The Effect of Autologous Platelet Concentrates on Maxillary Sinus Augmentation: A Meta-Analysis of Randomized Controlled Trials and Systematic Review

Introduction To assess the efficacy of the autologous platelet concentrates (APCs) combined with autologous bone or bone substitute for the maxillary sinus floor lifting by a meta-analysis. Materials and Methods Electronic databases (PUBMED, Web of Science, EMBASE through OVID, and Cochrane Library) were searched until Dec 31, 2019, and only randomized controlled trials (RCTs) in English were identified. Outcome variables included histologic evaluation, the implant stability quotient values, and radiographic evaluation. Data were analyzed by Revman5.3; the estimate of effect sizes was expressed as the 95% confidence interval; and the risk of bias was evaluated using the Cochrane Collaboration tool. Results 11 RCTs involving 141 patients (214 sites) were included in our meta-analysis, which indicated that the differences in the percentage of contact length among newly formed bone (2.61%, 95% CI, -1.18% to 7.09%), soft tissue area (-0.15%, 95% CI, -0.54% to 0.24%), and residual bone substitute material (-5.10%, 95% CI, -10.56% to 0.36%) in the APC group lacked statistical significance. Besides, there was the same effect on the implant stability quotient (ISQ) values of APC group who underwent implant placement 4 months after sinus augmentation and control group who received implant placement 8 months after sinus augmentation (-0.48, 95% CI, -1.68 to 0.72). No significant effect of APCs on the bone density was found (1.05%, 95% CI, -1.69% to 3.82%). Conclusions The use of APCs in sinus augmentation may be further shorten the time required for bone graft maturation and allow earlier implant placement, but cannot enhance the bone formation in the long term. It is not currently recommended for routine use APCs as an osteoinductive material to bone grafting in sinus augmentation.


Introduction
Due to the aesthetic requirements of the patient and/or functional problems, one of the big challenges for implant dentistry is atrophic maxilla, which arises from the deficiency of hard and soft tissues [1]. Tooth loss in the posterior maxilla induces a progressive resorption of the alveolar bone, pneumatization of the maxillary sinus, and periodontal disease, making it very hard to place dental implants without the support of fixed prosthetic rehabilitation [2,3]. Transalveolar osteotome sinus floor elevation is also called the osteotome technique, capable of increasing the density of the soft maxillary bone and elevating the sinus membrane with grafting substitute materials [4][5][6]. There are different kinds of osteoconductive materials to ele-vate the maxillary sinus floor, including autografts, freezedried bone allografts (FDBA), deproteinized bovine bone mineral (DBBM) xenografts, and alloplastic materials (hydroxyapatite, HA, and beta-tricalcium phosphate, β-TCP) [7]. However, these materials may cause the new bone formed to lack osteoinductive properties and osteogenic capacities, prolong the healing time, and trigger immune responses, etc. Some osteoinductive materials have none of these disadvantages and can enhance bone regeneration, such as autologous platelet concentrates (APCs), bone morphogenetic protein-2 (rhBMP-2), and mesenchymal stem cells (MSCs).
The APCs are prepared by the centrifugation of autologous blood, and three generations of APCs have been developed so far, including platelet-rich plasma (PRP), platelet-rich fibrin (PRF), and concentrated growth factors (CGFs) [8]. The PRP, as the first generation of APCs, was introduced into the field of stomatology by Marx in 1998 [9]. The concept of PRF, a new second generation of platelet concentrate, was raised by Choukroun in 2006 [10], and it is greatly different from PRP. Compared with PRP, there was no chemical additive added for the separation of PRF, and PRF could avoid the possibility of immune rejection and allergic reactions for its three-dimensional molecule network enabled a favorable cell migration, attachment, and differentiation. CGF is the latest generation of APCs introduced by Rodella et al. in 2006 [11], which contains a higher concentration of growth factors and releases growth factors and matrix proteins slowly. In order to verify the capability of APCs' osteoinductive effect on sinus augmentation with osteoconductive materials, we carried out a systematic literature review and meta-analysis on the efficacy of APCs in sinus augmentation.

PICOS Question.
Our study was conducted and reported according to the PRISMA (Preferred Reporting Project Guidelines for Systematic Review and Meta-analysis) protocols [12]. The following statements were used to conduct a systematic search. The participants (P) included healthy adults with sinus augmentation and/or implants; the intervention (I) was APCs combined with osteoconductive materials for sinus augmentation; the comparison (C) was conducted without the APCs; the outcomes (O) comprised radiographic, histomorphometric, clinical, and postoperative implant stability quotient assessment; the study (S) was designed for humans, and only randomized controlled trials (RCTs) were enrolled.

