The Incidence of Maxillary Sinus Membrane Perforation During Endoscopically Assessed Crestal Sinus Floor Elevation: A Pilot Study

Maxillary sinus floor elevation is a reconstructive surgical procedure commonly used to augment a deficient posterior maxilla to accommodate dental implant placement when pneumatizaton of the maxillary sinus is present. The conventional method for maxillary sinus elevation requires surgical access through the zygomatic buttress of the maxilla, followed by elevation of the sinus membrane itself and direct insertion of bone graft material. Although Boyne and James published this technique in 1980, it is reported that Tatum described this surgical method first during a 1977 lecture and later in a publication.^1,2  This procedure is also referred as the lateral approach to the maxillary sinus floor.¹  Since the technique was first introduced, numerous grafting materials have been evaluated by a number of authors.^2–9 

Although the sinus graft is considered to be a relatively invasive surgical procedure, the incidence of reported surgical and postsurgical complications is relatively low.^10,11  In fact, reported perforation rates vary from as low as 7% to as high as 58%.^{11,14,16–25} The most commonly reported surgical complication for the lateral approach technique is perforation of the sinus membrane.^10–15  Maxillary sinus membrane perforations, as described in the literature, also are strongly linked to the development of postoperative complications such as acute or chronic sinus infection, edema, bleeding, wound dehiscence, loss of the bone graft material, and a disruption of normal sinus physiologic function.^{11–14,17,26–28}  Conversely, some authors have reported no association between sinus membrane perforations and implant survival,^{10,11,14,17,29}  while others have attributed implant failure directly to the sinus membrane perforation.^13,16,30 

In 1994, Summers³¹  introduced a less invasive procedure for sinus membrane elevation along with dental implant placement. The “Summers technique,” often referred to “osteotome/crestal sinus membrane elevation,” or OCSME, is recommended for patients with at least 5.0 to 6.0 mm of adequate alveolar bone below the sinus floor.³¹  During the surgery, the sinus membrane is elevated with osteotomes from a crestal approach through the osteotomy prepared for dental implant placement, thereby avoiding surgical access through the zygomatic buttress of the maxilla. Consequently, OCSME is a visually restrictive procedure and is considered to be a sensitive technique with inherent limitations, especially when direct visual examination of the maxillary sinus membrane is desired or required. Despite the widespread clinical application of this surgical technique, along with the advent of multiple surgical variations, few studies have evaluated the incidence of maxillary sinus membrane perforations.^31–36  In a 2008 systematic review, Tan et al³⁷  reported transcrestal membrane perforation varied widely between 0% and 21.4%. In addition, the incidence of potential postoperative complications may be greater than in the lateral approach technique, possibly due to recognized visual limitations with the OCSME method.³⁸  In 2000, Wiltfang et al³⁸  reported that endoscope-controlled sinus floor augmentation may actually have a lower postoperative complication rate for the transcrestal procedure in patients with 4.0 to 8.0 mm of vertical bone height below the sinus floor.

The intraoperative use of sinuscopy as described by Grunenberg and Gerlach in 1990 for maxillary sinus elevation procedures allows for exclusion of sinus pathology intraoperatively, control of the bone graft position, a reduced risk of sinus membrane perforations, and fewer postoperative complications.^39,40 

The aim of the present pilot study was multifold. The primary goal was to evaluate the surgical outcomes of 3 transcrestal maxillary sinus membrane elevation procedures (Summers' osteotome sinus floor elevation [OSFE], Dask crestal bone planing, and the Implantium System with the Dentium Osteotome Kit). A second goal was the determination of the Schneiderian membrane perforation frequency and its prevalence using a maxillary sinus endoscope. The third goal was to compare and recognize membrane perforations using direct visual endoscopic assessment with periapical radiographs and cone-beam computerized tomography (CBCT) in fresh human cadaver's heads.

