Abbreviations
CT
computed tomography
HU
Hounsfield units
ICC
intraclass correlation coefficient
IGF
insulin-like growth factor
IMF
intermaxillary fixation
ORIF
open reduction and internal fixation
PDGF
platelet-derived growth factor
PRF
platelet-rich fibrin
PRP
platelet-rich plasma
ROI
region of interest
TGF-β1
transforming growth factor-beta 1
INTRODUCTION
The mandible is the most commonly fractured facial bone due to its anatomical prominence and exposure to trauma [1], accounting for 48%–73% of all maxillofacial fractures [1,2]. The goal of management is to restore occlusal and functional stability through anatomical reduction and rigid fixation. Open reduction and internal fixation (ORIF) using miniplates and screws, based on AO/ASIF (Arbeitsgemeinschaft für Osteosynthesefragen/Association for the Study of Internal Fixation) principles, remains the gold standard for displaced fractures of the symphysis, parasymphysis, body, and angle [3]. Bone union typically occurs within 2 weeks, with complete remodeling achieved by 6–12 weeks [4-6]. However, optimizing the biologi-cal environment to accelerate bone regeneration remains clinically important, particularly in complex or delayed-healing cases [7-9].
Platelet-rich fibrin (PRF) is a second-generation autologous platelet concentrate produced by simple centrifugation without anticoagulants [10]. It contains platelets, leukocytes, and growth factors such as transforming growth factor-beta 1 (TGF-β1), platelet-derived growth factor (PDGF), and insulin-like growth factor (IGF), which promote osteoblastic proliferation, angiogenesis, and collagen synthesis [11-13]. Its fibrin matrix provides sustained release of bioactive molecules for up to 14 days, potentially enhancing bone regeneration compared with platelet-rich plasma (PRP).
PRF has been successfully applied in socket preservation, sinus augmentation, periodontal regeneration, and the management of osteonecrosis [14-17], yet its use in maxillofacial trauma remains underexplored. Prior studies on mandibular fractures are limited by small sample sizes, variable protocols, and reliance on two-dimensional imaging [18-23]. Al Rayess et al. [20] and Al-Khawlani et al. [22] reported improved bone density at 3 months, while Elhamshary et al. [23] confirmed enhanced bone formation up to 6 months using computed tomography (CT). Despite encouraging results, existing evidence lacks statistical power and generalizability. CT provides a quantitative assessment of bone mineral density via Hounsfield units (HU), offering superior accuracy and reproducibility compared with two-dimensional radiography. HU measurements have been validated for evaluating bone regeneration, implant planning, and fracture healing [24,25]. Moreover, a split-mouth design minimizes interindividual variability, since each patient serves as their own control [26]. This prospective split-mouth clinical trial evaluated the effect of autologous PRF on bone regeneration in mandibular fractures using serial CT-based HU measurements preoperatively and at 2 weeks, 1 month, and 3 months. We hypothesized that adjunctive PRF with standard ORIF would significantly increase bone density compared with ORIF alone.
METHODS
Study design
This study was a prospective, single-center, single-blinded, split-mouth clinical trial conducted between December 2023 and September 2025. Each patient with bilateral mandibular fractures served as their own control: one side received ORIF+ PRF (experimental), and the contralateral side received ORIF alone (control).
Prior to study initiation, a coin toss was performed to determine side allocation. According to the result, the left side was designated as the intervention side (PRF), while the right side served as the control for all participants. This within-subject design reduced biological variability and enhanced statistical power. Although fracture locations were not necessarily identical bilaterally, the split-mouth design allowed each patient to serve as their own control, thereby minimizing interindividual variability in systemic and biologic healing responses. Given that biomechanical forces and vascular characteristics vary among mandibular regions, baseline bone density (HU) was evaluated between sides to confirm initial comparability prior to intervention, as reported in the Results section.
Outcome assessment was performed by two independent plastic surgeons (W.B. and K.A.) blinded to allocation. They evaluated postoperative CT scans to determine bone density. Due to the nature of the procedure, the operating surgeon was not blinded.
Sample size calculation
The primary outcome was bone density at 3 months. Based on Al Rayess et al. [20], mean HU values were 547.8±74.7 in the PRF side and 374.7±109.8 in controls (mean difference=173.2 HU). For a split-mouth design, within-subject correlation (r=0.267) was obtained from a pilot study (n=10). The standard deviation of paired differences (SDdiff) was calculated as:
Sample size was estimated using the paired-samples t-test formula:
With α=0.05 (two-sided), power=0.80, Zα/2 =1.96, and Zβ=0.84, the minimum required sample size was 12. Adjusting for 10% attrition yielded 14 participants. The calculation was confirmed using Stata MP 18 (StataCorp). Ultimately, 18 patients completed the study, exceeding the required number and improving statistical reliability.
