Impact of Bone Particle Size on Alveolar Ridge Preservation in Periodontally Compromised Molar Extraction Sites

Abstract

Introduction and aims

The clinical efficacy of alveolar ridge preservation (ARP) is well-documented in the literature. However, there is a lack of evidence regarding ARP in the posterior area and in periodontally compromised sites. There is no clinical decision-making guideline on the choice of bone substitute particle size for ARP in periodontally compromised molar sockets. This study aimed to compare the radiographic outcomes and implant feasibility of two deproteinized bovine bone mineral (DBBM) particle sizes for alveolar ridge preservation (ARP) in periodontally compromised molar extraction sockets.

Methods

A total of 96 patients (96 sites) with stage III/IV periodontitis requiring molar extraction were divided according to DBBM particle size (small particle [SP]: 0.25-1 or large particle [LP]: 1-2 mm). Dimensional and volumetric changes of the alveolar ridge and level of implant placement difficulty (LIPD) were evaluated. Multivariable linear regression was used to adjust for potential confounders.

Results

LP-DBBM showed a non-significant trend toward increased vertical central socket height (VC; 8.54 ± 2.34 vs 7.76 ± 1.93 mm; P = .077). No significant differences between SP and LP groups were observed for the remaining alveolar dimensional and volumetric changes. However, multivariable linear regression found that socket integrity (number of socket walls with height ≥5 mm; P < .001), jaw (P = .007), and particle size (P = .041) showed significant effects on VC.

Conclusion

A trend of higher VC using LP after ARP in periodontally compromised molar extraction sockets was observed. The choice of DBBM particle size should be considered alongside site-specific factors, especially socket integrity.

Clinical Relevance

Understanding the impact of graft particle size on ridge dimensional changes and implant placement feasibility can provide clinical decision-making guidance for ARP in periodontally compromised molar sockets. Larger particle size facilitates superior VC and increases the predictability of subsequent implant treatment.

Introduction

Resorption of the alveolar ridge after tooth extraction is a well-documented and inevitable process.¹ Although dental implants are a reliable option for oral rehabilitation,² long-term outcomes are dependent on initial bone quantity and quality. In the posterior region, it has been reported that in the absence of additional intervention after tooth extraction, both horizontal and vertical bone resorption was prominent in molars ranging from 1.20 to 3.61 mm.³ The proximity of critical anatomical structures in the posterior area including the maxillary sinus and inferior alveolar nerve further complicates implant placement.

The clinical efficacy of alveolar ridge preservation (ARP) in maintaining alveolar ridge dimensions and enabling prosthetically-driven implant placement is well-documented in the literature.4, 5, 6, 7, 8, 9, 10 However, there is a marked lack of evidence regarding ARP in the posterior area and particularly in patients with severe periodontitis.^11,12 According to a secondary analysis of the Global Burden of Disease 2021 study, severe periodontitis is projected to affect 1.56 billion individuals in 2050,¹³ highlighting that further understanding of treatment options for this patient population is of paramount importance. It has been reported that substantial reduction in ridge height in periodontally compromised sockets (∼60%) is present compared to intact sockets (35%).^14,15 A previous retrospective study also found that delayed cortical changes were present in extraction sites with periodontal bone damage.¹⁶ Although ARP in periodontally compromised sockets is safe,¹⁷ results on dimensional changes remain controversial.^18,19

Autogenous bone is recognized as the material of choice for bone grafting procedures due to its biocompatibility and osteoinductive properties. However, the difficulty and increased patient morbidity during bone harvesting has necessitated the use of bone substitute materials.20, 21, 22 Deproteinized bovine bone mineral (DBBM) is commonly used as an autogenous bone substitute in ARP.²³ Although DBBM is commercially available in two particle sizes (0.25-1 and 1-2 mm), there is no clinical decision-making guideline on the choice of particle size for ARP in periodontally compromised molar sockets, which present marked anatomical variations compared with intact non-molar anterior sockets. Radiographic outcomes of different bone particle sizes have been compared for lateral sinus floor elevation (LSFE) and lateral ridge augmentation (LRA).^24,25 Large bone particle size significantly improved augmentation stability after one-stage LSFE, but no statistically significant differences in dimensional change were observed between groups after LRA, underscoring the need to investigate its impact at different anatomical locations and baseline ridge morphologies.

To the best of the authors’ knowledge, no studies have documented the effect of bone particle size on ARP in periodontally compromised molar extraction sites. Therefore, the present retrospective study aimed to compare the radiographic outcomes and implant feasibility of two DBBM particle sizes for ARP in periodontally compromised molar extraction sites.

