Bone vitality and vascularization of mandibular and maxillary bone grafts in maxillary sinus floor elevation: A retrospective cohort study

Abstract Objectives Mandibular retromolar (predominantly cortical) and maxillary tuberosity (predominantly cancellous) bone grafts are used in patients undergoing maxillary sinus floor elevation (MSFE) for dental implant placement. The aim of this retrospective cohort study was to investigate whether differences exist in bone formation and vascularization after grafting with either bone source in patients undergoing MSFE. Methods Fifteen patients undergoing MSFE were treated with retromolar (n = 9) or tuberosity (n = 6) bone grafts. Biopsies were taken 4 months postoperatively prior to dental implant placement, and histomorphometrically analyzed to quantify bone and osteoid area, number of total, apoptotic, and receptor activator of nuclear factor‐κB ligand (RANKL)‐positive osteocytes, small and large‐sized blood vessels, and osteoclasts. The grafted area was divided in three regions (caudal‐cranial): RI, RII, and RIII. Results Bone volume was 40% (RII, RIII) higher and osteoid volume 10% (RII) lower in retromolar compared to tuberosity‐grafted areas. Total osteocyte number and number of RANKL‐positive osteocytes were 23% (RII) and 90% (RI, RII) lower, but osteoclast number was higher (retromolar: 12, tuberosity: 0) in retromolar‐grafted areas. The total number of blood vessels was 80% (RI) to 60% (RIII) lower, while the percentage of large‐sized blood vessels was 86% (RI) to 25% (RIII) higher in retromolar‐grafted areas. Number of osteocyte lacunae and apoptotic osteocytes were similar in both bone grafts used. Conclusions Compared to the retromolar bone, tuberosity bone showed increased bone vitality and vascularization in patients undergoing MSFE, likely due to faster bone remodeling or earlier start of new bone formation. Therefore, tuberosity bone grafts might perform better in enhancing bone regeneration.

What is known: • Cortical bone grafts are considered to have less bone regeneration potential than cancellous bone grafts, due to lack of osteogenic cells and less osteoconductive matrix surface.
• Retromolar (cortical) and tuberosity (cancellous) bone grafts are used in patients undergoing maxillary sinus floor elevation (MSFE) for dental implant placement, but their bone regeneration potential has not been compared.
What this study adds: • Tuberosity bone grafts show enhanced bone vitality and vascularization in MSFE compared to retromolar bone grafts.
• Tuberosity bone grafts result in more osteoid deposition, blood vessel formation, and active bone remodeling, indicating that tuberosity bone might perform better as autologous graft in MSFE than retromolar bone.