Twenty fresh human cadaver heads, ranging in age from 33 to 74 years, were used for an endoscopic and a radiographic evaluation of the maxillary sinus membrane perforation during crestal sinus floor elevation. The inclusion criteria for the selection of the cadaver heads were (1) partial or complete edentulism of the posterior maxilla limited to the premolar-molar areas, (2) a preoperative residual alveolar ridge of at least 5.0 mm in width and with at least 5.0 mm to 7.0 mm of crestal bone height beneath the maxillary sinus floor as measured on the baseline preoperative periapical and CBCT radiographs, and (3) posterior dentate cadavers with premolars and maxillary molars, without periapical pathology, that could be extracted and the interadicular boney septum instrumented. The point of reference for interadicular bone height measurement was the most coronal bony crest within the furcation that measured on CBCT between 5.0 and 7.0 mm in height to the sinus floor. The presence of pathologic alteration of the maxillary sinus, such as large polypoidal mucosal thickening, mucous retention cyst, scarred sinus mucosa from previous operation, or sinus opacification resulted in exclusion of those sinus cavities from the study.

A total of 40 sinuses were randomly divided into 3 groups: 10 maxillary sinuses were assigned to the control group, 15 sinuses were assigned to the experimental group A, and 15 sinuses also were assigned to the experimental group B. Using the inclusion/exclusion criteria described above, a total of 15 sinus cavities were excluded from the study, resulting in 7 sinus cavities remaining in the control group, 12 in experimental group A, and 6 in experimental group B, for a total of 25 sinus cavities being used (Table 1).

The specimens receiving the control treatment (control group) were treated with the osteotome-crestal sinus floor elevation (OCSFE) protocol, also commonly referred to as the “Summers' technique.”³¹ 

The objective of this procedure is to conserve all of the bone from the site and to impact it upward. This selective displacement requires the use of a special set of osteotomes (Biomet 3I, Global Headquarters, Palm Beach Gardens, Fla) that creates an osteotomy as the instrument advances past the sinus floor. In the OCSFE technique, the bone mass, which is impacted upward, elevates the sinus floor. In this procedure, there is no dissection of the Schneiderian membrane with other instruments. However, for the present study, Summers' technique was slightly modified to allow the osteotome to penetrate inside the sinus cavity. Alloplastic graft material was added before surgical implant placement (Figure 1).

Specimens in experimental group A were treated with the DASK Sinus Lift and Elevation Kit (Dentium, Cypress, Calif), which used a series of specially designed surgical drills with stoppers and curettes to reach to the sinus boundary, eliminate the cortical bony floor, and displace the maxillary sinus membrane by mechanical and hydraulic pressure for subsequent bone grafting and implant placement (Figure 2).

Specimens in experimental group B received a surgical protocol in which the crestal bone was eliminated by site preparation. Sequential conventional drills with stoppers from the Implantium Surgical Kit (Dentium) were used with extreme care to penetrate to a depth of 1.0 to 2.0 mm from the maxillary sinus boundary. Convex osteotomes were used to displace the antral floor as part of the procedure using a Dentium Osteotome Kit (Dentium) followed by bone grafting and implant placement (Figure 3).

In all of the aforementioned surgical procedures, bone graft was inserted in the receptor site in a standard volume not exceeding a complete syringe of about 0.25 mL in each instrumented site (Osteon Lifting, 70% HA scaffold coated with 30% β-TCP, particle size 0.5 to 1.0 mm, Dentium). Implants (Dentium, SuperLine Fixture Platform 4 mm/body 3.8 mm with lengths of 10.0 mm and 12.0 mm) were placed at bone level in all the respective prepared sites using an insertion torque value between 30 and 35 N/cm. One investigator performed all the procedures. In relation to the sensitivity and the predictability of the values measured with periapical films and CBCT scans to asses sinus membrane perforation, 3 dentists, with different levels of clinical experience (E1, E2, and E3), compared their readings using the endoscope video recording findings as the predicate in the assessment.

Each of the remaining 25 cadaver sinus cavities was scanned in a Cone Beam Volumetric Tomography and Panoramic Dental Imaging System Unit (i-CAT Imaging Sciences International, Hatfield, Penn). Periapical digital radiographs (Dental X-Ray, Trophy Radiologie CCX Digital, City, Tex) were taken of each of the 25 the posterior maxillary areas as well. These images were used to determine the anatomical integrity, dimensions, and relevant conditions of the sinus floor prior to the surgical procedures and also used to ensure the inclusion and exclusion criteria of the specimens. Digital radiographs were accomplished using the long cone paralleling technique.