Search Strategy and Terms.
A computer-assisted literature search of the PUBMED, Web of Science, EMBASE through OVID, and Cochrane Library electronic databases was performed (last access Dec 31, 2019) to identify studies on the use of the APCs in sinus augmentation. The focus was on the hypothesis that the APCs could have a potentially positive effect on bone regeneration. The search strategy was performed by using the following terms: ("platelet concentrates" OR "platelet rich plasma" OR "platelet rich fibrin" OR "concentrated growth factors") AND ("maxillary sinus augmentation" OR "maxillary sinus floor lift" OR "maxillary sinus floor elevation"). An additional hand search of the following journals was performed on the official websites: , and then a mean and a standard deviation were calculated using the method of Hozo et al. [13]. The data were expressed 2 BioMed Research International as mean difference (MD) and 95% confidence interval (CI). Besides, we also calculated the standardized mean differences (SMD) for different measures of outcome and gave their definitions. Heterogeneity was interpreted as recommended by the Cochrane Handbook, and χ 2 and Higgins index (I 2 ) were used to judge whether there was heterogeneity. The fixed effects model (FE) was used for meta-analysis if the heterogeneity between studies was low (I 2 ≤ 50%), and the random effects model (RE) was employed if the heterogeneity between studies was significant (I 2 > 50%). The heterogeneity between studies in radiographic, histomorphometric, and ISQ assessment was compared through subgroup analysis. Furthermore, the related forest plots were also generated.

Results
The article selection process is shown in Figure 1. A total of 300 articles were identified by the outline and hand searching. 160 records remained after the exclusion of duplicates, and 20 potential studies that met the eligibility criteria were selected. After full-text reading of these 20 articles, 11 articles were finally included for meta-analysis [14][15][16][17][18][19][20][21][22][23][24], and the other 9 articles were excluded because of the incomplete results. Finally, a total of 141 patients with 214 maxillary sinuses under treatment were enrolled, including 87 patients with 102 maxillary sinuses in PRF group and 54 patients with 112 maxillary sinuses in PRP group. Four articles adopted a parallel design and seven articles adopted a split-mouth design. The main characteristics and outcome data of the included studies are shown in Tables 1 and 2, respectively. The sensitivity analysis of results is shown in Table 3.
3.1. Quality Assessment of the Studies. The quality analysis of the comparative studies for risk of bias is shown in Figure 2. Five articles [17,18,[21][22][23] were highly risky in sequence generation for they did not explain the methods of random generation. Allocation concealment was considered an unclear risk of bias in 10 articles because the method of allocation concealment was elevated, and only one article [19] was at low risk of bias for the investigator responsible for enrolling participants, who did not know in advance group, and the next person would join in the allocation concealment process. All performance biases were at high risk since the patients and operator were not made blind in the surgery. All the articles were at low risk of detection bias except one [20] without blinded operators in radiographic and histologic analyses. The risk of attrition bias was high in three articles because the bone biopsy of 2 ruptured sinuses was not feasible for histomorphometric analysis [18], and the postoperative pain and oedema were not evaluated [19], and one BioMed Research International   [23]. The reporting bias and other biases of all articles were at low risk. In the end, five articles [14-16, 19, 24] were classified as moderate risk (two criteria were not met or unclear) and six articles [17,18,[20][21][22][23] as high risk (three or four criteria were not met or unclear).

Implant Stability Quotient (ISQ)
Values. The implant stability quotient values (ISQ) were found in 2 articles [14,15] with 3 periods in the two groups compared. And the random effects model (RE) was adopted owing to the high heterogeneity ( Figure 6). The ISQ at implant loading (76:08 ± 5:86) was significantly higher than that after  7 BioMed Research International implant placement in the test group (60:9 ± 9:35); the results of the included studies showed that the use of PRF determines a greater ISQ than osteoconductive materials alone, but there was no statistically significant difference between the two groups (-0.48, 95% CI, -1.68 to 0.72; p = 0:43),

Radiographic Evaluation.
Four studies [15,19,22,23] reported the postoperative radiographic evaluation, and due to the high heterogeneity (I 2 = 95%), a random effects model was selected. There was no significant difference concerning the bone density by radiological analysis between the groups, with an SMD of 1.06% less bone density observed in the APCs added to osteoconductive materials compared with the osteoconductive materials alone (1.06%, 95% CI, -1.69% to 3.82%; p = 0:45). The details of each study can be found in Figure 7. 3.3. Sensitivity Analysis. Sensitivity analyses were carried out by discarding one research every time to assess the impact of single research on the general outcomes. The overall stability of our results is shown in Table 3.