| INTRODUCTION
Maxillary sinus floor elevation (MSFE) is a frequently performed surgical procedure to restore insufficient bone height in the posterior maxilla allowing dental implant placement. [1][2][3][4] In MFSE, the space created between the maxillary alveolar process, the elevated Schneiderian membrane, and the inwardly rotated lateral sinus wall is filled with graft material. Autologous bone is considered as the gold standard grafting material in MSFE, 5,6 due to its osteoconductive as well as osteoinductive properties. Moreover, it contains osteogenic cells, and does not evoke immunogenic responses. Histologically, autologous bone grafts in MSFE result in predominantly a mature, lamellar type regenerated bone with higher mineralized bone volumes compared to bone substitutes which result in regenerated bone with lower mineralized bone volumes with a more immature, woven type of bone. [7][8][9] Therefore, autologous bone demonstrates increased bone regenerative potential compared to other grafting materials, such as synthetic, xenograft, or allograft bone substitutes with only osteoconductive properties. 5 Various donor sites are available to harvest autologous bone, including iliac crest, calvaria, tibia, and intraoral sites (mandible, maxilla). [10][11][12][13] The choice of the donor site is based on the type and quantity of bone graft required, the ease of access to the donor site, and the time required with regard to the harvesting procedure and costs involved. 3,[12][13][14][15] Autologous bone grafts from intraoral sources are widely used in MSFE, either applied purely or mixed with a bone substitute. 3 A major advantage of intraoral sites for bone harvesting compared to extraoral sites, is that the graft can be harvested under local anesthesia. 13,14 The mandibular retromolar and maxillary tuberosity regions are favorable donor sites due to low morbidity compared to other intraoral sites. 13,14,16 There are multiple clinical and biological differences between bone harvested from the retromolar versus the tuberosity region. Bone from the retromolar region is predominantly cortical with a high mineral density, while bone from the tuberosity is more cancellous with a lower mineral density. 16,17 Cortical bone grafts are considered to have less bone regeneration potential than cancellous bone grafts, due to the lack of osteogenic bone marrow cells and less osteoconductive matrix surface. [18][19][20] Cortical bone grafts show delayed vascularization due to their lack of porosity and consequent inhibition of vascular ingrowth, resulting in reduced diffusion of oxygen and nutrients through the cortical matrix.
Therefore, cells in cortical grafts, compared to cancellous grafts, are less likely to survive grafting procedures. It has been suggested that primitive osteogenic cells surviving transplantation and forming mature osteoblasts are crucial for the formation of new bone. [20][21][22] Moreover, cortical bone grafts contain fewer osteoprogenitors than cancellous bone.
Finally, the remodeling period of cortical bone graft takes longer, due to longer resorption time preceding osteogenic new bone formation. 21,22 The majority of histologic and histomorphometric studies evaluating different sites and methods of autologous bone grafting in MSFE investigated bone grafts from the iliac crest and chin. 12 Only four studies investigated purely retromolar bone graft in MSFE. 8,[23][24][25] No studies investigated purely tuberosity bone graft in MSFE. Comparison of the bone regeneration potential of retromolar bone grafts with tuberosity bone grafts in MSFE by means of histological and histomorphometrical analysis has not been performed so far. Therefore, this study aimed to investigate possible differences in bone vitality and vascularization in patients undergoing MSFE using retromolar or tuberosity bone grafts through histomorphometrical analysis of bone biopsies. Four months after the MSFE, we evaluated the biopsies prior to dental implant placement. It was hypothesized that tuberosity compared to retromolar bone graft will show enhanced new bone formation in patients undergoing MSFE. In this study, we report the first comparison of retromolar and tuberosity bone grafts for bone vitality and vascularization in patients undergoing MSFE.

| Study approval
The protocol was approved by the medical ethics committee (IRB) of the VU University Medical Center in Amsterdam (#2016.105). All patients signed a written informed consent before participation in the study. The study was performed according to the STROBE guidelines. 26

| Patient selection
Fifteen patients (4 females and 11 males), who were partially edentulous in the posterior maxilla and required dental implants for prosthetic rehabilitation between 2003 and 2012 were selected consecutively for this study (Table 1). All patients required an MSFE due to insufficient vertical bone height (≤3 mm) in at least one of the planned dental implant positions. Since some biopsies of these locations broke apart and could not be reconstructed properly, we sometimes had to switch to adjacent biopsies instead.
The average age of the patients was 56 ± 2 years (mean ± SEM).
Nine patients undergoing MSFE received mandibular retromolar bone graft, and six patients received a maxillary tuberosity bone graft. The average residual bone height was 6 ± 1 mm (mean ± SEM), with an average residual bone height in patients grafted with retromolar bone of 5 ± 1 mm (mean ± SEM) and with tuberosity bone of 7 ± 1 mm (mean ± SEM). Patient demographics are summarized in Table 1.
The patients included in this study had a healthy periodontium and were non-smokers or moderate smokers (<10 cigarettes/day).
Patients who required horizontal bone augmentation, and patients with specific conditions, for example, systemic diseases, drug abuse, heavy smokers, other semi-invasive dental treatments, and/or pregnancy, were not included in this study. One oral and maxillofacial surgeon performed all surgical procedures both in the Alrijne Hospital, Leiderdorp, and in Amsterdam UMC, location VUmc, Amsterdam, The Netherlands.

