1 INTRODUCTION

Dimensional changes of the alveolar ridge following natural healing of tooth extraction sites are well documented in the literature (Amler et al., 1960; Pietrokovski & Massler, 1967). Schropp et al. (2003) showed that 50% of the ridge width was lost during the first year, with two thirds of that loss taking place during the first 3 months following extraction. A systematic review reported a mean ridge width reduction of 3.87 mm and mid-buccal height loss of 1.67 mm post extraction in sockets left to heal naturally (van der Weijden et al., 2009).

Unfavorable dimensional changes resulting from this healing process may necessitate technique sensitive guided bone regeneration (GBR) procedures prior to dental implant placement. Ridge preservation procedures (RP) may limit these dimensional changes (Araújo & Lindhe, 2009; Darby et al., 2009; Iasella et al., 2003; Vignoletti et al., 2012). A systematic review evaluating the effect of RP following extraction of non-molar teeth showed a mean reduction of 1.89 mm in buccolingual width and 2.07 mm in buccal height, which was significantly less than natural healing in the control group (Avila-Ortiz et al., 2014). Another study found that 58% of non-preserved premolar and molar extraction sites required bone grafting at time of implant placement compared to only 7% of the sites having RP (Cardaropoli et al., 2015). To date, no one specific biomaterial or technique for RP has been shown to be superior (Atieh et al., 2015; Corbella et al., 2017). The use of a dPTFE membrane over the socket of an extracted tooth without bone graft material has been shown to be clinically and histologically effective as RP technique when the membrane was left to heal exposed (Hoffmann et al., 2008). However, there is limited evidence examining the use of a non-resorbable membrane with no bone graft as RP technique. Therefore the primary purpose of this study is to examine soft and hard tissue dimensional changes and the histologic parameters of healing when RP was performed at molar extraction sites using a dPTFE membrane alone. Secondary objectives were to compare these dimensional changes to the results in three previously published study arms of molar RP from the same research group, including: (a) No RP (control; Walker et al., 2017); (b) RP using mineralized freeze-dried bone allograft (FDBA) with dPTFE membrane (test 1; Walker et al., 2017), and (c) RP using FDBA with collagen wound dressing (CWD; test 2; Al Harthi et al., 2019).

In addition, histologic outcomes of wound healing for the current study arm were compared to the same histologic parameters in the abovementioned treatment arms (Duong et al., 2019).

2 MATERIALS AND METHODS

This study originated as a randomized controlled clinical trial (RCT) comparing natural healing without RP to RP using FDBA with dPTFE membrane at molar extraction sites (Walker et al., 2017). Registered under Clinicaltrials.gov ID: NCT02543398). Consecutive to the completion of the RCT, the authors enrolled patients in a new study arm that examined RP at molar sites using FDBA with an overlying CWD (Al Harthi et al., 2019). The histologic parameters of healing including the percentage of vital bone, residual graft (when applicable), and connective tissue/other (CT/other) for bone cores obtained from the abovementioned study arms were then reported in a separate paper (Duong et al., 2019). Therefore, the current study is best described as a four-arm cohort prospective study.

This study was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2013. The protocol was approved by the institutional review board of UT Health San Antonio (Protocol number: HSC20170339H). Written informed consent was obtained for each patient by the Graduate Periodontics research team.

2.1 Participant enrollment protocol

One hundred and six patients, 27 in the present study and 79 from the previous three arms (Al Harthi et al., 2019; Walker et al., 2017), referred to the Graduate Periodontics department at the UT Health School of Dentistry, San Antonio, Texas who required extraction of a single molar due to non-restorable carious lesion, prosthetic or endodontic failure, or root fracture followed by replacement with a dental implant were enrolled between January 2014 to November 2018.