Discussion
In the past several decades, APCs have attracted the attention of scholars as a potential regenerative material in the treatment of tissue healing [25]. There were lots of researches on improving sinus augmentation with the use of regeneration  [26][27][28][29], but there was a lack of meta-analysis on these clinical results. The present meta-analysis was aimed at evaluating the additional effect of APCs with osteoconductive materials on the sinus augmentation.
Generally, the histologic evaluation of sinus augmentation was mainly reflected in the percentage of new bone formation, percentage of residual bone substitute material, and percentage of soft tissue area. The results showed that 9 BioMed Research International the quantities of histologic evaluation were equivalent between the APC group and the non-APC group, and the result was similar to the findings of Liu's analysis [30], which indicated that there were also no statistical differences in new bone formation and newly formed bone substitute between the non-PRF and PRF groups. Besides, in the research of Pichotano et al. [14], the healing time between sinus augmentation and implant placement could be considerably reduced     Heterogeneity: Tau 2 = 77.94; Chi 2 = 4.05, df = 1 (P = 0.04); l 2 = 75% Test for overall effect: Z= 1.14 (P = 0.25) Figure 4: Percentage of residual bone substitute material (%).

10
BioMed Research International to 2-4 months by using PRF. The autogenous bone combined with PRP could also achieve the same favorable results based on the research report of Thor [21]. Recent studies [10] also showed similar results that adding PRF to the graft materials used for the maxillary sinus lifting had no beneficial effect on regeneration and new bone formation. These studies could not confirm the potential effect of PRF in promoting bone formation due to the long graft healing time [8,26]. The radiological analysis of bone density was conducted on four studies, three of which were in PRP group [19,22,23] and only one in PRF group [15]. Olgun and Consolo reported the bone density of 86:66 ± 43:57 and 890:7 ± 74:25 at four months after grafting in APC group, respectively, compared to 160:81 ± 63:65 and 427:5 ± 60:76 at six months after grafting in the control group. Khairy showed the bone density was 151:5 ± 37:8 in sinus lifted by autogenous bone with PRP at 3 months and 144:5 ± 35:5 at 3 months after implantation in control group. However, there was no significant difference in sinus floor elevating between the APC group at about 3 months and the control group at about 6 months after grafting, indicating that APCs might influence the early bone healing even though they had no potential effect of accelerating bone formation in the long term. This result was confirmed by the animal studies of Miron and Gerard [31,32]. There was no direct correlation between the application of APCs in sinus augmentation and the radiologic and histological variables, which might be due to the small size of RCTs.
In general, dental implant is placed in the maxillary sinus after 8 months of the floor augmentation healing, and the implant stability quotient (ISQ) was used to detect the contact of bone-to-implant that is important to the effect of sinus augmentation. Pichotano et al. [14] claimed that compared with the control group, the PRF plus DBBM group installed the implant 4 or 8 months after augmentation, and such a shorter healing time for implant played a vital role in increas-ing the implant stability. It was also found by Pichotano et al. that the sinus augmentation with DBBM plus PRF could shorten the time required for bone graft maturation and was conducive to the earlier implant placement. The study of Cömert Kılıç et al. [17] also proved that the addition of L-PRF improved implant stability and allowed for faster osseointegration, and ISQ values at loading demonstrated a significant increase in the test group compared to the initial value at implant placement (60:90 ± 9:35 and 76:08 ± 5:86). Through the histologic examination of the mini-implants retrieved, Aimetti et al. [33] found that at the abutment connection, a higher bone-to-implant contact rate was observed on the autogenous bone added with PRP than on the autogenous bone alone despite of similar clinical and radiographic healing patterns (46:75% ± 13:6% vs. 20:5% ± 5:57%, respectively). However, there was no statistically significant difference between the two groups in the result.