Materials and methods

Study design

This retrospective cohort study included patients diagnosed with periodontitis stage III/IV who underwent tooth extraction and subsequent ARP in molar sites at the Second Dental Center of Shanghai Ninth People’s Hospital between January 2020 and December 2024. The study protocol was approved by the Institutional Review Board at Shanghai Ninth People’s Hospital (SH9H-2025-T423-1) and adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.²⁶ All patients signed an informed consent form prior to tooth extraction in accordance with the Declaration of Helsinki (1975, revised in 2013).

Inclusion criteria

The following inclusion criteria were applied:

(a) ≥18 years of age; (b) need for molar extraction due to periodontitis stage III/IV; (c) the extraction socket was present with a bone dehiscence on at least one wall extending over 50% of the root length, (d) presence of one adjacent tooth, (e) cone-beam computed tomography (CBCT) scans acquired within 3 months of tooth extraction (T0) and 4 to 8 months after ARP (T1).

Exclusion criteria

The following exclusion criteria were applied: (a) presence of systemic diseases; (b) consumption of medication (eg, bisphosphonates, immunosuppressives, etc.) that may alter bone metabolism; (c) history of chemotherapy or radiotherapy in the past 5 years; (d) smoking; (e) pregnancy or lactation; (f) low quality CBCT data or presence of artifacts that may impact measurement accuracy.

Treatment procedure

Prior to and during the surgical and healing process, individualized oral hygiene instructions were provided.

After local anaesthesia, minimally traumatic flapless extraction was performed using periotomes, elevators, and forceps. Granulation tissue was thoroughly removed during extraction socket debridement, and the socket was rinsed with a sterile saline solution. The extraction sockets were subsequently filled with particulate DBBM (Bio-Oss [granulometry 0.25-1 mm or 1-2 mm], Geistlich Pharma AG, Wolhusen, Switzerland), covered with a resorbable collagen membrane (Bio-Gide, Geistlich Pharma AG, Wolhusen, Switzerland), and secured with a combination of simple interrupted and criss-cross sutures (5−0 Vicryl, Ethicon, Somerville, NJ, USA). Healing occurred through secondary intention. All sockets healed uneventfully.

Digital imaging assessments

To ensure measurement accuracy and quality, a single calibrated examiner (X.Y.) measured both linear and volumetric outcomes in 10 randomly selected patient samples at two time points 2 weeks apart. Formal data assessment commenced once an intra-class correlation coefficient of over 0.9 was achieved. An illustration of the digital imaging measurement methodology is displayed in Figure.

Figure.

Radiographic measurements. A, Linear measurements. B, Volumetric measurements. HW₁, HW₃, HW₅, horizontal alveolar ridge width at 1 mm, 3 mm, and 5 mm below the crest; T0, 0 to 3 mo before tooth extraction; T1, 4 to 8 mo after alveolar ridge preservation; VB, vertical buccal plate height; VC, vertical central socket height; VL, vertical lingual/palatal plate height.

Linear measurements

Paired (T0 and T1) CBCT DICOM (Digital Imaging and Communications in Medicine) files were imported into an imaging software (Romexis 6.4.7, Planmeca OY, Helsinki, Finland), superimposed after matching at least five anatomical landmarks, and manually adjusted to ensure precision.

Buccal and lingual/palatal bone plate thickness was measured at 1 mm below the alveolar crest. Horizontal alveolar bone width changes were measured at 1 (HW₁), 3 (HW₃), and 5 (HW₅) millimetres below the mid-buccal/lingual/palatal alveolar crest at T0 according to the method described by Aimetti and colleagues.²⁷ Vertical alveolar bone height changes were measured from the horizontal reference line drawn at the socket apex level to the alveolar bone crest at the buccal (VB) and lingual/palatal (VL) bone plates in millimetres and percentages. Vertical mesial and distal alveolar bone height were also measured in the same manner for the evaluation of socket integrity. The primary outcome, central height (VC), was measured from the socket apex to the most coronal point of the grafted socket at T1. Linear changes were calculated in millimetres (absolute) and percentages (relative).

Volumetric measurements

Volumetric measurements were performed using an open-source image computing software (3D Slicer 5.7.0).²⁸ Paired CBCTs (T0 and T1) were imported and a dento-maxillo-facial segmentation tool, DentalSegmentator,²⁹ was utilized to create accurate models of the jaws at T0 and T1. Subsequently, the maxillary or mandibular models were precisely superimposed using the Surface Registration tool and cropped using the Crop Volume tool to obtain the socket-of-interest in both models using the exact same region-of-interest bounding box. Finally, the Segment Statistics tool was utilized to quantify the volume. Volumetric changes were calculated in cubic millimetres (absolute) and percentages (relative).

