Associations of Sinus Width, Septa, and Schneiderian Membrane Thickness with Bone Formation After Lateral Window Sinus Augmentation

2.1. Study Participants

2.2. Study Design

This prospective secondary analysis included 60 sinus augmentation sites from 42 patients with severely atrophic posterior maxillae (RBH ≤4 mm) treated between January 2022 and February 2024. All procedures followed a standardized lateral window sinus augmentation protocol. Ethical approval was obtained from the Ethics Committee for Biomedical Research at Can Tho University of Medicine and Pharmacy (Decision No. 504/PCT-HĐĐĐ).

Sample-size sensitivity analysis: This study was a secondary analysis of a fixed prospective cohort. Because the primary outcomes (NB and R) were continuous percentages, a single-proportion formula was not appropriate. For a two-sided correlation test with α=0.05 and 80% power, 59 independent observations are required to detect an absolute correlation of 0.36 using Fisher’s z transformation. Accordingly, 60 sites provide approximately 81% power to detect |r|=0.36 before accounting for within-patient clustering. The analysis therefore had sensitivity for moderate associations but could not exclude smaller effects.

The biopsy site was the observational unit. Because some patients contributed more than one site, site-level observations may be correlated. Primary analyses were therefore interpreted as exploratory. This cohort partially overlaps with a previously published descriptive histological study; the present analysis addresses a distinct question concerning CBCT-derived anatomical predictors [12].

2.3. Radiographic Assessment of Maxillary Sinus Anatomy

Preoperative cone-beam computed tomography (CBCT) scans were used to evaluate anatomical characteristics of the maxillary sinus. All CBCT images were obtained using standardized imaging parameters and analyzed using digital imaging software.

Measurements were performed on cross-sectional images perpendicular to the alveolar ridge at the planned implant location. The following radiographic variables were assessed:

• Residual bone height (RBH): vertical distance from the alveolar crest to the sinus floor
• Sinus width (SW): horizontal distance between medial and lateral sinus walls measured 10 mm above the sinus floor
• Schneiderian membrane thickness at baseline (S0)
• Membrane thickness changes after sinus augmentation and during follow-up (S1–S5)
• Presence or absence of sinus septa

Two calibrated examiners independently performed the radiographic measurements following a calibration exercise using CBCT sections that were not included in the principal analysis. Disagreements were resolved by consensus. Inter-examiner reliability was assessed in a randomly selected subset of 20 sinus sites. Continuous measurements were evaluated using a two-way random-effects intraclass correlation coefficient for absolute agreement based on single measurements [ICC(2,1)], whereas agreement for sinus septa detection was evaluated using Cohen’s κ and percentage agreement.

2.4. CBCT Acquisition Protocol

CBCT examinations were acquired using an ORTHOPHOS XG 3D system (Sirona Dental Systems GmbH, Bensheim, Germany). Images were obtained using an 8 × 8 cm field of view, an isotropic voxel size of 0.20 mm, a tube voltage of 90 kVp, a tube current of 8 mA, and an acquisition time of 17 seconds. Images were reconstructed and evaluated using SIDEXIS software. Patients were positioned according to the manufacturer’s instructions, and all measurements were performed on cross-sectional reconstructions perpendicular to the alveolar ridge at the planned implant location. CBCT imaging was performed using the smallest diagnostically acceptable field of view in accordance with ALADA/ALADAIP principles.

Sinus width was measured on a cross-sectional image perpendicular to the alveolar ridge at the planned implant site, as the horizontal distance between the medial and lateral sinus walls at a line 10 mm superior to the sinus floor, following the approach used in previous sinus-dimension studies [20, 21]. Figure 1 illustrates the measurement landmarks.

2.5. Surgical Procedure

All sinus augmentation procedures were performed using a standardized lateral window approach under local anesthesia and antibiotic prophylaxis. A lateral window was prepared using a piezoelectric device, followed by Schneiderian membrane elevation and graft placement according to the previously published protocol [12]. Three intraoperative membrane perforations occurred (5.0%); all were managed during surgery. No postoperative infections or graft-related complications were recorded during healing [13].

Autologous platelet-rich fibrin was prepared using a standardized centrifugation protocol. The use of a uniform surgical approach across all patients aimed to minimize procedural variability and allow evaluation of anatomical factors as primary predictors of healing outcomes.

After a 6-month healing period, bone core biopsies were obtained from the planned implant sites immediately before implant placement, which was performed in accordance with the manufacturer’s instructions.