Participants and eligibility criteria
Inclusion criteria
Adults aged ≥18 years presenting with bilateral mandibular fractures involving two discrete fracture sites on opposite sides of the mandible were eligible. Fractures had to be located at the symphysis, parasymphysis, body, or angle regions and be associated with malocclusion requiring surgical intervention.
Operative procedures
All surgical procedures were performed under general anesthesia using standard aseptic techniques. Temporary intermaxillary fixation (IMF) was applied to establish proper occlusion. Fracture sites were exposed via intraoral or extraoral approaches, depending on the anatomical location and the condition of any pre-existing wounds.
Once the fracture lines were identified, bony fragments were mobilized, and any interposed soft tissue or fibrin was carefully debrided. Anatomical reduction of the mandible was achieved, followed by rigid internal fixation using two 2.0-mm titanium miniplates and screws on each side. The control side was fixed first, followed by the experimental (PRF) side. This sequence was standardized in all cases to minimize potential biologic cross-contamination; specifically, fixing the control side first prevented unintended diffusion or spillover of PRF-derived growth factors to the contralateral fracture site and ensured that both sides were treated under comparable mechanical conditions before PRF application. This approach helped maintain the internal validity of the split-mouth design. After satisfactory reduction and bilateral confirmation of fixation stability, preparation of PRF was initiated.
Preparation of PRF
Ten milliliters of venous blood were drawn into two sterile 5-mL clot-activator tubes without anticoagulant and centrifuged at 3,000 rpm (=400 ×g) for 10 minutes (MRC SCEN-206). Three layers formed: platelet-poor plasma (top), PRF clot (middle), and red cells (bottom). The PRF clot was retrieved with sterile forceps, trimmed to remove red cells, and gently compressed between sterile gauze to form a membrane (Fig. 1).
Application of PRF to the fracture site
Following ORIF, the PRF membrane was placed along the fracture line of the experimental side before closure, adapted to bone, and secured with fine sutures to adjacent tissue or miniplate to prevent displacement. All steps were performed under sterile conditions (Fig. 2).
Postoperative management and radiographic evaluation
Standard care included intravenous antibiotics for 24–72 hours, oral antibiotics for 7 days, analgesics as needed, and a soft/liquid diet for 2–4 weeks. Sutures were removed at 7 days. Patients were reviewed at 2 weeks, 1 month, and 3 months for clinical and radiographic assessment.
Outcome measures and assessment
Primary outcome
Bone healing was quantified by mean HU on preoperative and 2-week, 1-month, and 3-month CT scans.
CT scanning protocol
All scans were performed on a Toshiba 128-slice CT (120 kVp, 200 mAs, 2-mm slice thickness) with bone-algorithm reconstruction.
HU measurement technique
Three 0.06 cm² regions of interest (ROIs), specifically the proximal, central, and distal regions, were placed along each fracture line (Fig. 3); the mean value of these ROIs represented site density. Standard bone window settings (level 350 HU, width 2,700 HU) ensured consistency. Two blinded plastic surgeons independently measured HU using the same protocol; inter-rater reliability was assessed via intraclass correlation coefficient (ICC; two-way mixed-effects, absolute agreement). The mean of both readings was used for analysis.
Secondary outcomes
Occlusal alignment was assessed and categorized as typical, minor malocclusion, or significant requiring correction. Postoperative complications included infection (purulent discharge, erythema, swelling), hardware failure, nonunion/malunion, or PRF-related adverse reactions. All evaluations were conducted by a blinded assessor at each follow-up.
Statistical analysis
Continuous data were summarized as mean±standard deviation or median (interquartile range), as appropriate; normality of paired differences was assessed using the Shapiro-Wilk test and visual diagnostics. Repeated-measures analysis of variance (ANOVA) with within-subject factors side (PRF vs. control) and time (preoperative, 2 weeks, 1 month, 3 months) evaluated omnibus effects; Greenhouse-Geisser corrections were applied when sphericity was violated. Upon a significant side by time interaction, PRF versus control was compared at each time point using pre-specified conditional follow-up tests with Holm adjustment for four comparisons; adjusted p-values were reported. A sensitivity analysis at 2 weeks used the Wilcoxon signed-rank test when assumptions were doubtful, with rankbased effect sizes and Hodges-Lehmann estimates (95% confidence interval [CI]). Additionally, percent change in HU was computed for three intervals (from preoperative to 2 weeks, from 2 weeks to 1 month, from 1 month to 3 months) per side per subject and compared between PRF and control using paired t-tests (or Wilcoxon if non-normal), with Holm adjustment across the three interval-wise tests. Missing data were handled per comparison (pairwise exclusion). All tests were two-sided with α=0.05; estimation (mean/median differences with 95% CIs) was emphasized alongside hypothesis testing. Analyses were conducted using JASP (version 0.95.2.0).
