Transient swelling of the Schneiderian membrane after transversal sinus augmentation: a pilot study
For figures, tables and references we refer the reader to the original paper.
In the posterior area of the upper jaw, the bone available for implant placement often is limited due to the presence of the maxillary sinus. The height of the alveolar process can, however, be increased via sinus augmentation procedures.
The maxillary sinus is internally lined with a thin mucosa of “ciliated” respiratory epithelium, which is continuous with that of the nose (Ritter & Lee 1978; McGowan et al. 1993). The normal antral mucosa, however, is thinner (approximately 1 mm thick) and less vascular than the nasal mucosa. The ciliated respiratory epithelium has a transport function for fluids like pus and mucus towards the internal ostium (Stamberger 1986). This ostium is situated at the cranial side and connects the maxillary sinus to the middle meatus of the nasal cavity (May et al. 1990). The mucosa of the sinus received several names: antral mucosa, maxillary sinus lining/mucosa or Schneiderian membrane.
The average dimensions of the adult maxillary sinus are as follows: 25–35 mm in width, 36–45 mm in height and 38–45 mm in length (Eckert-Mobius 1954), with a total volume of 15 cm3 (van den Bergh et al. 2000). The thickness of the Schneiderian membrane, when measured on cone-beam CT (Janner et al. 2011), showed a wide range (minimum of 0.2 mm, maximum of 35 mm). The highest mean values (143 patients included), ranging from 2.2 to 3.1 mm, were found in the mid-sagittal regions of the sinus, whereas the mean values on the lateral and medial aspects varied between 0.9 and 1.8 mm. Aimetti et al. (2008) discovered, via biopsies, a positive correlation between sinus membrane thickness and gingival biotype (1.3 mm in individuals with thick, 0.6 mm with thin gingival tissues).
Today, the lateral window sinus lift is widely used and considered reliable, particularly when autogenous bone is used, with implant survival rates comparable with implants placed in pristine bone (for review, see Wallace & Froum 2003; Del Fabbro et al. 2004; Jensen & Terheyden 2009; Nkenke & Stelzle 2009; Esposito et al. 2010). Summers (1994) described a less invasive, one-stage, technique for sinus floor elevation called the osteotome sinus floor elevation. Osteotomes, applied via a crestal approach, are used to elevate the floor of the sinus and to advance bone substitute beyond the level of the original sinus floor, as such elevating the mucosal lining. Also for this approach, high implant survival rates have been reported (for review see Tan et al. 2008; Jensen & Terheyden 2009; Nkenke & Stelzle 2009; Esposito et al. 2010).
It should, however, be noted that the grafting material is placed blindly into the space below the sinus membrane, with the risk of perforating it (Nkenke et al. 2002; Berengo et al. 2004). An endoscopic study revealed that the sinus floor may be successfully elevated up to 5 mm without perforating the membrane (Engelke & Deckwer 1997). Berengo et al. (2004) performed eight Summers' interventions (16 implants) under sinuscopic monitoring. A small-sized perforation of the sinus membrane was visualized at two implant sites without significant loss of graft material and without negative long-term consequences. At 12 implants, a clear distention of the membrane was observed enfolding the implant tips, the in between and circumferential areas. For four implants, the vertical distension was restricted to the area surrounding the implant apices, with two of them showing a perforation. None of the cases with perforation showed clinical signs of ongoing sinus pathology.
Before placing any grafting material, the condition of the Schneiderian membrane should, however, be checked, for example, via the Vasalva manoeuvre. In a recent study from Pjetursson et al. (2009a), this test was found to be positive in 11% of the cases. In such situations, implant insertion should be aborted, or the implant should be placed without grafting material. Membrane perforation has indeed been identified as the most frequent complication encountered (3.8%) with trans-alveolar sinus floor elevation, while postoperative infection is rare (0.8%) (Tan et al. 2008).
The longitudinal response of the Schneiderian membrane to a trans-alveolar sinus floor elevation has, to our knowledge, not been explored. Therefore, this longitudinal study aimed to follow changes in the thickness of the Schneiderian membrane over time, after a summers' technique with the use of a bone substitute and simultaneous implant placement.