An otolaryngologist, using a 4.0-mm Storz-Hopkins endoscope with a 30° viewing angle (Karl Storz, Tuttlingen, Germany), connected to a high-definition camera and flat screen monitor, visualized the middle meatus. The uncinate process was removed using endonasal instruments, and the natural ostium of the maxillary sinus was identified and enlarged (Figure 4). A 70° endoscope (Karl Storz, Tuttlingen, Germany) was then used to visualize the undisturbed floor of the maxillary sinus. Both the still photographic images and the video recordings were collected throughout the subsequent surgical procedures using a digital recording system (Dyonics Digital 3-Chip Video Camera Console Endoscopy, Zoi Surgical Inc, Chesterland, Ohio). The deformation capacity of the Schneiderian membrane during instrumentation, progressive addition of the bone graft, and implant placement were actually observed in real time by the otolaryngologist, who did not give any feedback to the surgeon when making these observations.

The sinus membrane elevation was standardized and limited to 5.0 mm in all situations. Each elevation was measured with a millimetric-marked osteotome, sinus dome curette, or a depth gauge throughout the procedure. The inward sinus floor border of the osteotomy served as landmark. The intended height was recorded in the center of the osteotomy as soon as sinus membrane resistance was detected with light pressure.

Visualization of the gradual elevation of the maxillary sinus membrane at instrumentation, graft insertion, and implant placement was recorded from multiple angles. The presence or absence of a membrane tear or perforation was determined and scored from 0 to 2. A score of 0 was given when no perforation was detected, a score 1 for a 0.5- to 5.0-mm size perforation located apical or slightly lateral to the most elevated membrane area, or a score of 2 for a perforation larger than 5.0 mm with a loss of the dome shape and exposure of the implant fixture within the sinus cavity. In addition, to ensure that membrane tears too small to be visualized by the endoscope were not missed, methylene blue dye was instilled into the sinus lumen and followed by direct visualization intraorally to detect any penetration.

The otolaryngologist assessed the size of the perforation based on a comparison with an object of a known size present in the same surgical field (ie, 2.8 mm as implant apex platform diameter, 3.8 mm as implant body diameter, or 3.0 mm as the suction tip diameter). A linear measurement was made of each perforation in the value of the largest diameter present (Figure 5).

All the cadaver heads were scanned postoperatively with the CBCT unit to inspect and identify any sinus membrane perforations using an individual-labeled putty station to support the cadaver head. Periapical digital radiographs were also taken and read for the presence or absence of sinus membrane perforations. Two evaluators (E1 and E3) made the initial postoperative radiographic interpretations, and 3 evaluators (E1, E2, and E3) completed the second set of readings. Three levels of expertise used in this study were 1 evaluator with advanced experience (E1), 1 with intermediate-level experience (E2), and 1 novice evaluator (E3).

The degree of perforations and timing were evaluated with the aid of a nasal endoscopy as described before. These endoscopic findings were then compared to the perforation readings obtained from the CBCT and the periapical digital radiographs. Sinus conditions were scored, and those scores were then recorded.

All readers were calibrated for diagnosing the perforation size according to the ranking scores. A thickness of 0.04 mm was present on every postoperative scan slice. No measurements were taken on any CBCT sections or digital radiographs. The evaluators were instructed to visually screen the corresponding CBCT sections using the serial thin-slice transplanar images, in all 3 dimensions (midsagittal, coronal, and axial), at different degrees of density and also to estimate the size of the perforation and assign a score. The radiographic evaluation was performed twice: first, 4 weeks postoperatively, and second, 12 weeks after the surgical experiment.