Feasibility of implant placement

A previously proposed level of implant placement difficulty (LIPD) criteria was used to categorize the feasibility of implant placement and need for ancillary bone augmentation.³⁰ Briefly, Level I: standard implant placement without the need for ancillary bone augmentation (1.5 mm of bone surrounding the implant); Level II: implant placement with simultaneous bone augmentation and/or osteotome sinus floor elevation (OSFE; ≥5 mm of residual bone height [RBH]); Level III: staged guided bone regeneration (GBR) or lateral sinus floor elevation (LSFE; RBH <5 mm); and Level IV: implant placement not possible (RBH <6 mm, extensive vertical augmentation needed, nerve transposition, etc.).

Statistical analysis

To calculate the minimum sample size needed, a power calculation was performed a priori using G*Power 3.1 (Heinrich-Heine University, Duesseldorf, Germany). The criterion for significance was set at 0.05 and the power at 0.90. Assuming a mean difference in VC of 0.7 mm between particle size groups, with a standard deviation (SD) of 0.345²⁴; an effect size of 0.88 was determined, and a two-tailed analysis found that at least 56 patients (28 per group) was necessary.

Data was analysed using dedicated software (SPSS 30.0, IBM Corp, Armonk, NY, USA; MATLAB R2025A, The MathWorks Inc, Natick, MA, USA). Descriptive statistics were obtained for all variables and the Shapiro–Wilk test was used to test for normality. Categorical variables were compared using the chi-squared test. Continuous variables were compared using independent t-tests. Mann–Whitney U test was performed if the data was not normally distributed. The difference between T0 and T1 was analysed using paired t-tests for normally distributed data or Wilcoxon signed rank sum tests when the data was not normally distributed. A multivariate linear regression model was constructed to analyse potential confounding factors including age, gender, jaw, tooth position, healing duration, socket wall thickness, socket integrity (number of socket walls with height ≥5 mm), and particle size on VC. The significance level was set at P < .05.

Results

Demographics

Initially, 589 patients were screened for eligibility (Appendix Figure 1). A total of 96 patients (96 extraction sockets) were included in this study (37 female, 59 male), with a mean age of 48.89 years (SD: 10.66, range: 25-73) at the time of extraction. Given a 1:1 distribution, 48 patients were treated with ARP using exclusively small DBBM particles (SP) and 48 patients were treated with ARP using exclusively large DBBM particles (LP). Baseline extraction socket variables including healing duration, jaw, tooth location, socket wall thickness, and number of socket walls with height ≥5 mm were comparable between groups with no statistically significant differences observed. Bone resorption to the root apex at the furcation zone was observed for all cases. Detailed description of the patient population and extraction socket samples is shown in Table 1. No significant differences were found with regards to patient/extraction socket demographics.

Table 1.

Demographic data of subjects and extraction sockets at baseline.

Variable	Total	Small Particle Group	Large Particle Group	P
Number of patients/extraction sockets	96/96	48/48	48/48	–
Age (years), mean (SD)	48.89 (10.66)	50.17 (10.53)	47.60 (10.75)	.241
Gender (female/male), N (%)	37/59 (38.5/61.5)	18/30 (37.5/62.5)	19/29 (39.6/60.3)	.835
Healing duration (months), mean (SD)	6.40 (1.05)	6.35 (1.04)	6.44 (1.07)	.696
Jaw
Maxilla/mandible, N (%)	46/50 (47.9/52.1)	24/24 (50/50)	22/26 (45.8/54.2)	.684
Tooth location
First/second molar, N (%)	54/42 (56.2/43.7)	27/21 (56.3/43.7)	27/21 (56.3/43.7)	1.000
Socket wall thickness
Buccal (mm), mean (SD)	0.98 (0.82)	1.02 (0.83)	0.94 (0.83)	.574
Lingual/palatal (mm), mean (SD)	0.95 (0.92)	0.92 (0.95)	0.98 (0.90)	.633
Socket integrity (Number of socket walls with height ≥5 mm)				.862
4 walls, N (%)	46 (48)	22 (46)	24 (50)
3 walls, N (%)	30 (31)	18 (38)	12 (25)
2 walls, N (%)	15 (16)	7 (15)	8 (17)
1 wall, N (%)	5 (5)	1 (2)	4 (8)

SD, standard deviation.