Material and methods
Ten consecutive patients, referred to the Department of Periodontology (University Hospital, Catholic University Leuven) for implant therapy in the maxillary premolar/molar area in combination with a sinus augmentation, were included in this pilot study. Besides standard inclusion criteria for implant therapy (e.g. general health, periodontitis free, good oral hygiene,), the following extra conditions had to be fulfilled: pathology-free sinus (as visualized on X-ray images), ≥4 mm alveolar bone height and ≥5 mm alveolar bone width at implant site. In all patients, one or more implants were placed together with the trans-alveolar sinus floor elevation, and each time BioOss® (Geistlich, Wolhusen, Switzerland) was used as bone substitute. All patients took antibiotics for 4–5 days after the intervention, mostly amoxicillin.
Prior to implant insertion, and at 1 week and at 1 month after the implant placement/sinus augmentation, a small field, low-dose cone-beam CT was taken, visualizing the maxillary sinuses (three patients bilateral sinuses, seven patients single sinus).
The protocol was approved by the Ethical committee of the Catholic University Leuven (Leuven, Belgium), and informed consent was obtained.
The images were obtained with a 3D Accuitomo 170 CBCT (Morita, Kyoto, Japan) with a voxel size of 0.15 mm. Operating parameters were set at 5.0–7.0 mA, 80 kV, and exposure time was 17.5 s. For all CBCT images, a limited field of view (FOV) of 6 × 6 cm or 8 × 8 cm was selected. The data were reconstructed with slices at an interval of 0.5 mm.
Evaluation of the images
The CBCT images were evaluated by a trained and experienced postgraduate student (DL), not directly involved in the treatment and follow-up of the patients. For calibration and evaluation of intra-examiner reliability, 10 randomly selected cases were measured twice on two different days, resulting in a mean difference of 0.12 mm per image (range of 0–0.4 mm). For this project, each measurement was repeated twice (Janner et al. 2011), and the mean was calculated. When the difference between two values was ≥0.2 mm, a third measurement was performed (Bornstein et al. 2011).
The dimensions of the Schneiderian membrane (in millimetres) were scored on cross-sectional images at nine distinct positions (Fig. 1a,b), as suggested by Janner et al. (2011). In summary, this means:
(a) Landmarks for the measurement of the mucosa thickness in cross-sectional images in the (i) anterior [a: between the root tips of the maxillary premolars (PM1-2)], middle [b: the apex of the first molar (M1)] and posterior region [c: region between the apex of the second and third maxillary molars (M2-3)]; (ii) for each location at: the onset of the zygomatic process for the lateral measurement (lateral); the deepest point of the sinus floor for the mid-sagittal measurement (middle); the ipsilateral bony floor of the nose for the medial measurement (medial). (b) Schematic overview of the nine measurement points in the evaluated maxillary sinuses (axial slice). Modified according to Janner et al. (2011). (c) Schematic overview of the augmented site with measured distances: A = entire augmentation, including the cortical plate, B = the gap between the augmentation and the former cortical plate, C = thickness of the Schneiderian membrane.
in a sagittal plane: between the root tips of the maxillary premolars (PM1-2), the apex of the first molar (M1) and the region between the apex of the second and third maxillary molars (M2-3), respectively. In edentulous areas, distances between premolar roots were set at 7 mm and between molar roots at 8 mm.
in coronal plane: the onset of the zygomatic process for the lateral measurement (lat), the deepest point of the sinus floor for the mid-sagittal measurement (mid) and the ipsilateral bony floor of the nose for the medial measurement (med).
All measurements of the mucosal thickness were performed with a special software tool to the nearest 0.1 mm (lightbox, IMPAX5, Agfa Healthcare, Mortsel, Belgium) and always perpendicularly to the underlying bone, starting at the underlying bony plate and ending at the mucosal surface (Janner et al. 2011).
At implant locations, where the Summers' technique was applied, following parameters were measured: the obtained sinus augmentation including the cortical plate, the gap between cortical plate and augmentation and the thickness of the Schneiderian membrane above the augmented area (Fig. 1c).
With the same software, the density of the Schneiderian membrane (artificial Hounsfield units) was also scored, at the same locations as described earlier.
To analyse the changes in the thickness of the Schneiderian membrane, the values at week 1 and month 1 were subtracted from the baseline values, to examine whether these values were statistically different from zero (Wilcoxon signed rank test and Bonferroni correction for simultaneous testing). The null hypothesis of no change was rejected at P < 0.05. For the aforementioned analyses, a S-plus 8.0.4 software for Linus has been used.
Prior to the surgical intervention, the Schneiderian membrane was thin (Table 1). In the middle, a mean thickness of 1.3–1.4 mm was measured (SD ranging from 0.7 to 1.4) for the three respective sites (PM1-2, M1, M2-3). The corresponding values for medial measurements were 0.7 up to 1.0 mm (SD ranging from 0.2 to 0.7), and for lateral sites 0.5–0.7 mm (SD ranging from 0.2 to 0.3).