2.6. Histomorphometric Outcome Assessment

Biopsy processing and histomorphometric procedures followed the previously published protocol [12]. Briefly, specimens were fixed in buffered formalin, decalcified in EDTA, embedded in paraffin, sectioned at approximately 5 μm, and stained with hematoxylin and eosin.

Histomorphometric analysis was conducted using ImageJ software (National Institutes of Health, USA) to quantify the proportion of newly formed bone (NB) and residual graft material (R). For each specimen, three non-consecutive sections corresponding to coronal, middle, and apical regions were analyzed to improve representativeness.

Areas not identified as mineralized tissue or residual graft particles were categorized as connective tissue or marrow-like structures. All measurements were performed independently by two blinded examiners to ensure reproducibility.

Histomorphometric variables were considered outcome parameters for evaluating the relationship between sinus anatomical characteristics and bone regeneration.

A representative hematoxylin and eosin–stained histological section obtained six months after lateral window sinus augmentation, showing newly formed bone, residual graft particles, and intervening connective tissue/marrow-like spaces. Scale bar = 100 μm (Fig. 2).

2.7. Statistical Analysis

Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Continuous variables were summarized as mean ± standard deviation and range, and categorical variables as counts and percentages. Normality was evaluated using the Shapiro-Wilk test.

Because several variables were not normally distributed, exploratory two-group comparisons were conducted using the Mann-Whitney U test. Site-level associations between continuous CBCT-derived predictors and the histomorphometric outcomes NB and R were explored using Spearman's rank correlation coefficient (ρ), with approximate 95% confidence intervals derived via Fisher's z-transformation.

Because some patients contributed more than one biopsy site, site-level observations were potentially correlated. Therefore, linear mixed-effects models with a patient-specific random intercept were additionally fitted for NB and R to account for within-patient clustering. Continuous anatomical predictors were entered per 1-mm increase, and sinus septum status was entered as present versus absent; adjusted β coefficients with 95% confidence intervals were reported. Intraclass correlation coefficients were calculated from the random-intercept variance components. These cluster-aware models were interpreted as exploratory because this was a secondary analysis of a fixed cohort with a limited number of patients. NB and R are compositional percentages and are mathematically dependent on the remaining tissue component, which further supports cautious interpretation of inferential results.

All statistical tests were two-tailed, and p < 0.05 was considered statistically significant.

Inter-examiner reliability was excellent for sinus width, baseline Schneiderian membrane thickness, and post-augmentation membrane thickness, with ICC (2,1) values of 0.94, 0.91, and 0.93, respectively. Agreement for sinus septa detection was almost perfect, with a Cohen’s κ of 0.89 and an overall percentage agreement of 96.7%. All reliability estimates were statistically significant, and the corresponding 95% confidence intervals indicated good to excellent reproducibility.

3.1. Study Population and Distribution of Implant Sites

A total of 42 patients (60 sinus augmentation sites) were included in the study. The mean patient age was 53.57 ± 12.90 years (range: 23–72 years). Sixteen patients were male (38.1%), and 26 were female (61.9%). The majority of implant sites were located in the molar region (78.3%), whereas 21.7% were positioned in the premolar region. All sites presented residual bone height ≤ 4 mm and were therefore considered appropriate for lateral window sinus augmentation. Baseline characteristics of implant location are summarized in Table 1.

CBCT analysis revealed variability in the anatomical morphology of the maxillary sinus among implant sites. The molar region demonstrated a significantly greater sinus width measured 10 mm above the sinus floor than the premolar region (19.75 ± 3.44 mm vs 15.63 ± 4.40 mm, p = 0.005).

No statistically significant differences were observed between premolar and molar regions in baseline Schneiderian membrane thickness (S0) or post-augmentation membrane thickness (S1). Furthermore, changes in membrane thickness at follow-up time points (S2–S5) did not differ significantly between anatomical regions (all p > 0.05).

Although sinus width differed between premolar and molar sites, no statistically significant differences in membrane-thickness measures were detected (Table 2). These findings were interpreted as exploratory due to unequal group sizes and site clustering.

Sinus septa were identified in a subset of implant sites. Comparison of histomorphometric outcomes between sites with and without septa showed no statistically significant differences in newly formed bone (NB) or residual graft material (R) (NB: p = 0.897; R: p = 0.327).

No statistically significant difference in NB or R was detected between sites with and without septa (Table 3). Because this was a null result in a modest, unbalanced sample, it should be interpreted as failure to detect an association rather than evidence that septa have no biological effect.