Comparison between particle size groups

As shown in Table 2, no statistical significance was observed in both SP and LP groups at T0 and T1 for all variables of interest (P > .05). All postoperative data for VB, VL, HW₁, HW₃, and HW₅ decreased significantly at T1 while volume increased significantly at T1 for both particle size groups (P < .05). At T1, greater VC in the LP group (8.54 ± 2.34 mm) was observed compared to the SP group (7.76 ± 1.93); however statistical significance between groups was not observed (P = .077). Changes in VB, VL, HW₁, HW₃, HW₅, and volume were not statistically significant between SP and LP groups for both absolute (mm) and relative (%) values (P > .05).

Table 2.

Radiographic linear and volumetric changes in the extraction sites.

	Small particle group	Large particle group	P
VC
T0 (mm), mean (SD)	0.00 (0.00)	0.00 (0.00)	–
T1 (mm), mean (SD)	7.76 (1.93)	8.54 (2.34)	.077‡
Change (mm), mean (SD)	7.76 (1.93)	8.54 (2.34)	.077‡
VB
T0 (mm), mean (SD)	6.19 (2.02)	7.01 (2.53)	.081‡
T1 (mm), mean (SD)	4.22 (2.12)	5.11 (2.82)	.084‡
Change (mm), mean (SD)	−1.97 (1.65)	−1.90 (2.12)	.803§
Change (%), mean (SD)	−31.69 (26.25)	−23.90 (44.12)	.465§
P	.000†	.000†
VL
T0 (mm), mean (SD)	6.17 (2.00)	6.69 (2.74)	.289‡
T1 (mm), mean (SD)	5.06 (1.93)	5.73 (2.67)	.168‡
Change (mm), mean (SD)	−0.99 (1.15)	−0.97 (1.00)	.347§
Change (%), mean (SD)	−18.74 (27.01)	−16.42 (19.90)	.183§
P	.000†	.000†
VM
T0 (mm), mean (SD)	8.82 (2.71)	8.46 (3.04)	.668§
VD
T0 (mm), mean (SD)	7.34 (2.53)	7.56 (2.96)	.695‡
HW1
T0 (mm), mean (SD)	12.38 (2.20)	12.84 (1.96)	.286‡
T1 (mm), mean (SD)	8.34 (2.88)	9.34 (2.99)	.098‡
Change (mm), mean (SD)	−4.04 (3.07)	−3.50 (2.46)	.550‡
Change (%), mean (SD)	−31.61 (21.28)	−27.56 (19.75)	.448‡
P	.000†	.000†
HW3
T0 (mm), mean (SD)	13.62 (2.15)	13.95 (2.20)	.453‡
T1 (mm), mean (SD)	11.62 (2.63)	12.01 (2.77)	.486‡
Change (mm), mean (SD)	−2.00 (1.97)	−1.94 (1.47)	.689§
Change (%), mean (SD)	−14.49 (13.53)	−14.44 (11.83)	.823§
P	.000†	.000†
HW5
T0 (mm), mean (SD)	14.52 (2.18)	14.73 (2.31)	.647‡
T1 (mm), mean (SD)	13.67 (2.34)	14.06 (2.73)	.453‡
Change (mm), mean (SD)	−0.85 (0.91)	−0.67 (0.89)	.120‡
Change (%), mean (SD)	−5.88 (6.35)	−4.98 (7.18)	.144‡
P	.000†	.000†
Volume
T0 (mm3), mean (SD)	1272.18 (521.39)	1393.00 (590.51)	.244‡
T1 (mm3), mean (SD)	1511.31 (688.85)	1762.44 (871.44)	.171‡
Change (mm3), mean (SD)	239.13 (292.80)	369.44 (383.25)	.068‡
Change (%), mean (SD)	17.73 (23.02)	23.78 (23.15)	.202§
P	.000*	.000*

HW₁, HW₃, HW₅, horizontal alveolar ridge width at 1 mm, 3 mm, and 5 mm below the crest; SD, standard deviation; T0, 0 to 3 mo before tooth extraction; T1, 4 to 8 mo after alveolar ridge preservation; VB, vertical buccal plate height; VC, vertical central socket height; VD, vertical distal ridge height; VL, vertical lingual/palatal plate height; VM, vertical mesial ridge height.

^⁎Wilcoxon signed-rank test.

^†Paired t-test.

^‡Independent sample t-test.

^§Mann-Whitney U-test.

A stratification analysis comparing extraction sockets with a baseline VB and VL discrepancy threshold of 2 mm was performed at the intergroup (Appendix Table 1) and intragroup (Table 3) levels. Changes in all variables of interest were not statistically significant in the intergroup comparison (Appendix Table 1; P > .05). Changes in VC, VB, VL, HW₅, and volume were not statistically significant in the intragroup comparison (Table 3; P > .05). Statistically significant absolute and relative changes in HW₁ (P < .05) and HW₃ (P < .05) were present in both SP and LP groups. Specifically for the LP group, statistically significant differences were observed between VB–VL discrepancies >2 and ≤2 mm for HW₁ and HW₃ values at T1 (P < .05).