| Maxillary sinus floor elevation
All 15 patients underwent MSFE as previously described. 2 A preoperative clinical photograph ( Figure 1A) and a radiograph ( Figure 1B) were taken, and a lateral bony window was prepared and turned inward and upward leaving the lifted Schneiderian membrane intact ( Figure 1C). The generated cavity within the maxillary sinus was filled with pure autologous bone harvested from either the retromolar or tuberosity region. Wound closure was performed with Gore-Tex sutures (W.L. Gore and Associates, Newark, DE, USA), which were removed after 10-14 days. All patients received antibiotic prophylaxis, consisting of 500 mg amoxicillin, 3 times daily starting 1 day preoperatively and continuing 7 days postoperatively. After a healing period of 4 months (post-MSFE), prior to dental implant placement, a panoramic radiograph was made to determine the increase in vertical tissue height at the planned dental implant positions ( Figure 1D).

| Autologous bone graft harvesting technique
The retromolar bone grafts were harvested in half-cylinder shape with explantation trephines (inner diameter 4.2 mm; Institute Straumann AG, Basel, Switzerland), with a drilling speed of 500 rpm with minimal pressure and using sterile saline for copious irrigation, from the external oblique ridge of the mandible. The harvested half-cylinder bone cores were used as a cylinder to fill the recipient site. The halfcylinders were not milled but placed as such in the maxillary sinus

| Dental implant surgery
Four months after MSFE, dental implant surgery was performed under local anesthesia ( Figure 1E). A crestal incision was made with mesial and distal buccal vertical release incisions. A full-thickness mucoperiosteal flap was raised to expose the underlying alveolar ridge, which was inspected visually for sufficient bone volume for the intended dental implant placement. Bone biopsies were obtained during dental implant surgery, using trephine drills with a length of 40.5 mm, and with an external diameter of 3.5 mm matching the outer core diameter of the dental implants and an inner diameter of 2.5 mm (Institute Straumann AG, Basel, Switzerland), with a drilling speed of 500 rpm and using sterile saline for copious irrigation, prior to dental implant insertion. Immediately after dental implant placement, a panoramic radiograph was made to check dental implant positions ( Figure 1F). Panoramic radiographs taken pre-MSFE, as well as before dental implant placement, were used for morphometric measurements to determine the increase in vertical tissue height at the planned dental implant positions, using digital software. Calculations were performed with the use of a conversion factor (1.25Â) that adjusted for magnification of the panoramic radiograph. After 3 months of osseointegration of the dental implants, the superstructures were fabricated and placed by the patient's dentist.

| Bone biopsies
The bone biopsies taken during dental implant surgery with a trephine drill were fixated in 4% phosphate-buffered formaldehyde solution dome-shaped, crumbled cranial side (Figure 2A,B). These histologic features were used to identify the apicocoronal orientation of the biopsy. Subsequently, the whole research team verified whether the apicocoronal orientation of the biopsy corresponded with the histological appearance. Consensus was reached for all specimens.
Seventeen biopsies from gap, multiple gap, and free-ending locations were evaluated ( Table 1). The following biopsy location definitions were used: (1) single gap location: a natural tooth is present at both sides of the dental implant location; (2) multiple gap location: a natural tooth is present on either side of at least two dental implants next to each other; multiple bone biopsies can be retrieved in this type of gap; and (3) free-ending location: there is only one natural tooth present at the mesial side of the dental implant location(s); multiple bone biopsies can be retrieved in this situation.

| Histology and histomorphometry
After dehydration in descending alcohol series, the bone specimens were embedded without prior decalcification in low-temperature poly- with respect to the bone regeneration and blood vessel formation in the augmented maxillary sinus. [28][29][30] For each separate area of interest, the histomorphometrical measurements were performed with a computer using an electronic stage  32 The total number of lacunae over bone area (N.Tt.Lac/B.Ar n mm À2 ) and the total number of osteocytes over total number of lacunae (N.Ot/N.Tt.Lac%) were calculated. Only sharp and clearly displayed lacunae with and without osteocytes were included for analysis.
Blood vessel numbers, taking into account the blood vessel size, were determined as mean value of two separate blinded counts.
Blood vessel size was calculated as the total blood vessel area expressed in μm 2 . According to their diameter, blood vessels were divided into small (0-400 μm 2 ) or large vessels (>400 μm 2 ).
Tartrate-resistant acid phosphatase (TRAcP) staining was used to visualize bone resorbing multinuclear cells (osteoclasts) and was per-

| Immunohistochemistry
A previously described protocol for immunostaining was used. 28,29,34 To visualize and calculate the number of apoptotic osteocytes, immu-
In retromolar bone biopsies, compared to tuberosity bone biop-  Figure 3D). Moreover, no significant differences in total number of TRAcP-positive osteoclasts per bone area were found between the regions per bone graft ( Figure 3D).  Figure 4A,B). The total number of lacunae was also similar between the different regions per bone graft ( Figure 4B).