Sensitivity analyses were performed to detect the heterogeneity by the method of discarding research. The heterogeneity of histological assessment among these studies showed no significant difference except the percentage of soft tissue area (I 2 = 0). Furthermore, APCs had no significant effect on the maxillary sinus elevation in the long term. The sensitivity analyses indicated the heterogeneity of the implant stability quotient values (I 2 = 0%) existed between the studies due to the period of implant placement by discarding the study of Pichotano et al. [14] (implant placement 4 or 8 months after augmentation). The ISQ of the test and control groups who underwent implant placement 4 and 8 months after sinus augmentation, respectively, was measured. The results showed significantly higher ISQ values after implant placement in the control group (75:13 ± 5:69) than those in the test group (60:90 ± 9:35), but a significant increase in ISQ was observed in the test group at the time of implant loading (60:9 ± 9:35 to 76:08 ± 5:86, p = 0:0014). The above results might be attributed to sinus augmentation with PRF   11 BioMed Research International shortening the time required for implant healing, which agree with Canellas [34]. In their systematic review, they showed PRF increased implant stability 1 week and 1 month after implant surgery and no difference between the use of PRF mixture with bone substitute or bone substitute alone in sinus lift procedure.
Concentrated growth factors (CGFs) are the PRF derivatives developed by Rodella et al. in 2006, and there is a much larger, denser, and richer growth factor fibrin matrix in CGFs than PRF [11]. In our analysis, CGFs are not present in subgroups for lack of randomized controlled studies, but some other studies indicated its importance in sinus floor augmentation. A study on canine model aiming at assessing the effect of Bio-Oss plus concentrated growth factors (CGFs) on bone regeneration for maxillary sinus floor augmentation showed that grafting with Bio-Oss in combination with CGFs could increase new bone formation more efficiently than using Bio-Oss alone [35]. Autologous fibrin-rich blocks with concentrated growth factors (CGFs) were found in a retrospective study to be the predictability of new bone formation in both the maxillary sinus augmentation and implant placement. Anitua et al. reported that autologous platelet aggregates [36], such as platelet-rich plasma and platelet-rich fibrin gel in CGFs, have been used to accelerate new bone formation associated with guided bone regeneration and sinus grafting for many years.
At present, there are several meta-analyses on APCs for alveolar ridge preservation, gingival recession, and tooth extraction. However, there is less analysis on APCs for sinus augmentation and just few systematic reviews about PRF or PRP on maxillary sinus augmentation [37][38][39] and one meta-analysis [30] on PRF. In their reviews, PRF or PRP had effect on soft tissue healing and postoperative symptomatology, but it was conflicting on new bone formation during maxillary sinus augmentation when combined with other biomaterials. Similar to the results of this study, the percentages of residual bone substitute material and new bone formation were no significant difference between APC group and non-APC group in our results. There are more effective analyses to the subgroup of APC group and less postoperative symptomatology analysis in our research, compared with Liu's analysis and other reviewers.
However, there were some limitations to this metaanalysis. Firstly, the inherent heterogeneity could not be avoided between the included articles, 7 articles were splitmouth design and 4 articles were parallel design, which affect the methodological analysis. Secondly, only PRF and PRP subgroup analysis and lacking RCTs of CGFs were enrolled in this analysis, which may have reduced the power in detecting significant differences. Lastly, limitation of this paper was the high risk of bias in the selected studies. Because of the methodological shortcomings for the production processes of APCs, it was impossible to conduct the allocation concealment strictly or to be blind to the personnel in the surgery. Our risk of bias assessment results showed that among 11 included studies, 5 studies were classified as moderate risk and 6 as high risk. Taken together, the conclusions of our analysis are limited, and more studies with low risk of bias in this field are needed in the future.

Conclusion
According to the system review and meta-analysis, we can initially get the following conclusions: (1) Based on the limited studies, it seemed that PRF or PRP failed to show additional effect on new bone formation and implant stability when combined with osteoconductive materials. It is not currently recommended for routine use PRF or PRP as an osteoinductive material to bone grafting in sinus augmentation (2) The addition of APCs to osteoconductive materials in sinus augmentation may help to reduce the healing time and postoperative symptomatology and conducive to shortening the time required for bone graft maturation and allows earlier implant placement, but it need more RCTs to confirm (3) The limited RCTs for CGFs and high risk of bias of studies and more low-risk correlational studies were needed

Conflicts of Interest
The authors have declared that no competing interest exists.