Table 3.

Radiographic linear and volumetric changes in extraction sites with a baseline buccal and lingual/palatal plate height discrepancy threshold of 2 mm.

	Small particle group		P	Large particle group		P
VB–VL Discrepancy	>2 mm (N = 21)	≤2 mm (N = 27)		>2 mm (N = 21)	≤2 mm (N = 27)
VC
T0 (mm), mean (SD)	0.00 (0.00)	0.00 (0.00)	–	0.00 (0.00)	0.00 (0.00)	–
T1 (mm), mean (SD)	7.69 (2.15)	7.81 (1.79)	.841†	8.81 (2.06)	8.33 (2.55)	.490†
Change (mm), mean (SD)	7.69 (2.15)	7.81 (1.79)	.841†	8.81 (2.06)	8.33 (2.55)	.490†
VB
T0 (mm), mean (SD)	6.02 (2.61)	6.32 (1.44)	.640†	6.74 (3.09)	7.23 (2.03)	.534†
T1 (mm), mean (SD)	4.23 (1.97)	4.22 (2.27)	.988†	4.89 (2.77)	5.28 (2.90)	.643†
Change (mm), mean (SD)	−1.80 (1.86)	−2.10 (1.50)	.332‡	−1.85 (2.69)	−1.95 (1.59)	.581‡
Change (%), mean (SD)	−26.37 (22.54)	−35.83 (28.54)	.453‡	−15.99 (59.75)	−30.05 (26.12)	.328‡
P	.000*	.000*		.005*	.000*
VL
T0 (mm), mean (SD)	5.80 (2.60)	6.46 (1.35)	.303†	5.84 (3.28)	7.36 (2.06)	.073†
T1 (mm), mean (SD)	4.97 (2.39)	5.14 (1.52)	.777†	5.14 (3.00)	6.18 (2.35)	.182†
Change (mm), mean (SD)	−0.84 (1.13)	−1.32 (1.13)	.149†	−0.70 (0.64)	−1.17 (1.17)	.275‡
Change (%), mean (SD)	−16.37 (15.57)	−20.58 (17.14)	.511‡	−15.57 (22.71)	−17.07 (17.84)	.578‡
P	.003*	.000*		.000*	.000*
VM
T0 (mm), mean (SD)	8.92 (3.62)	8.74 (1.80)	.825†	8.96 (2.41)	8.06 (3.45)	.293†
VD
T0 (mm), mean (SD)	7.12 (2.60)	7.51 (2.51)	.599†	7.11 (2.54)	7.91 (3.25)	.361†
HW1
T0 (mm), mean (SD)	12.91 (2.59)	11.97 (1.77)	.142†	12.84 (1.79)	12.84 (2.12)	.992†
T1 (mm), mean (SD)	7.57 (3.25)	8.94 (2.44)	.101†	8.11 (2.89)	10.30 (2.75)	.010†
Change (mm), mean (SD)	−5.35 (3.92)	−3.03 (1.66)	.018†	−4.73 (2.86)	−2.54 (1.56)	.004†
Change (%), mean (SD)	−39.12 (26.21)	−25.78 (14.46)	.045†	−36.66 (23.35)	−20.48 (12.92)	.008†
P	.000*	.000*		.000*	.000*
HW3
T0 (mm), mean (SD)	13.72 (2.45)	13.54 (1.93)	.781†	13.46 (1.75)	14.33 (2.46)	.175†
T1 (mm), mean (SD)	10.98 (3.00)	12.12 (2.24)	.136†	10.93 (2.45)	12.84 (2.76)	.016†
Change (mm), mean (SD)	−2.74 (2.60)	−1.42 (1.02)	.037†	−2.53 (1.72)	−1.49 (1.06)	.021†
Change (%), mean (SD)	−19.31 (17.40)	−10.74 (8.05)	.046†	−19.15 (14.60)	−10.79 (7.58)	.024†
P	.000*	.000*		.000*	.000*
HW5
T0 (mm), mean (SD)	14.33 (2.34)	14.67 (2.08)	.601†	14.10 (1.94)	15.22 (2.49)	.096†
T1 (mm), mean (SD)	13.25 (2.49)	14.00 (2.20)	.273†	13.17 (2.53)	14.76 (2.73)	.046†
Change (mm), mean (SD)	−1.08 (1.13)	−0.66 (0.66)	.144†	−0.92 (1.06)	−0.47 (0.68)	.075†
Change (%), mean (SD)	−7.47 (7.81)	−4.64 (4.72)	.154†	−7.10 (9.06)	−3.34 (4.87)	.072†
P	.000*	.000*		.001*	.001*
Volume
T0 (mm3), mean (SD)	1212.07 (532.93)	1318.93 (517.45)	.487†	1276.95 (551.01)	1483.26 (614.37)	.234†
T1 (mm3), mean (SD)	1443.51 (635.89)	1564.04 (734.93)	.553†	1631.45 (735.90)	1864.32 (965.17)	.364†
Change (mm3), mean (SD)	231.44 (286.78)	245.11 (302.71)	.875†	354.51 (287.85)	381.06 (448.84)	.815†
Change (%), mean (SD)	19.81 (26.32)	16.11 (20.46)	.586†	26.67 (21.76)	21.63 (24.34)	.451†
P	.001*	.000*		.000*	.000*