Participants were considered eligible if they had ≥10 mm of radiographic alveolar bone height, without impinging on the maxillary sinus or inferior alveolar canal and adequate restorative space for dental implant-retained restoration. Exclusion criteria included: (a) active localized or systemic infection and untreated periodontal diseases; (b) pregnant women or intended to become pregnant during the study period; (c) medical conditions or taking medications that may negatively affect soft and hard tissue healing; (d) smoking more than 10 cigarettes per day; (e) root of the tooth to be extracted was not in alignment with proposed implant osteotomy; and (f) more than 50% dehiscence of the total socket depth on either buccal or lingual wall at time of extraction.

2.2 Surgical protocol

Following enrollment, alginate impressions were made for fabrication of a customized stent (Essix ACE plastic 0.04″ [1 mm] thickness, DENTSPLY, York, PA) with a radiographic fiducial marker (US Patent 20060241406 A1, Dr. Noujeim, UT Health, San Antonio, TX) to be used at the time of CBCT scans as described by Walker et al. (2017)). On the day of extraction, keratinized tissue (KT) width for mandibular molars was measured to the nearest 0.5 mm using a periodontal probe (UNC-15, G. Hartzell & Son, Concord, CA) from the gingival margin to the mucogingival junction (MGJ) at the mid-buccal and mid-lingual sites of the molar to be extracted. Baseline KT was calculated by adding the width of the socket recorded immediately following extraction at the middle of the socket to the preoperative buccal and lingual KT width. After obtaining anesthesia, full-thickness flaps were reflected not to exceed 3 mm from the crest, and a minimally traumatic molar extraction was performed. Following extraction, the integrity of the socket was evaluated for the presence of dehiscence or fenestration. If a dehiscence greater than 50% of the socket depth was present, the patient was exited from the study and standard care therapy was delivered. The inter-radicular bone, if present, was removed to a level at least 6 mm apical to the bony crest to prevent collecting native bone at the time of bone core harvest.

The socket was covered with a dPTFE membrane (Cytoplast TXT-200 Single, Osteogenics Biomedical, Lubbock, TX) extending 3 mm beyond the buccal and lingual bony crest. No bone graft was used. Flaps were approximated and sutured together using 4–0 PTFE sutures (Cytoplast PTFE sutures, Osteogenics Biomedical, Lubbock, TX) with no attempt for primary closure. A CBCT scan (3DX Accuitomo, J. Morita USA, Irvine, CA) was taken with the radiographic stent in place within 72 h of extraction.

All patients were prescribed amoxicillin 500 mg or clindamycin 300 mg three times a day for a week. A 0.12% chlorhexidine gluconate mouth rinse was used twice a day for 4 weeks following extraction. First postoperative visit was scheduled at 14 days post extraction followed by a second visit for removal of non-resorbable sutures and membrane between 21 and 28 days postoperatively.

Three months following extraction, participants returned for the second CBCT (final) taken with the same radiographic stent. Within 3 weeks from the final CBCT, implant surgery was scheduled. On the day of implant placement surgery, the width of KT was measured at the prospective mandibular implant sites by using a piece of floss placed passively over the KT of the healed ridge and marked at the MGJ on both the buccal and lingual aspects, then measured to the nearest 0.5 mm with the periodontal probe. Full thickness flaps were reflected, and the implant osteotomy initiated with a trephine drill (Salvin Dental Specialties, Charlotte, NC) to allow retrieval of a bone core of at least 6 mm in length and 2 mm in diameter at the middle of the former socket. The harvested bone core was immediately placed in 10% formalin. Thereafter, implant osteotomy was continued with consecutive drilling using the implant manufacturer's recommendations. A dental implant with a diameter of 4.7 mm (Zimmer Biomet, Warsaw, IN) or 4.8 mm (Straumann, Basel, Switzerland) was placed. In cases where there was implant surface exposure >4 mm2 (2 x 2 mm), bone grafting at time of implant placement was performed and recorded. A cover screw or healing abutment was attached based on the surgeon's discretion. Following implant placement, patients completed the study and standard care was provided thereafter.