Baseline Schneiderian membrane thickness was comparable between sites without and with sinus septa [median: 1.00 mm (IQR, 0.40–1.53) versus 1.00 mm (IQR, 0.50–1.35); Mann–Whitney U = 268.50; p = 0.715]. Similarly, no significant difference was observed in post-augmentation membrane thickness [2.80 mm (IQR, 2.08–5.10) versus 3.20 mm (IQR, 2.33–4.13); U = 284.00; p = 0.941]. (Table 4).

3.2. Associations between CBCT-derived Variables and Histomorphometric Outcomes

At the biopsy-site level, sinus width was not significantly correlated with newly formed bone percentage (Spearman ρ = 0.064; 95% CI: −0.193 to 0.312; p = 0.629) or residual graft material percentage (ρ = 0.106; 95% CI: −0.152 to 0.351; p = 0.420) (Table 5).

Baseline Schneiderian membrane thickness demonstrated a weak positive association with newly formed bone that approached, but did not reach, statistical significance (ρ = 0.252; 95% CI: −0.002 to 0.476; p = 0.052). Baseline membrane thickness was not significantly associated with residual graft material (ρ = −0.062; 95% CI: −0.311 to 0.195; p = 0.637) (Table 5).

Post-augmentation membrane thickness was not significantly associated with newly formed bone (ρ = 0.106; 95% CI: −0.152 to 0.350; p = 0.420) or residual graft material (ρ = −0.056; 95% CI: −0.305 to 0.201; p = 0.672). These site-level findings were considered exploratory because multiple biopsy sites may have originated from the same patient (Table 5).

In patient-clustered mixed-effects models, sinus width was not significantly associated with newly formed bone percentage (β = 0.22 percentage points per 1-mm increase; 95% CI: −0.53 to 0.97; p = 0.560) or residual graft material percentage (β = 0.19; 95% CI: −0.30 to 0.68; p = 0.443) (Table 6).

Baseline Schneiderian membrane thickness showed a weak positive association with newly formed bone that approached, but did not reach, statistical significance (β = 1.36 percentage points per 1-mm increase; 95% CI: −0.08 to 2.80; p = 0.064). No significant association was observed with residual graft material (Table 6).

Post-augmentation membrane thickness and sinus septa status were not significantly associated with either histomorphometric outcome. The estimated within-patient correlation was modest, with ICC values of 0.105 for newly formed bone and 0.069 for residual graft material (Table 7).

4.1. Anatomical Variability as a Potential Determinant of Regenerative Outcomes

Anatomical characteristics of the maxillary sinus have been proposed as potential modifiers of bone regeneration due to their influence on vascular supply, graft containment, and spatial distribution of biomaterials [6, 7]. In particular, wider sinus cavities have been hypothesized to reduce the proximity of graft particles to vascularized sinus walls, potentially increasing the diffusion distance for angiogenic and osteogenic signaling [20, 21]. From a mechanistic perspective, reduced vascular proximity could theoretically delay mineralization and maturation of newly formed bone.

However, the results of the present study do not support a clinically meaningful association between sinus width and histomorphometric outcomes. Despite statistically significant differences in sinus width between premolar and molar regions, newly formed bone and residual graft material did not differ significantly across anatomical configurations. These findings suggest that dimensional variation of the sinus cavity alone may not be a strong determinant of bone regeneration when surgical technique and graft stability are standardized.

One possible interpretation is that sinus augmentation creates a contained healing environment in which clot stability and scaffold properties may mitigate the influence of anatomical variability. The lateral window approach enables controlled placement of graft material and may reduce the functional impact of differences in sinus dimensions. This interpretation is consistent with previous studies reporting that bone formation may occur predictably across a range of sinus morphologies [13, 22].

4.2. The Present Exploratory Analysis did not Detect a Moderate Site-level Association between Sinus Width and Histomorphometric Outcomes

Because of limited power for small effects and the limited precision of the patient-clustered estimates, this finding should not be interpreted as evidence that sinus width is biologically irrelevant.

Sinus septa are frequently considered a technical challenge during sinus elevation procedures due to their association with increased risk of membrane perforation [10, 23]. However, their influence on regenerative outcomes remains unclear.

In the present study, septa were not associated with statistically significant differences in NB or R. This finding addresses biological healing rather than surgical difficulty. Three membrane perforations occurred in the cohort, but the available aggregate data did not support stratified or adjusted analysis by perforation status. Because perforation may confound healing, future analyses should include it as a covariate when individual-level data are available.