HW₁, HW₃, HW₅, horizontal alveolar ridge width at 1 mm, 3 mm, and 5 mm below the crest; SD, standard deviation; T0, 0 to 3 mo before tooth extraction; T1, 4 to 8 mo after alveolar ridge preservation; VB, vertical buccal plate height; VC, vertical central socket height; VD, vertical distal ridge height; VL, vertical lingual/palatal plate height; VM, vertical mesial ridge height.

^⁎Paired t-test.

^†Independent sample t-test.

^‡Mann-Whitney U-test.

Feasibility of implant placement

Based on the LIPD criteria, standard-length (≥8 mm) implant placement without bone augmentation (Level I) could be achieved in most sites (70/96, 73%; Table 4). A total of 26 ARP sites required ancillary bone augmentation procedures, with 13 sites designated as Level II and 13 as Level III. A slightly larger number of sites were designated as Level III in the SP group compared with the LP group (8/48 vs 5/48). All Level III designations were in maxillary molar sites requiring LSFE. No sites were designated as Level IV. The LIPD distribution was similar between groups (P = .500).

Table 4.

Need for ancillary bone augmentation.

	Total	Small particle group	Large particle group	P
LIPD				.500†
Level I, N (%)	70 (73)	35 (73)	35 (73)
Level II, N (%)	13 (13.5)	5 (10)	8 (17)
Level III, N (%)	13 (13.5)	8 (17)*	5 (10)*
Level IV, N (%)	0 (0)	0 (0)	0 (0)

LIPD, level of implant placement difficulty.

Level I: Implant placement without ancillary bone augmentation (BA).

Level II: Implant placement with simultaneous BA or osteotome sinus floor elevation.

Level III: Staged guided bone regeneration or lateral sinus floor elevation.

Level IV: Implant placement not possible.

^⁎All cases required lateral sinus floor elevation.

^†Chi-square test.

Multivariate linear regression analysis

A multivariate linear regression analysis was conducted by adjusting for potential confounders with VC as the independent variable (Table 5). Age, gender, jaw, tooth position, healing duration, and socket wall thickness were found to have no significant effect on VC. However, socket integrity (number of socket walls with height ≥5 mm; Coefficient [C] = 1.027, SE = 0.220, P < .001), jaw (C = 1.077, SE = 0.391, P = .007), and particle size (C = 0.788, SE = 0.380, P = .041) were found to significantly influence VC.

Table 5.

Multivariable linear regression analysis for vertical central socket height.

	Coefficient	SE	β	t	P
Intercept	1.584	2.488	0.000	0.636	.526
Age	−0.025	0.019	−0.124	−1.349	.181
Gender	0.001	0.396	0.000	0.003	.998
Jaw	1.077	0.391	0.249	2.755	.007
Tooth Position	0.169	0.391	0.039	0.433	.666
Healing Duration	0.241	0.185	0.117	1.303	.196
Buccal Wall Thickness	0.208	0.250	0.079	0.832	.408
Lingual/Palatal Wall Thickness	−0.138	0.225	−0.059	0.613	.542
Socket Integrity (No. of socket walls with height ≥5 mm)	1.027	0.220	0.425	4.660	.000
Particle Size	0.788	0.380	0.182	2.073	.041

SE, standard error.

Discussion

It is well known that ridge width resorption of ∼50% will occur after tooth extraction and extensive bone resorption will occur in the molar regions,³¹ therefore necessitating timely intervention procedures to facilitate subsequent implant placement.