The statistical analyses were performed using the commercially available software SPSS 16.0 V (SPSS Inc, Chicago, Ill). The postoperative findings were reported with descriptive statistics, frequency tables, and bar charts. Pearson chi-square, Kendall's tau (τ) coefficient, and Spearman's rho nonparametric test were used to analyze the experimental findings and to evaluate the postoperative radiographic results (CBCT and digital periapical radiographs) of the sinus membrane perforation readings among the evaluators and also against the endoscope findings on 2 different sets of readings. The level of significance was set at P < .05.

After the intragroup CBCT preoperative evaluation, 15 sinus cavities did not meet the inclusion criteria and were excluded from the study. The remaining 25 sinuses were distributed into the following 3 categories: (1) completely edentulous arches, (2) partially edentulous arches, and (3) sites that required extractions (Table 2). According to the exclusion criteria, the bilateral sinus membrane elevation procedure was performed on 9 specimens (n = 18), and a unilateral sinus elevation was performed on 7 specimens (n = 7). Upon endoscopic evaluation on all sinus floor elevation procedures (n = 25), uneven tension of the mucosa was observed and judged to be responsible for the membrane overexpansion, which in 10 cases was followed by tissue laceration. One specimen received a score of 2, and 9 perforations received a score of 1 (Table 3). None of the sinuses with a 0 score showed any evidence of penetration of methylene blue dye intraorally.

According to the Pearson chi-square analysis, no significant differences were found among the 3 groups in terms of perforation size (P = .79). Such an outcome suggests that there was no difference between the sinus floor elevation methods at the level of significance of P < .05 with respect to the data obtained from endoscope observation. According to the endoscope video-recording comparative results, membrane perforations mainly occurred during implant placement and were located in the area where the greatest amount of membrane distension occurred (P = .04; Table 4). In those perforation areas, the mucosa was observed to become significantly thinner prior to disruption. All the sinus membrane perforations were video recorded and correlated with the corresponding group and subgroup (Table 5). A gradual enlargement of the perforation also was observed and noted at the final implant position.

Overall, the presence of sinus membrane perforation was recorded in 40% of the elevations. Thirty-six percent of the specimens in the small perforation group had a score of 1, and 4% had a perforation score of 2. Endoscopy results indicated a perforation rate of 4% correlated with the surgical instrumentation alone, 12% linked to the graft insertion procedure, and 24% associated with the implant placement.

The Pearson chi-square test reached a value of P = .04, indicating that perforations were statistically more likely to occur during the time of implant insertion. Of 25 implants placed (7 in the control group, 12 in group A, and 6 implants in group B), 4 implants were placed as immediate implants in fresh extraction sites.

Intraexaminer reliability and precision were determined comparing the first and second radiographic readings (CBCT and periapical radiographs) with Kendall's test. Evaluator E1 was highly correlated with himself on CBCT (P = .006, r = .551) and periapical radiographic readings (P = .001, r = .640). Evaluator E1's first CBCT readings were also strongly correlated with the endoscope findings (P = .016, r = .473), where P < .05. This suggests that an advanced level of experience facilitates the diagnosis of membrane perforation on CBCT. On the other hand, evaluator E3 did not agree with his own determinations of whether perforations occurred on repeated CBCT readings (P = .068, r = .367), nor did his radiographic findings correlate with the endoscope findings (P = .740, r = .065). He did show consistency on periapical radiographs (P = .023, r = .445) readings, however. There was significant interexaminer correlation between the E1 and E2 evaluators on the CBCT readings where all 3 examiners participated (P = .017, r = .481). However, evaluators E1 and E2 were not correlated with the third evaluator E3 (P1 = .920, P2 = .144). When the periapical radiographs were screened, there was again a valid interexaminer correlation between evaluators E1 and E2 (P = .005, r = .545) at a significance level of P < .05 on Kendall's nonparametric analysis. The third evaluator, E3, seldom correlated with the other 2 examiners (Table 6; Figure 6).