2.3 Radiographic measurement protocol

Two CBCT scans were taken for each patient from September 2017 to February 2019 at the UT Health San Antonio, Oral and Maxillofacial Radiology clinic. Methodology of the radiographic measurement technique was described in detail previously (Walker et al., 2017).

2.4 Histologic processing and analysis

The histologic processing and histomorphometric analysis of harvested bone cores followed the same protocol previously described by Beck and Mealey (2010). Briefly, the processed core was viewed under ×4 microscope magnification and digital pictures of the core were captured. The pictures were then exported to imaging software (Adobe Photoshop CC 20.0.2, Adobe, San Jose, CA) and merged into one image. Vital bone and CT/other were identified and traced separately. Individual images of each tissue component were generated and converted into binary images (Image J, National Institutes of Health, Bethesda, MD) to calculate the total pixels in each image. The percentage of total area of each tissue component was then calculated based on the total number of pixels.

2.5 Statistical analysis

Based on the results reported by two previous studies (Leblebicioglu et al., 2013; van der Weijden et al., 2009) the distribution of bone width changes was expected to be normally distributed with a mean of −1.21 mm and a standard deviation (SD) of 0.54 mm when no RP is performed. We assumed that a clinically significant effect of RP will be reducing the loss of ridge width by 1 mm, resulting in a mean of 0.21 mm in the treated group.

Eight subjects in each group would result in a power of 0.90 to detect such a difference in width using a Student's t test for independent samples in a two-tailed test with alpha = 0.05. The mean crestal bone height change reported by van der Weijden et al. (2009) was −1.53 mm with a SD of 0.88 mm. If a difference of 1 mm is again used as clinically important, 18 subjects per group resulted in a power of 0.90 to detect such a difference in ridge height using a Student's t test for independent samples in a two-tailed test with alpha = 0.05. Use of a larger number of 18 subjects per group thus resulted in sufficient power for both outcomes. It was anticipated that approximately 70% of patients would complete the study; thus, the planned enrollment was 26 subjects. If the true population mean difference is 1 mm or more and the population standard deviations are as above, then the proposed sample is sufficient to detect a clinically significant bone width change mean difference and crestal height changes between patients who received ridge preservation with a dPTFE membrane (test 3) compared to no RP (control) and other RP techniques (test 1, test 2) reported previously by the same research group (Al Harthi et al., 2019; Walker et al., 2017).

Analysis of variance comparing the four groups was done for each outcome measured as well as the computed change scores. This was followed with pairwise multiple comparison t tests where p-values less than 0.05 were reported. Means and standard deviations are given for descriptive statistics on these analyses. Residuals from the analysis of variance were plotted to help assure a bell shaped distribution approximating a normal distribution was present for a valid analysis.

Frequency data on demographics, molar and bone type distribution data were compared for the four groups with the exact chi-square test. Intraclass correlation coefficients (ICC) were computed with variances for raters and subjects compared using restricted maximum likelihood method to describe rater reliability. The relationship between ridge width change and buccal plate thickness was analyzed with Pearson's correlation coefficient. All analyses were done using statistical software (SAS Version 9.4 for Windows, SAS Institute, Cary, NC). Intraexaminer agreement for radiographic measurement was determined by repeating 10% of the previous radiographic measurements performed by the research primary investigator (A.M.A.).

3 RESULTS

Eighteen of 27 enrolled patients, 8 females and 10 males with a mean age of 52.8 ± 16.1 years, completed the study. Nine patients were exited due to the following reasons: one patient decided not to have an implant, one patient did not get the first CBCT scan within the first 72 h following extraction, two patients withdrew from the study due to financial reasons, and five patients had greater than 50% buccal plate dehiscence. One core biopsy could not be harvested because of poor tissue quality; however, the implant was placed with good primary stability. Therefore, a total of 17 bone cores were harvested. None of the enrolled patients smoked or reported a history of uncontrolled systemic disease. Three patients required grafting at time of implant placement due to exposure of >4 mm2 of the implant surface. One patient did not receive an implant after successful bone core harvest due to lack of primary stability at time of implant placement. The site was grafted with bone allograft and implant placement was performed 3 months following graft healing. The remaining 17 patients had implants placed without any complications encountered during the 3 months healing period following implant placement and were referred to their restorative dentists thereafter.