To the best of the authors’ knowledge, this retrospective cohort study is the first to evaluate the radiographic changes of the alveolar ridge and implant feasibility after ARP in periodontally compromised molar extraction sockets using two DBBM particle sizes. Our findings revealed the following: (1) a trend that LP DBBM graft resulted in greater central socket height and volume compared with SP DBBM graft, but no statistically significant differences were observed; (2) changes in buccal and lingual ridge height and horizontal ridge width were similar between groups; (3) in sockets with vertical buccal and lingual/palatal bone plate discrepancies of >2 mm, significant reductions in horizontal ridge width measured at 1 and 3 mm below the crest were found compared to discrepancies of ≤2 mm for both groups; and (4) although similar, fewer sites required complex bone augmentation procedures in the LP group.

The present study supports previous conclusions that ARP maintains alveolar ridge dimensions after extraction of periodontally compromised teeth.³² In addition, this study expands the current body of evidence that was determined to be lacking and of low certainty in a recent systematic review.¹¹ Our results demonstrated the efficacy of two commercially available DBBM particle sizes to maintain dimensions of periodontally compromised molar extraction sockets after ARP.

Socket height is a critical factor for subsequent prosthetically-driven implant placement. Vertical central socket height gain was substantial and comparable between groups (7.76-8.54 mm), in line with previous clinical studies. Two controlled clinical trials both reported a central socket height gain of ∼8.5 mm after ARP in periodontally compromised molars, supporting our findings.^33,34 Although significant socket wall resorption was inevitable after a mean healing duration of 6.4 months, superior buccal bone height resorption was observed for both groups compared with lingual bone height resorption (−1.90 and −1.97 mm [−23.9 and −31.69%] vs −0.97 and −0.99 mm [−16.42 and −18.74%]). This finding is supported by a randomized controlled trial (RCT) that reported similar absolute and relative changes in buccal and lingual bone height changes (−2.1 mm [−36.36%] vs −0.88 mm [−10.1%]).³²

No significance was detected between SP and LP groups for horizontal ridge width changes measured at 1, 3, and 5 mm subcrestally. Both particle sizes were effective in maintaining alveolar ridge width after ∼6.4 months of healing. The amount of absolute and relative horizontal change were comparable with a previous study using particulate DBBM and collagenated DBBM for ARP in periodontally compromised molars at 1, 3, and 5 mm subcrestally.³⁵ Horizontal ridge width changes at periodontally compromised non-molar ARP sites were substantially different at HW₁ (−2.6 vs −4.04 mm), HW₃ (−0.86 vs −2 mm), and HW₅ (0.34 vs −0.85) compared with our results at molar sites.²⁷ Our results were in line with a previous consensus report that found a mean reduction of 3.8 mm in horizontal ridge width and 1.24 mm in vertical ridge height.³⁶ Similar comparative studies in LRA also found no significant differences between particle size groups for horizontal ridge width change.²⁴ Sockets with baseline buccal and lingual/palatal wall height discrepancies over 2 mm showed significantly greater HW₁ and HW₃ reduction for both particle sizes (P < .05). This could be explained by the influence of socket anatomy and the assumption that a more contained defect better protects the site from collapse.³⁷ Interestingly, it was observed that at sites exhibiting a buccal/palatal bone height discrepancy of >2 mm as compared to ≤2 mm, the horizontal ridge width change measured at 1 mm (−4.73 ± 2.86 vs −2.54 ± 1.56, P < .01) and 3 mm (−2.53 ± 1.72 vs −1.49 ± 1.06, P < .05) below the crest reached a statistically significant difference in the LP group but not in the SP group. However, this finding was likely due the difference between the mean baseline lingual socket height for the LP subgroups (>2 mm: 5.84 mm; ≤2 mm: 7.36 mm) which may have skewed the results.

The clinical benefit of ARP in reducing the need for complex bone augmentation procedures such as staged GBR and LSFE at the time of implant placement is extensively documented in the literature.³⁸ Although exclusively analysing non-molar teeth, evidence has suggested that a 17.8 fold reduction in the need for ancillary bone augmentation procedures after ARP is present compared to spontaneous healing sites.³⁹ However, a recent RCT exclusively evaluating the posterior maxilla found that 38.9% (7/18) of patients treated with ARP using DBBM and DBBM with 10% collagen still required LSFE compared to 90% (9/10) of patients in the control group after 4 months of healing.⁴⁰ Previous studies have reported that 8.3% of periodontally compromised molar ARP sites required LSFE during implant placement compared to 37.2% of spontaneous healing sites (P = .001).³⁴ The frequency of ARP sites necessitating LSFE in our study falls within the range of previous studies (8.3%-38.9%) but differences could be attributed to heterogeneity of the analysed sockets and sample size variations. Clinical recommendations from the XV European Workshop in Periodontology consensus report have specifically emphasized the importance of ARP in posterior sites close in proximity to the sinus.⁴¹

