracture union, after analyzing capillary density and angiogenic markers, such as eNOS and SDF-1 [13].
Recently, three pilot studies in humans have been published using simvastatin for bone regeneration [26][27][28]. In one of them [26], the healing of maxillary third molar postextraction sockets was compared clinically and radiologically after the application of different preservation materials: deproteinized bovine bone mineral with 10% collagen, poly(D,L-lactide-co-glycolide) with a hydroxyapatite/β-tricalcium phosphate scaffold (PLGA/HA); poly(D,L-lactide-co-glycolide) and hydroxyapatite/β-tricalcium phosphate with a 2% simvastatin scaffold (PLGA/HA/S); and spontaneous healing. The results were not as expected, as there were more failures in sockets filled with PLGA/HA scaffolds with and without simvastatin. Scaffolds with simvastatin showed better results, with less clinical complications than scaffolds without simvastatin. The same conclusion comes from the study by Papadimitriou et al., conducted on 14 New Zealand white rabbits, not included in this review for not meeting criteria, when it suggests that the local application of simvastatin, combined with an appropriate carrier, can promote new bone formation [27].
The two other pilot studies compared the use of β-tricalcium phosphate with or without simvastatin [28], or the use of a bovine bone substitute with or without simvastatin [29], in maxillary sinus augmentation. In the first [28], radiographic follow-up was complemented with a biopsy after 9 months. Histomorphometric results showed that the amount of newly formed bone was statistically significantly higher in the simvastatin group. Because the patients of the intervention group showed an intense postoperative inflammatory reaction, the authors emphasize that the dose of simvastatin should be re-evaluated. The second pilot study [29] performed a radiographic follow-up until 9 months, evaluating alveolar bone height, the vertical height of the grafted bone and density. The results did not show any significant positive effect for simvastatin in maxillary sinus augmentation based on radiographic examination. Perhaps the histological study could have shown positive results, as the previous paper.
No other reviews evaluating the effect of local application of statins on the regeneration of non-periodontal bone defects have been found, but a recent systematic review on the use of statins in implantology for animal models (rats and dogs), obtained a significant increase in the osteogenesis around the implants, in cases where the drug was administered locally, applied directly to the surface of the implant [30]. In a similar way to the present work, the articles included in this review also mainly used simvastatin in different concentrations, although in some cases fluvastatin was used, and the follow-up periods ranged from 14 to 84 days.
There are meta-analyses that refer to the use of statins as an adjunct to scaling and root planning in humans. As two of the most recent analyses, it was concluded that the application of statins together with mechanical periodontal treatment significantly reduces the clinical attachment level and periodontal bone defects [31,32]. Both analyzed studies with long follow-up periods of up to 9 months, in which the statins used were simvastatin, rosuvastatin and atorvastatin, the latter not used in any of the studies of the present review. Although most of the studies that analyze statins used gels with concentrations of 1.2% and 2%, some work in which the drug is administered orally exist for both cases.
As far as limitations of this review go, it should be noted that the studies are different in terms of filler material (statin used, dose and complementary substances), the animal model and the parameters analyzed. In addition, the samples are small and the follow-up periods are short. Because some quantitative data are not provided in the studies, only two studies have been included in the meta-analyses.
Even the positive results obtained in animal model, considering the scarce and incipient pilot studies with ambiguous results, more homogeneous human studies, with larger and randomized samples and histological evaluation, are needed.
In conclusion, the local application of statins could be a promising therapeutic strategy for the regeneration of bone defects due to its apparent osteogenic and angiogenic effect. Further randomized clinical trials with larger samples and histological studies are necessary to establish its effect.