Table 1. Thickness of Scheiderian membrane (expressed in mm) prior to the surgical intervention and the difference in thickness of Scheiderian membrane in mm (subtracting baseline from observations at week 1 and month 1; mean, standard deviation, median, 95% confidence interval and P-value for difference from zero) scored at nine specific reference points (in mesiodistal direction: between the root tips of the maxillary premolars (PM1-2), the apex of the first molar (M1) and the region between the apex of the second and third maxillary molars (M2-3), in lateromedial direction: the onset of the zygomatic process for the lateral measurement (lateral), the deepest point of the sinus floor for the mid-sagittal measurement (middle); the ipsi-lateral bony floor of the nose for the medial measurement (medial), respectively
One week after the trans-alveolar sinus floor elevation (Table 1), the swelling of the Schneiderian membrane showed mean values of 9.2 mm in the middle (SD 0.3; P < 0.005), 7.2 mm medially (SD 2.5; P < 0.005) and 3,9 mm laterally (SD 0,3; P < 0.05), respectively, resulting in an overall mean of 6.7 mm (SD 2.6). The largest swellings were found in the PM1-2 and M1 region, but the entire mucosa in contact with the lower border of the sinus was swollen. These effects are clearly illustrated by a clinical case (Fig. 2).
CBCT images from a patient who received implants in first and second quadrant, together with a trans-alveolar sinus floor elevation, in conjunction with of BioOss application. The first CBCT (a) was taken 4 months before implant insertion, and the 17 was extracted due to terminal periodontitis. The Schneiderian membrane showed a clear swelling caused by the aforementioned infections (white arrow). An intra-oral image showed that the swelling of the membrane disappeared soon afterwards. One week after the surgical intervention (b), both sinuses show a clear swelling of their membrane (≥10 mm), which fully disappeared after 1 month (c). The yellow arrows show the cortical border of the sinus that has been pushed cranially. Cross-sectional CBCT image from the implant position 15 of the previous patient, with a nice fill of the augmented area in the sinus (d). The Schneiderian membrane is nearly invisible. The two intra-oral images show the implant 1 week after insertion (e) and 6 months after implant loading (f). Also, these images indicate a healthy situation. These images were only taken for research purposes and should, based on the ALARA principle, not routinely be taken in combination with a CBCT.
One month after the intervention, the Schneiderian membrane re-obtained its normal thickness, with differences compared with baseline of ≤1 mm (Table 1).
At the apex of the inserted implants (Table 2), similar observations were made at week 1. At month 1, a negligible swelling of 1–2 mm could be seen, and the low number of observations did, however, not allow statistical analysis. The mean elevation of the floor of the sinus, using this technique, was 4.7 mm after 1 week. One month after the intervention, the height decreased to 3.3 mm.
Table 2. Mean height (in mm) of the area that was elevated and grafted with the bone substitute material after 1 week and 1 month (measurement A in Fig. 1c) and the mean distance (in mm) between the augmentation and the cortical plate that was used to line the cortex of the maxillary sinus and that was elevated with the Summers' technique (measurement B in Fig. 1c), measured from the apex of the inserted implant
Our data indicate a clear transient swelling of the Schneiderian membrane after a trans-alveolar sinus elevation, disappearing within 1 month after the intervention. To the knowledge of the authors, this is the first “longitudinal” report on the response of the Schneiderian membrane to a sinus augmentation procedure. The clinical importance of this observation remains unclear. On one hand, it shows the tremendous response of the Schneiderian membrane after a “minor” trauma, with a swelling of 5–10 times its size. On the other hand, it also became obvious that this swelling disappeared after 1 month, without obvious sequelae. The fact that none of the patients reported any side effects (sinus pathology, sinus infection, signs of inflammation, pain, data not shown) is even more reassuring. One might of course question what will happen in an infection-prone (over reacting) patient, in a case of a narrow sinus, a small ostium, etc. More research is needed to answer these questions.
The reason for this swelling is also not well understood. On could think of a postoperative bleeding, pushing the membrane away from the bony floor of the sinus. This is, however, not confirmed by the endoscopic monitoring, clearly pointing to a very local loosening of the membrane, with sometimes even a depression between two adjacent augmented implant sites (Berengo et al. 2004). The arbitrary Hounsfield values measured for the swollen area, when compared with those of the soft tissue in the tuber maxillaries, nearly always scored higher values (higher density) in most sinuses. In those cases where the elevated cortical border of the sinus floor was clearly detectable, it became obvious that the swelling occurred primarily above it, thus within the membrane itself (Table 2).