As highlighted in previous consensus reports and systematic reviews, studies on ARP should also report volumetric dimensional outcomes instead of solely relying on linear data.^6,41 Overall, volumetric changes in our study were larger compared with previous studies.⁴² Heterogeneity in socket location, the extent of initial periodontal damage, and varying methods of measurement could have caused this difference. Increase in alveolar ridge volume was superior in the LP group (369.44 mm³ [23.78%] vs 239.13 mm³ [17.73%]), but no statistical significance was observed. The advantages of LP grafts have been attributed to better mechanical resistance that may prevent graft displacement,²⁵ greater porosity size which may hasten new bone formation,⁴³ and reduced osteoclastic bone resorption compared with SP grafts.⁴⁴

The effect of local factors on healing outcomes should be prioritized during clinical decision making. While age, gender, tooth position, healing duration, and socket wall thickness did not significantly influence vertical central socket height, significance was observed for jaw (P = .007), socket integrity (number of socket walls with height ≥5 mm; P < .001), and particle size (P = .041). Regarding jaw position, the mean VC after ARP was 7.394 mm in the maxilla and 8.844 mm in the mandible, significantly favouring mandibular sites (P = .007; Appendix Table 2). Although the impact of maxillary sinus proximity is absent in mandibular sites, variability present in the current sample may also explain this finding. Socket integrity, quantified by the presence of intact socket walls, showed a profound impact on central socket height. For each additional intact bone wall per socket, VC increased by 1.027 mm. Further analysis of the specific number of walls ≥5 mm found a statistically significant difference between groups (P = <.001; Appendix Table 3) and the differences were most pronounced in severely compromised sockets (1 or 2 bone walls ≥5 mm in height; Appendix Table 4). Favourable socket defects facilitate an increased quantity of bone cell migration, blood supply, and hard and soft tissue stability.⁴⁵ Although particle size significantly influenced central socket height, though modest, its effect was less pronounced compared to jaw and socket integrity. For clinical implications, our findings suggest that although the use of LP DBBM grafts for ARP in periodontally compromised molar extraction sockets may offer advantages for central socket height gain, anatomical factors are more important. These findings are in line with a consensus narrative review which highlighted that factors such as bone loss extent, bony wall presence, closed healing environment, space provision, and wound stability significantly impacts healing.⁴⁶

Several limitations of the present study should be mentioned. First, the retrospective nature of the study made it difficult to calibrate the surgical procedures and baseline patient and socket characteristics could not be entirely consistent. Although an effort was made to standardize the patient and socket characteristics between groups, studies with prospective design would provide increased overall evidence strength. Second, CBCTs were not taken immediately after extraction. Potential damage to the socket due to trauma during extraction, although unlikely in teeth diagnosed with stage III/IV periodontitis, were unable to be assessed. However, CBCT acquisition immediately after extraction is not common practice and would warrant an additional dose of radiation for the patient. Third, soft tissue, histomorphometric, and patient reported outcomes were unavailable and not within the scope of this radiographic study. Previous histomorphometric studies on ARP in compromised molar sites using collagenated DBBM have shown that, at 6 months, the mean percentages of mineralized tissue, non-mineralized tissue, and residual graft material were 30.87%, 55.05%, and 14.08%, respectively.³² However, it should be mentioned that the influence of new mineralized tissue formation on implant outcomes remains inconclusive and further research is needed to elucidate its impact for ARP in compromised molar sites. A recent narrative review highlighted the clinical importance of cost-effectiveness analyses for ARP.³⁸ There is a need for further evaluation in compromised molar sites, in which material utilization is greater than in single-root extraction sites. In this study, systemic factors were controlled for as potential confounders, and smokers were excluded. Previous studies have also demonstrated that smoking may adversely affect extraction site healing due to the vasoconstrictive effects of nicotine on the microvasculature, which may inhibit the expression of cytokines associated with neovascularization and osteoblast differentiation.⁴⁷ Further investigation in this population is warranted to determine the applicability of the present findings to smokers. The sample size for this study was limited and future large sample RCTs addressing these limitations are warranted to confirm the results of this study. With recent advancements in artificial intelligence, future studies could also develop tools to assist in identifying which patients could benefit the most from ARP prior to tooth extraction, considering various baseline anatomical conditions and biomaterial choice.

Conclusion

Within the limitations of this retrospective study, it can be concluded that alveolar ridge dimensions could be maintained regardless of DBBM particle size in periodontally compromised molar extraction sockets that have undergone ARP. While large particle DBBM appears to support central socket height gain after ARP, baseline anatomical factors such as socket integrity are more critical and should be prioritized during treatment planning.

Conflict of interests

The authors report no conflicts of interest related to the subject of this review.