Nkenke et al¹⁷  reviewed the literature for reports on membrane perforations with OSFE and concluded that endoscopic control is recommended when the sinus membrane is lifted more than 3.0 mm. Another endoscopic study⁴²  found that the sinus floor may be elevated up to 5.0 mm without perforating the sinus membrane, while other endoscopic studies have illustrated the risk of membrane perforation while performing transalveolar sinus floor elevation.^17,43  Based on the observations from our study, there are several factors that should be taken into account when instrumenting the sinus membrane. The elevation height and level of sinus membrane distension should be regarded as technique sensitive, and both parameters are affected by sinus membrane condition (in vivo or ex vivo). Factors related to the graft material used, such as the amount, the level of hydration, the size and configuration of bony particles, and so forth, also may influence the outcomes. Equally important is the implant apex geometry (sharp or rounded apex) as it may play an important role when the implant placement is in close vicinity to the elevated sinus membrane.

Nkenke et al¹⁷  also reported that the augmentation volume of the maxillary sinus should be kept to a minimum. They suggested that with the lateral window technique, an average bone augmentation volume of 3.50 ± 1.33 cm³ is needed when the residual bone height is about 5.0 mm. However, according to some clinicians, when the osteotome sinus floor elevation is performed, a minimum bone augmentation volume of only 0.5 cm³/mL is required.¹⁷ 

As a surgical observation from this study, the endoscopic recordings demonstrated that in 20 of the cadaver specimens, a thin layer of graft material was present over the implant apex after placement. This interposed graft layer will increase the initial planned elevation height by an additional 0.5 to 2.0 mm. Because of the irregular and sharp geometry of the bone graft particles, the risk of membrane disruption was potentiated, especially in thinner areas of the expanded or stretched membrane. As the implant reached the final position, the graft was displaced in an apical or lateral direction, creating perforations or enlargement of previously detected membrane tears at a total elevation height of approximately 6.5 to 7.0 mm.

One of the principal findings of this pilot study was the response of the Schneiderian membrane to the different surgical events of the 3 sinus membrane elevation procedures. The 6 perforations that occurred during the implant insertion were distributed among the preexisting anatomic status. In fact, there were 2 perforations for each subgroup: edentulous quadrants, extraction sites, and partially edentulous sites. Nonetheless, 5 of the 6 perforations occurred in specimens of group A, which is noteworthy.

The osteotome penetration is difficult to control during the application of mallet impact pressure, and this may result in an unwanted advancement of the instruments and/or the bone graft material into the maxillary sinus cavity with the increased potential for membrane perforation. On the other hand, one perforation was detected during the instrumentation in a group B specimen, while 2 perforations occurred during bone graft insertion specimens in the control group. The use of manual and rotating instruments, with adjustable stop devices, restricted the working action to the vertical amount of residual bone. Crestal bone planing burs were proved to be safe even when in close proximity to the sinus membrane. The stopping devices of variable heights were able to control the drill advancement every 2.0 mm and prevent the accidental penetration into the sinus cavity. Schneiderian membrane detachment was initiated by internal irrigation of the crestal drills that exerted a light hydraulic pressure. This hydraulic force introduced to the sinus cavity was believed to apply pressure to the deeper layer of periosteum-like connective tissue but not to blood vessels. Experimental Valsalva maneuver could not be performed in cadavers, but with the manual insertion of the disc curette (diameter of 3.3 mm; Dentium Sinus Kit), the elasticity of the sinus membrane could be easily felt and verified. The use of disc curette in this fashion was considered the most reliable test of membrane integrity.

In separate publications, Nkenke et al¹⁷  described the use of a maxillary sinus endoscope. When the procedure was carried out without visualization, there was little opportunity to detect membrane perforations. Consequently, Nkenke et al¹⁷  concluded that the perforation that was visible by endoscopy was related to a negative Valsalva maneuver, which showed the limited effectiveness of this test. With regard to the use of osteotomes in human formalin-fixed cadavers, Reiser et al⁴⁴  observed sinus membrane perforation in 6 of 25 implants (24%), with the risk of perforation increasing according to the increase in the extent of sinus floor elevation needing to be obtained. The increase in the height of the implant sites by the osteotome technique alone, up to the point at which the concomitant spontaneous dissection of the sinus membrane in the periphery of the elevated region stopped and the tension of the sinus membrane revealed the risk of rupture, ranged from 2.0 to 5.0 mm according to Nkenke et al.¹⁷  Two studies claim the OSFE methods limit elevation height to approximately 5.0 mm.^33,45 

In the present study, the OSFE technique confirmed the critical elevation of approximately 5.0 mm, with membrane tears observed during the graft placement (3 of 4 perforations) and mechanical instrumentation (1 of 4) in the control group and experimental group B, respectively. In this technique, an even-pressure distribution of the graft material under the elevated membrane is necessary.