Clinical, radiographic and histologic analyses were performed for the four treatment arms, which included a total of 79 patients, all treated by the same research group. From a previous RCT (Walker et al., 2017), 20 molar sites received a CWD sponge (CollaPlug, Zimmer Dental, Warsaw, IN) alone (Control) and 20 received RP using FDBA (enCore, Osteogenics Biomedical, Lubbock, TX) with a dPTFE membrane (Test 1). From a previous a case series (Al Harthi et al., 2019), 21 molar sites received RP using FDBA with an overlying CWD. The current study included 18 sites receiving dPTFE membrane alone. Differences in demographic characteristics (age and gender) and molar type distribution between the four groups (Table 1) were not statistically significant (p > 0.05). There was no significant difference in the KT change among the groups (p > 0.05; Table 2).

TABLE 1. Patient demographics and molar type distribution Group Patient age (years; mean ± SD) Gender Molar type Male Female MX1 MD1 MX2 MD2 Walker et al. (2017) Control (no RP) 53.3 ± 10.7 7 13 3 13 0 4 Test 1 (FDBA + dPTFE) 54.1 ± 11.6 7 13 2 16 0 2 Al Harthi et al. (2019) Test 2 (FDBA + CWD) 49.8 ± 14.6 7 14 2 15 0 4 Present study Test 3 (dPTFE alone) 52.8 ± 16.1 10 8 3 12 0 3 Abbreviations: MX1, maxillary 1st molar; MX2, maxillary 2nd molar; MD1, mandibular 1st molar; MD2, mandibular 2nd molar. TABLE 2. Keratinized tissue width change for mandibular molars Group Pre KT Post KT Loss of KT width Walker et al. (2017) Control (n = 17) 19.9 ± 3.7 12.1 ± 5.1 6.8 ± 3.0 Test 1 (n = 18) 18.9 ± 3.6 13.1 ± 5.4 5.2 ± 3.7 Al Harthi et al. (2019) Test 2 (n = 19) 22.1 ± 3.9 17.2 ± 3.4 4.9 ± 2.7 Present study Test 3 (n = 15) 21.9 ± 3.9 15.8 ± 4.4 7.1 ± 5.0 Note: Keratinized tissue at baseline (Pre KT), 3 months after extraction (Post KT) and loss of width. All values expressed in mm (mean ± SD). No significant differences in change in KT among groups. 3.1 Radiographic observations

The time between CBCT scans was 94.5 ± 6.6, 95.1 ± 8.5, 89.4 ± 5.0, and 90.2 ± 5.0 days for control, test 1, test 2, and test 3, respectively. While there was a statistically significant difference between groups (p < 0.05) in the number of days between CBCTs, the maximum between group difference in the mean number of days was less than 6 days.

Baseline buccal and lingual plate thickness is seen in Table 3. At the mesial aspect of the socket the baseline buccal plate thickness was significantly greater in test 3 group compared to the other groups at 1, 3, and 5 mm from the bony crest. This was also true for the distal aspect of the socket at 1 and 3 mm from the bony crest. Statistically significant differences among the groups in the mean change of buccal ridge height were seen between the groups at all aspects of the socket (p < 0.05; Table 4). The greatest loss of ridge height at buccal measurement points was in the control group (no RP), with no differences among the three groups receiving RP. The average loss of lingual ridge height between the four groups was 0.6–1.0 mm, with no significant difference among the groups (p > 0.05; data not shown).