The mean thickness of the Schneiderian membrane was evaluated in a post-mortem study with values ranging from 0.3 to 0.8 mm in 10 unfixed fresh cadavers without signs of sinusitis (Tos & Mogensen 1979). In a similar study on 20 fresh cadavers, Pommer et al. (2009) found a mean thickness value of 0.1 ± 0.05 mm (range 0.02–0.35 mm). In a study analysing sinus biopsies from healthy subjects, Aimetti et al. (2008) measured a mean thickness of 1.0 ± 0.5 mm. All three papers reported large interindividual variability.
Only one paper reported longitudinal observations on changes in the sinus membrane (Peleg et al. 1999). 21 patients with a lateral wall sinus augmentation were examined via msCT pre- and 8–10 months postoperatively, and no major thickening of the membrane could be observed after the intervention. They unfortunately had no observations within the first weeks after the intervention.
Radiographic observations on the dimensions of the Schneiderian membrane are scarce. Most studies applied different measurement scales and classifications, from the maximum mucosal thickness in millimetres (Rak et al. 1991) to the simple description of the shape of mucosal thickening (Soikkonen & Ainamo 1995; Patel et al. 1996). The early reports even revealed difficulty in visualizing normal mucoperiosteal structures in the paranasal sinuses through CT (Min et al. 1994) or MRI (Patel et al. 1996), due to a limited spatial resolution and a low detection limit for mucosal tissues. The mucosa could only be seen at a thickness of 2 mm or above, and therefore, historically 2 mm was considered a reliable threshold for pathological mucosal swelling (Cagici et al. 2009). More recently, Janner et al. (2011) examined 143 consecutive patients, referred for implant therapy, via CBCT and observed large interindividual variability in the thickness of the Schneiderian membrane, with values ranging from 0.2 mm (minimum) to 34.6 mm (maximum). They also reported a clear discrepancy between mean and median values, suggesting that some of the subjects exhibited very high mucosal thickness values. The median values for all nine regions, as used in this article, ranged from 0.5 to 1.4 mm. Only gender was identified by them as an influencing factor for maxillary sinus mucosa thickness in the multivariate regression analysis, with male subjects having higher mean values. Endodontic, periodontal and peri-apical status of the dentition in the region of interest had no statistically significant influence. Another analysis of 330 maxillary sinuses on CT scans revealed a significantly higher prevalence of mucosal thickening in the proximity to restored teeth (Connor et al. 2000). Vallo et al. (2010) reported a thicker Schneiderian in maxillary sinuses adjacent to periodontal and endodontic lesions. Furthermore, smoking and the presence of rhinologic diseases correlated with increased mucosal thickness. The month in which images were taken was reported to have an influence on pathologic findings in the maxillary sinuses, with autumn having a higher prevalence of cystic lesions (Rodrigues et al. 2009; Vallo et al. 2010), and an increased prevalence of mucosal thickening during winter (Tarp et al. 2000).
In a study on intra-oral radiographs, Engström et al. (1988) found a mean sinus floor mucosal thickness of 5.6 ± 6.1 mm before and 0.6 ± 1.6 mm after extensive periodontal therapy (13 patients with advanced periodontal disease). Gingival thickness and gender were the only genetically determined parameters to reliably predict sinus membrane thickness, being higher in patients with a thick gingival biotype and lower in women (Aimetti et al. 2008; Vallo et al. 2010).
A clear reduction in sinus augmentation, apical to the implant, between week 1 and month 1 (1.4 mm) was observed. Although the small number of cases jeopardizes a firm statement, one might conclude that this should be seen as a healing process (shrinkage and remodelling). The latter was previously observed by Brägger et al. (2004) and Pjetursson et al. (2009b), who both reported a gradual reduction in augmented bone height at the apex of the implant up to 1 and 3 years, respectively.
In all the cases, Bio-Oss has been used as bone substitute. Whether this has any impact on the degree of membrane swelling is very questionable. Unfortunately, to the knowledge of the authors, no other studies are available with other bone substitutes, so that a comparison cannot be made.
Within the limitations of this pilot study, one can conclude that the Schneiderian membrane responds with a significant “transient” swelling (5–10× its size) during the first weeks of healing after a Summers' technique.
Conflict of interest and source of funding
This article has been prepared without any sources of institutional, private or corporate financial support, and there are no potential conflicts of interests.