Also in the present study, the fresh anatomic specimens were preserved only by freezing and not embedded with formalin. Chan and Titze⁴⁶  showed that postmortem tissue changes observed on fresh and frozen specimens have little impact on the mechanical properties of tissue. In a similar study using 20 fresh cadavers, Pommer⁴⁷  found that these changes usually increase tissue fragility; therefore, perforation of the Schneiderian membrane (mean thickness: 0.90 mm) occurred at a mean tension of 7.3 N/mm. The membrane could be stretched only to 132.6% of its original size in 1-dimensional elongation and to 124.7% in 2-dimensional elongation. Thicker membranes demonstrated significantly higher load limits (the mean modulus of elasticity accounted 0.058 GPa, and the mean adhesion force between sinus membrane and bone surface was 0.05 N/mm).⁴⁷  However, there are no guidelines for the assessment and classification of mucosal findings in the maxillary sinus before sinus floor elevation.⁴⁸ 

For visualization of the maxillary sinuses, conventional computerized tomography is considered the diagnostic method of choice,⁴⁹  but CBCT imaging is becoming more popular and an important technique for the diagnosis and treatment planning in dental practice.^50–53  The postoperative CBCT images obtained in this study were able to offer enough information for the experienced readers (E1, E2) to predict the presence of perforation findings with good correlation to the video results. In comparison with the endoscopic evidence of perforation, the periapical radiographs might not be a very reliable postoperative diagnostic method to detect perforations because all 3 readers had inaccurate results.

Based on the parameters of this pilot study, the following conclusions were drawn:

1. 1.

   Fresh human cadaver heads can be used in studies to simulate surgical procedures related to the crestal maxillary sinus membrane elevation.

2. 2.

   There were no statistically significant differences in the maxillary sinus membrane perforation rates among the 3 surgical methods: modified Summers technique (control group), crestal bone planing (group A), and implantium OSFE (group B).

3. 3.

   Maxillary sinus membrane perforations were not detected by the operator during any of the 3 surgical procedures.

4. 4.

   Six of the total 10 perforations (24%) occurred at the time of implant placement.

5. 5.

   The CBCT and periapical digital radiographs were judged to be less reliable than the endoscope for the detection of Schnederian membrane perforations. Overall, the CBCT readings were determined to be more accurate than the periapical radiographs when compared with the endoscope video results.

6. 6.

   False-positive or -negative readings on periapical radiographs and CBCT scans were influenced by the experience level of the individual evaluating these images.

7. 7.

   Clinicians with experience in sinus procedures were able to more predictably recognize the presence of the sinus membrane perforations.

8. 8.

   Endoscopy was considered the most reliable method to confirm actual perforations and was used to determine the interexaminer and intraexaminer precision and reliability compared with postoperative readings of the digital periapical radiographs and CBCT scans.

Within the limitations of this experiment, we identified the inherited problems of the sinus floor elevation technique. A study with a larger sample and ideal conditions of the maxillary sinus membrane may allow for more detailed evaluation of surgical techniques and outcomes. Additional live human studies with histological assessment of the Schneiderian membrane, correlating healthy and compromised membranes, may help to establish the clinical relevance of the crestal sinus floor augmentation in the event of perforations during the surgical procedures.

CBCT

cone-beam computerized tomography

OCSME

osteotome/crestal sinus membrane elevation

OCSFE

osteotome/crestal sinus floor elevation

OSFE

osteotome sinus floor elevation

SFE

sinus floor elevation

The authors wish to acknowledge Assistant Professor John B. Won, DDS, for his valuable assistance with the scientific writing.