TABLE 3. Baseline buccal and lingual plate thickness Group Buccal plate thickness (mm) Lingual plate thickness (mm) Mesial Distal Mesial Distal 1 mm 3 mm 5 mm 1 mm 3 mm 5 mm 1 mm 3 mm 5 mm 1 mm 3 mm 5 mm Walker et al. (2017) Control (no RP) 0.71 ±0.32* 1.01 ± 0.52§

1.52±

0.86**

1.16±

0.78*

1.82±

1.16§

2.55±

1.51**

1.98 ±1.18 3.30 ±1.48

4.04±

1.45*

2.24 ±1.50*

3.50±

1.60†,‡

4.10±

1.45‖,¶

Test 1 (FDBA+ dPTFE)

0.84±

0.37†,‡

1.12±

0.54‖

1.57±

1.06††

1.07±

0.67†

1.74±

1.31‖

2.50±

1.87††

1.70 ±0.72 2.88 ±1.11

3.57±

1.43

1.50 ±0.76

2.31±

1.30†

2.81±

1.42‖,††

Al Harthi et al. (2019) Test 2 (FDBA + CWD)

0.31±

0.52‡,†

0.56±

0.83¶

0.87±

1.14‡‡

0.33±

0.48*,†,‡

0.93±

0.83¶

1.33±

1.11**,††,‡‡

1.70 ±1.80 2.25 ±1.90

2.50±

1.85*,†

1.43 ±1.24*

2.06±

1.44‡,§

2.11±

1.41¶,**

Present study Test 3 (dPTFE alone) 1.57±1.42*,†,‡

2.58±

1.81§,‖,¶

3.05±

1.78**,††,‡‡

1.98±

1.61*,†,‡

2.75±

2.18§,‖,¶

3.31±

2.17‡‡

2.05 ±1.86 3.14 ±2.18

4.03±

2.25†

1.91 ±1.33

3.27±

1.85§

4.06 ±2.04**,†† Note: Radiographic baseline buccal and lingual plate thickness measured at 1, 3, and 5 mm from alveolar crest expressed in mm (mean ± SD). *,†,‡,§,‖,¶,**,††,‡‡: Similar superscript within a given column indicates a statistically significant difference (p < 0.05) between the groups at that measurement point. TABLE 4. Buccal alveolar ridge height change Group Mean buccal height change (mm) Mesiobuccal Midbuccal Distobuccal Walker et al. (2017)

Control

(no RP)

−3.02 ± 2.24*,† −2.60 ± 2.06‡,§ −2.34 ± 1.72‖

Test 1

(FDBA + dPTFE)

−1.11 ± 1.69* −1.12 ± 1.60‡ −1.01 ± 1.85‖ Al Harthi et al. (2019)

Test 2

(FDBA + CWD)

−1.94 ± 1.73 −1.55 ± 0.93§ −1.30 ± 1.47 Present study

Test 3

(dPTFE alone)

−0.92 ± 2.08† −1.60 ± 1.70 −1.65 ± 1.97 Note: Radiographic buccal ridge height changes expressed in mm (mean ± SD). *,†,‡,§,‖: Similar superscript within a given column indicates a statistically significant difference (p < 0.05) between the groups at that measurement point.

The change in ridge width at 3, 5, and 7 mm from the bony crest is presented in Table 5. At 3 mm from the alveolar crest, the loss of ridge width was significantly smaller in test 3 than in the other groups at all aspects of the socket (mesial, midsocket, distal; p < 0.05). At 5 mm from the crest at both the mesial and midsocket aspects of the socket, a statistically significant difference was seen between the control and test three groups, and between test 1 and test 3 groups; while the distal aspect of the socket showed a significant difference only between the control and test 3 groups (p < 0.05). At 7 mm from the crest the only significant difference was observed at the mesial aspect between the control and test three groups (p < 0.05), with no significant difference at the mid and distal aspects of the socket between the groups (p > 0.05; Table 5).