From a causal perspective, septa may act as procedural modifiers that influence surgical difficulty but do not directly affect osteogenesis once the graft material is successfully placed and stabilized. Therefore, the presence of septa should be interpreted as a factor that requires surgical adaptation rather than as a predictor of regenerative capacity [24, 25].

4.3. Schneiderian Membrane Response as an Intermediate Variable

Changes in Schneiderian membrane thickness following sinus augmentation may represent a physiological response to surgical manipulation rather than a pathological process [12, 21]. Membrane thickening has been described as a transient inflammatory or adaptive response associated with wound healing.

In the present cohort, no statistically significant differences in membrane thickness were detected by region or septa status, and no significant site-level associations with histomorphometry were identified. All follow-up CBCT examinations were obtained for clinically indicated treatment assessment rather than solely for research purposes, using standardized acquisition parameters and a limited field of view consistent with ALADA/ALADAIP principles [26].

Within a causal framework, membrane thickness may be considered a mediator reflecting tissue response to surgery, but its magnitude may not necessarily predict the extent of mineralized tissue formation.

4.4. Relative Importance of Anatomical Versus Procedural Factors

The absence of statistically significant associations should not be interpreted as proof that anatomy is unimportant. One possible explanation is that standardized graft placement and membrane management attenuated moderate anatomical effects; another is limited power, residual confounding, or measurement error [27]. Therefore, procedural factors may be important, but their relative contribution cannot be established by this observational analysis.

Clinically, CBCT measurements should be used primarily for diagnosis, surgical access planning, and risk assessment. The present data do not support the use of sinus width, membrane thickness, or septa alone as validated predictors of histologic bone formation.

4.5. Methodological Considerations in Causal Interpretation

Interpretation of causal relationships in the present study requires consideration of potential confounding factors. Histomorphometric outcomes are influenced by multiple biological variables, including patient-specific healing capacity, vascular supply, systemic conditions, and biomaterial properties. Because all patients were treated using the same grafting protocol, the present study does not allow separation of the individual effects of biomaterial composition and anatomical variability.

In addition, histomorphometric variables represent compositional data that sum to 100%, which may introduce mathematical dependency between tissue components. Consequently, statistical associations should be interpreted with caution and treated as exploratory.

Another methodological consideration is that biopsy specimens represent localized tissue environments rather than the entire augmented sinus. Although histomorphometry provides high-resolution evaluation of tissue composition, it may not fully capture three-dimensional structural changes within the grafted volume.

4.6. Clinical Implications

The present exploratory findings indicate that sinus anatomical variation alone was not a validated predictor of six-month histomorphometric outcomes in this cohort. Anatomical variation should therefore continue to be considered in surgical planning rather than interpreted as a contraindication or as a definitive predictor of regenerative capacity.

Clinical emphasis should remain on surgical precision, preservation and management of the Schneiderian membrane, and stable graft placement. CBCT remains essential for diagnosis, access planning, and risk assessment, but the present data do not establish sinus width, membrane thickness, or septa as independent predictors of histologic bone formation.

The findings support the feasibility of lateral window sinus augmentation across varied anatomical presentations when appropriate protocols are followed; however, they do not establish anatomical equivalence or causality.

4.7. Limitations and Future Directions

Several limitations should be acknowledged. First, this was a secondary analysis of a fixed cohort and had sensitivity primarily for moderate correlations (approximately |r|=0.36); smaller effects could have been missed. Second, although patient-level random-intercept mixed-effects models were applied to account for clustering of multiple sites within patients, the limited number of patients and the small number of repeated sites per patient may have reduced the precision of the variance-component and association estimates. Third, measurement reliability was high/substantial for the evaluated radiographic and histomorphometric variables, reducing-but not eliminating-the possibility that measurement error attenuated the observed associations. Fourth, membrane perforation was a potential confounder, and only aggregate complication data were available. Fifth, repeated CBCT time points require explicit clinical indications and acquisition parameters. Sixth, histomorphometry was obtained at one time point from localized 3-mm cores and may not represent the entire augmented volume. Finally, the single-center cohort, an unbalanced premolar/molar distribution, and the lack of multivariable adjustment limit generalizability and causal inference.

Future studies incorporating longitudinal imaging, volumetric analysis, and larger multicenter cohorts may provide additional insight into the relationship between sinus morphology and regenerative outcomes. Integration of CBCT-derived anatomical variables with biomechanical parameters and implant stability measurements may further clarify the role of anatomical variability in implant success.