RESULTS
Patient enrollment and follow-up
From December 2023 to September 2025, 20 patients with bilateral mandibular fractures were enrolled; all met the inclusion criteria. Two were lost to follow-up (relocation, non-compliance), yielding 18 evaluable participants (90% retention). No protocol deviations, exclusions, or adverse withdrawals occurred (Fig. 4).
Baseline characteristics
All 20 randomized patients were male (mean age 33.8±10.1 years, range 18–52). Fracture sites included the symphysis (7.5%), parasymphysis (42.5%), body (10%), and angle (65%). All underwent ORIF with bilateral titanium miniplates and screws. Baseline data are shown in Table 1. Preoperative HU values did not differ between PRF and control sides (p>0.05).
Reliability of HU measurements
Prior to the main study, inter-rater reliability was evaluated using 10 representative CT images, which were independently assessed by two plastic surgeons. The intraclass correlation coefficient (ICC(3, k)) for absolute agreement (two-way mixed-effects model, mean measures) was 0.957 (95% CI, 0.825–0.989), demonstrating excellent measurement reliability. Based on this high level of concordance, all HU measurements in the main study were performed independently by two trained assessors, and the mean of their values was used for subsequent statistical analysis to minimize observer bias and random error.
Primary outcome: bone density over time
Before conducting inferential analyses, statistical assumptions were verified. Shapiro-Wilk tests confirmed that HU values at all time points were approximately normally distributed (p>0.05), except for the paired HU difference at 2 weeks (p=0.022). Mauchly’s test of sphericity indicated that the assumption of sphericity was violated for the main effect of time (W=0.446, χ² (5)=12.70, p=0.027), but not for the side by time interaction (W=0.701, χ² (5)=5.59, p=0.349). Therefore, Greenhouse-Geisser corrections were applied to all time-related effects in subsequent analyses.
Table 2 summarizes the mean HU values for the PRF and control sides, as well as the results of paired comparisons at each time point. On both sides, bone density increased progressively from baseline through 3 months postoperatively (Fig. 5).
At 2 weeks, the PRF side showed a numerically higher mean HU (441.4±118.1) than the control side (399.1±130.9), but the difference was not statistically significant (mean difference=42.2 HU, p=0.203). Because the distribution of paired differences deviated from normality (Shapiro-Wilk p=0.022), a confirmatory Wilcoxon signed-rank test was performed, yielding a consistent non-significant result (p=0.074).
At 1 month, the PRF side maintained higher HU values (554.1± 180.7) than the control side (479.3±132.7), with a mean difference of 74.8 HU (95% CI, –8.9 to 158.6; p=0.077; Cohen’s d=0.44).
At 3 months, the PRF-treated side demonstrated significantly greater bone density (700.0±152.1) compared to the control side (567.8±135.7). This difference was statistically significant (mean difference=132.2 HU; 95% CI, 62.3 to 202.1; t(17)=3.99; p<0.001) with a large effect size (Cohen’s d=0.94).
Repeated-measures ANOVA results
Two-way repeated-measures ANOVA revealed a significant main effect of time (F(2.05, 34.91)=123.87, p<0.001, partial η²=0.879), indicating a substantial increase in bone density across postoperative periods. The main effect of side (PRF vs Control) showed a non-significant trend favoring PRF (F(1, 17)=3.83, p=0.067, partial η²=0.184). Notably, a significant side by time interaction was detected (F(2.44, 41.55)=12.73, p<0.001, partial η²=0.428), suggesting that bone healing patterns differed between PRF and control sides (Table 3).
Post hoc pairwise comparisons with Holm correction indicated that PRF-treated sides had significantly higher HU values at 3 months (mean difference, 132.2 HU; 95% CI, 62.3–202.1; p<0.001; Cohen’s d=0.94), while differences at preoperative, 2-week, and 1-month intervals were not statistically significant (p>0.05). Because the 2-week difference deviated from normality, a confirmatory Wilcoxon signed-rank test was conducted, yielding consistent non-significant results (p=0.074).
Interval-specific analysis revealed no significant differences in HU gain between PRF and control sides. Mean differences were 19.3 HU (95% CI, –3.0 to 41.7; p=0.086) from baseline to 2 weeks, 2.0 HU (95% CI, –8.3 to 12.4; p=0.683) from 2 weeks to 1 month, and 11.0 HU (95% CI, –4.6 to 26.5; p=0.154) from 1 to 3 months. These findings suggest that PRF mainly contributes to the overall increase in bone density by 3 months rather than accelerating early-phase mineralization.
Secondary outcome
All patients maintained stable occlusion without malocclusion or occlusal disharmony at all follow-up visits (2 weeks, 1 month, and 3 months). No surgical site infections, hardware failures, nonunion or malunion, or PRF-related adverse reactions occurred during the follow-up period. All wounds healed uneventfully, and no patient required reoperation.