| DISCUSSION
Four months after MFSE, differences in bone vitality and vascularization were observed between retromolar and tuberosity bone grafts. indicating that both graft origin and remodeling rate influence the mineralization degree. 25,38 In this study, lower osteoid volume at the cranial and center of the retromolar bone-grafted areas, compared to tuberosity bonegrafted areas, was found. The reason for this observation is currently unexplained. Interestingly, osteoid volume increased towards the cranial side of the grafted area, which may have resulted from active bone formation starting from the cranial side of the biopsy. This is in line with earlier observations that bone formation may start not only from the maxillary native bone, but from the cranial side as well. 29,35 Moreover, it has been shown that the Schneiderian membrane of the maxillary sinus, which is lifted during MSFE to insert the graft material, contains a cell population with potential for osteogenic differentiation. 40 We found similar total numbers of lacunae and apoptotic osteocytes in biopsies from retromolar and tuberosity bone-grafted areas.
Remarkably, we observed a 90% lower total number of RANKL- Bone is highly vascularized, and vascular development needs to be induced prior to osteogenesis. Our results showed that the total number of blood vessels in the grafted area was lower in retromolar versus tuberosity bone biopsies, which was accompanied by a lower percentage of osteoid volume. In contrast, the higher percentage of small-sized blood vessels and the lower percentage of large-sized blood vessels as we observed in the tuberosity bone-grafted areas indicates higher angiogenic activity in tuberosity bone graft. This is in agreement with earlier studies showing that bone formation is related to increased blood vessel formation. 29,39 The lower total number of osteocytes in the center of the grafted area of the retromolar grafts may consequently be the result of reduced diffusion of oxygen and nutrients due to delayed vascularization in these grafts.
This confirms findings in earlier studies, that is, that cortical bone grafts, compared to cancellous bone grafts, show delayed vascularization due to lack of porosity and consequent inhibition of vascular ingrowth. 21 The study was conducted retrospectively resulting in several limitations. A limitation of the present study was that we compared two different autologous bone grafts in patients undergoing unilateral MSFE. To exclude inter-patient variation, a bilateral sinus floor elevation model would be more appropriate to compare two different grafting materials. Another limitation of this study was that we only analyzed biopsies at one time point, preventing to assess the dynamics of the remodeling process in both types of bone grafts. Therefore, we can only deduce that retromolar grafts displayed a slower bone remodeling rate, but cannot rule out that remodeling might reach similar levels at a later time point. Another limitation of this study was the use of two bone harvesting techniques. However, the bone harvesting techniques were unlikely to affect the results of our study in terms of bone vitality of the graft, since the harvesting techniques were as "atraumatic" as possible. This appears to be histologically confirmed since we did not observe necrotic bone tissue. A limitation of this study was also that no follow-up data of the patients could be obtained. However, no complaints regarding all dental implants have been reported thus far.
In summary, we found that the use of tuberosity bone graft in human MSFE resulted in a 10% higher osteoid volume in the center and at the cranial side of the grafted area, and 150%-300% higher total number of blood vessels in the total grafted area compared to retromolar bone grafts. We conclude that tuberosity bone grafts showed enhanced bone vitality and vascularization in patients undergoing MSFE in comparison with retromolar bone grafts, either due to a faster bone remodeling rate or due to an earlier start of bone remodeling in tuberosity bone graft-treated patients. Based on our histological data, it appears that tuberosity bone might perform better as an autologous graft material in MSFE than retromolar bone, since more osteoid was deposited, more blood vessels were formed, and a more active remodeling process was initiated. A shorter healing period before dental implant placement and loading might be feasible, if tuberosity bone grafts are used.