TABLE 5. Alveolar ridge width change Group Ridge width change (mm) Mesial Midsocket Distal 3 mm 5 mm 7 mm 3 mm 5 mm 7 mm 3 mm 5 mm 7 mm Walker et al. (2017)

Control

(no RP)

−2.97 ± 3.26* −1.46 ± 2.34§ −0.73 ± 1.53¶

−3.11 ± 3.83*

−1.59 ± 2.23§

−0.53 ± 0.92

−2.52 ± 2.22*,‡

−1.07 ± 1.32§

−0.49 ± 0.79

Test 1

(FDBA+ dPTFE)

−2.12 ± 2.43†

−1.00 ± 1.40‖

−0.53 ± 0.83

−2.48 ± 2.86†

−1.16 ± 1.97‖

−0.55 ± 1.29

−2.10 ± 2.53†

−0.70 ± 1.77

−0.22 ± 0.61

Al Harthi et al. (2019)

Test 2

(FDBA + CWD)

−1.46 ± 1.48‡

−0.72 ± 0.80

−0.15 ± 0.79

−1.64 ± 1.10‡

−0.79 ± 0.57

−0.40 ± 0.45

−1.12 ± 1.07‡

−0.47 ± 0.61

−0.20 ± 0.42

Present study

Test 3

(dPTFE alone)

−0.60± 0.86*,†,‡

0.28 ± 1.35§,‖

0.12 ± 1.10¶

−0.13 ± 1.29*,†,‡

0.15 ± 1.77§,‖

0.03 ± 1.57

0.45 ± 1.70*,†,‡

−0.06 ± 1.50§

−0.36 ± 1.48

Note: Radiographic ridge width change measured at 3-, 5- and 7-mm from alveolar crest expressed in mm (mean ± SD). *,†,‡,§,‖,¶: Similar superscript within a given column indicates a statistically significant difference (p < 0.05) between the groups at that measurement point. 3.2 Dimensional change analysis

Pearson correlation analysis showed no significant correlation between lingual plate thickness and ridge width change. For the buccal plate thickness, there was no significant correlation on the distal root. However, there was a small but significant negative correlation on the mesial root between buccal plate thickness at 3 mm and ridge width change at 3 mm (R = 0.25, p = 0.03), and between buccal plate thickness at 5 mm and ridge width change at 5 mm (R = 0.34, p = 0.003), indicating greater loss of ridge width when a thinner buccal plate was present.

3.3 Radiographic measurement reproducibility

Overall, inter-examiner agreement of digital radiographic measurements was determined by 52 repeated measurements performed by three primary examiners (A.M.A, S.A., C.W.). The interclass correlation, mean ICC = 0.95, indicated excellent reproducibility between examiners. Intra-examiner reliability was determined by 51 repeated measurements performed by a single examiner (A.M.A.), with mean ICC = 0.88.

3.4 Histomorphometric analysis

Histologic comparison of wound healing among the groups showed significantly greater vital bone formation (62.10%) in the dPTFE membrane alone group (test 3) than the other groups (p < .05; Table 6). There was no significant difference in vital bone formation among the control, test 1, and test 2 groups.

TABLE 6. Histologic outcomes Group %Vital bone (mean ± SD) %Residual graft (mean ± SD) %CT/other (mean ± SD) Walker et al. (2017)

Control

(no RP)

37.97% ± 12.0a N/A *62.43% ± 12.5‖,¶

Test 1

(FDBA+ dPTFE)

25.41% ± 17.8† 10.58% ± 7.1§ *64.01% ± 13.4‖,¶ Al Harthi et al. (2019)

Test 2

(FDBA + CWD)

30.43% ± 21.4‡ 23.22% ± 15.2§ 46.35% ± 11.5‖ Present study

Test 3

(dPTFE alone)

62.10% ± 31.5*,†,‡ N/A 37.94% ± 31.5¶