DISCUSSION
This prospective split-mouth trial demonstrated that adjunctive autologous PRF significantly enhanced bone regeneration in surgically treated mandibular fractures. At 3 months, PRFtreated sides showed 23% higher mean bone density (mean difference, 132.2 HU; p<0.001; Cohen’s d=0.94) than control sides treated with ORIF alone, with no increase in complications, confirming both efficacy and safety.
Comparable baseline density between PRF and control sides (p=0.927) supports the internal validity of the design, indicating that subsequent differences stem from the intervention. The absence of early differences (2 weeks, 1 month) is consistent with fracture-healing biology [27], in which early recovery relies mainly on mechanical stability [28,29]. PRF’s biological effects, mediated by sustained growth factor release and progenitor recruitment, emerge during the hard callus and remodeling phases (weeks 4–12) [30,31], explaining the late divergence in bone density. Interval analysis showed no significant differences in HU gain rates between PRF and control sides (p>0.05 for all intervals), although PRF consistently exhibited higher values. This result indicates that PRF enhances cumulative bone regeneration by 3 months rather than accelerating early mineralization.
By 3 months, PRF-treated sides demonstrated a substantial and statistically significant increase in bone density (mean difference: 132.2 HU, representing a 23% improvement), with a large effect size (Cohen’s d=0.94). This finding suggests that PRF markedly accelerates the transition from woven to mature lamellar bone during the remodeling phase [32]. The significant side by time interaction observed in the repeated-measures ANOVA (p<0.001) confirms that PRF modified the overall healing trajectory, producing the greatest enhancement during the late consolidation phase, when osteoblastic activity and mineral deposition peak [28]. The mean HU value of approximately 700 HU observed in the PRF group falls within the range reported for mandibular cortical bone (>600 HU) [33], suggesting near-complete restoration of bone mineralization by 3 months.
These findings concur with previous evidence. Al Rayess et al. [20] reported a 46% higher HU in PRF sides, also appearing at 3 months. Other studies using radiographic or CT methods similarly found enhanced density at 3–6 months [18,21-23], though early effects varied with study power and PRF protocol. Collectively, results from multiple trials support that PRF consistently improves bone healing for around 3 months, with early discrepancies likely due to methodological rather than biological differences.
PRF’s regenerative capacity is attributed to its fibrin matrix, which entraps platelets, leukocytes, and growth factors such as PDGF, TGF-β1, vascular endothelial growth factor, and IGF [11-13]. Unlike PRP, PRF provides sustained release over 7–14 days, promoting angiogenesis and osteoblastic differentiation. Its fibrin scaffold mimics the extracellular matrix, facilitating cell attachment and proliferation [11], while leukocytes modulate inflammation to favor regeneration [12]. Together, these mechanisms accelerate the transition from inflammation to mineralized bone.
Histological and animal studies corroborate these effects, showing enhanced vascularization, collagen organization, and bone strength with PRF use [34-36], consistent with our radiologic evidence of higher HU values.
Clinically, PRF appears to be a safe, cost-effective adjunct to ORIF, potentially benefiting patients at risk of delayed healing, such as elderly patients or those with comminuted fractures [7-9]. Nevertheless, evidence remains insufficient for universal adoption. Protocol heterogeneity, especially in centrifugation parameters, necessitates standardization before broader use [37].
Strengths of this study include the split-mouth design, which minimized intersubject variability, and the objective CT-based HU assessment, which demonstrated excellent reliability (ICC=0.957). Limitations include the modest sample size, the single-center setting, and a 3-month follow-up period that precludes long-term conclusions. Although HU values provide a quantitative assessment, they remain surrogate indicators of bone quality rather than direct measures of biomechanical strength.
Anatomical heterogeneity related to fracture location also warrants consideration. Mandibular regions are subject to differing biomechanical stress distributions and muscular forces— particularly between the corpus and angle—which may influence healing dynamics. Rahajoe et al. [38] demonstrated that fractures in these regions are exposed to distinct loading conditions and may exhibit variations in healing rate. Biomechanical modeling studies further support regional differences in stress patterns across the parasymphysis, body, and angle of the mandible [39], while variations in vascular supply may also contribute to differences in regenerative capacity [40].
In the present study, fracture locations were not necessarily identical bilaterally; however, the prospective split-mouth design allowed each patient to serve as their own control, thereby minimizing systemic interindividual variability. Furthermore, the similarity of baseline HU values between the experimental and control sides reduced the likelihood that regional anatomical variability materially influenced the interpretation of the outcome.
Future research should employ multicenter randomized controlled trials with standardized PRF protocols and extended follow-up to determine whether radiographic improvements translate into functional or biomechanical benefits. Comparative investigations among different platelet concentrates may further refine clinical indications and optimize patient outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Cited
Print
