Relationship between cortical bone thickness or computerized tomography-derived bone density values and implant stability



Download 26.84 Kb.
Date conversion05.12.2016
Size26.84 Kb.
Relationship between cortical bone thickness or computerized tomography-derived bone density values and implant stability

For figures, tables and references we refer the reader to the original paper.

Osseointegration has been defined as ‘the direct structural and functional connection between living bone and the surface of a load-bearing implant, without the interposition of soft tissue’ (Albrektsson et al. 1983; Branemark et al. 1985). It has been identified as the main indicator of implant success (Laney et al. 1986; Lioubavina-Hack et al. 2006). Primary implant stability has been acknowledged as an essential criterion for later achievement of such osseointegration. Primary stability was traditionally assessed by the practitioner's manual examination. This method evidently suffered from being highly subjective and examiner dependent. Several alternative methods have been proposed to provide the practitioner with an objective indicator of the implant's stability. Among these, resonance frequency analysis (RFA) measured using the Osstell device and damping capacity assessment are considered as reliable, reproducible and easy to use (Olive & Aparicio 1990; Meredith et al. 1996).

The introduction of computerized tomography (CT) in implant therapy (Schwarz et al. 1987a, 1987b) allowed a tridimensional visualization of bone, especially in the bucco-lingual direction, which was not available on traditional panoramic images. Recording bone mineral density (BMD) and subsequent translation into Hounsfield units (HU) also became possible, provided that the necessary calibration had been carried out previously (Kalender & Suess 1987; Nickoloff et al. 1988; Hill et al. 2005). As a result of these properties, and especially after the development of reliable low-dose scan protocols (Ekestubbe et al. 1996; Loubele et al. 2005), CT can be considered as an acceptable tool for implant therapy besides the cone beam technology, which, thanks to the lower radiation dose, might be considered the standard tool for the future.

So far, only a few studies have assessed the relationship between CT parameters and implant stability (Beer et al. 2003; Ikumi & Tsutsumi 2005). This study aims to examine the relationship between different bone parameters (HU values and cortical bone thickness, type of anchorage, presence of bony dehiscencies) and implant parameters (length, diameter) on the one hand and primary implant stability on the other.

Material and methods

Patients, implants and surgery

Twenty-four patients (16 females, eight males, average age 58 years) who presented a fully edentulous upper jaw were selected for the present study. None of the patients selected suffered from a systemic disease. They were all scheduled for an implant-supported overdenture. Every patient received a CT scan before implant planning. Per patient, four to six Straumann SLA implants were inserted (two-stage procedure) in the upper jaw, summing to a total of 136 implants with diameters of 3.3 mm (n=43) or 4.1 mm (n=93), and lengths of 6 mm (n=23), 8 mm (n=13), 10 mm (n=14), 12 mm (n=29) or 14 mm (n=57) (Table 1). Implants were placed via a surgical template for optimal localization. During surgery, the presence of any bony dehiscence (n=22) and bi-cortical anchorage (n=80) was recorded. This study was approved by the ethical committee of the Catholic University of Leuven, and all patients gave written informed consent. These patients participated in a study on the outcome of GBR for the coverage of buccal dehiscencies.

Table 1.  Distribution of the implants by length and width

Bone density and cortical thickness estimation

All CTs were performed using a Siemens Volume Zoom helical CT (Siemens, Erlangen, Germany). Identical settings were applied for all patients: 120 kV, 90 mAs, 0.75 mm slice thickness and 0.3 mm slice increment. Data were stored in Dicom format. These Dicom files were loaded in a planning software (Simplant, Materialise dental, Leuven, Belgium), which included 3D reconstruction, segmentation and drawing of the reference curve.

Following abutment placement, an occlusal photography of the jaw was taken with the help of intra-oral mirrors. Using Photoshop software (Photoshop CS2, Adobe Systems incorporated, San Jose, CA, USA), this photograph was superimposed to the reference axial CT slice of the CT examination taken before the implant surgery (Fig. 1). This procedure allowed localization of the osteotomy sites on the reference axial slice and therefore the corresponding cross-sectional CT slices. The abutment connection was performed 6–9 months after implant insertion.

Figure 1.

Superposition of the occlusal photograph to computerized tomography reference axial slice.

Straumann implant models of appropriate diameters and lengths were selected in Simplant's implant library and placed at the corresponding sites. Implant inclination was determined after comparison with per-operative photographs and post-operative peri-apical radiographs and/or panoramic images.

HU values of the volume simulating the osteotomy site as well as the surrounding 1-mm-thick bone shell were measured. HU values of the spongious part and of the coronal cortical plate were also recorded separately. Finally, a cylinder simulating the coronal 3 mm of the implant was placed at the implant-receiving site and the corresponding density of the area was recorded (Fig. 2). Moreover, the cortical thickness was scored at four sites of implant–bone contact, distributed around the implant. For the latter, a mean value was calculated (Fig. 2).

Figure 2.

The measured bone sectors. Yellow, implant; pink, surrounding shell; red, cortical bone part; light green, spongious bone part; blue, coronal 3 mm of the implant; dark green & orange, cortical bone thickness.

Parameters for implant stability

RFA measurements were performed in two perpendicular directions [mesio-distal (M-D) and oro-facial (O-F)], twice in every direction, and a mean value was calculated. RFA was performed at insertion and at implant loading (6–9 months post-operatively). For this study, the wireless device was used (Ostell Mentor®, Integration diagnostics AB, Sävedalen, Sweden).

Periotest (PTV) measurements (Periotest®, Gulden-Medizintechnik, Bensheim an der Bergstraße, Germany) were performed at loading time with the stylus placed perpendicular to the abutment in the oro-facial direction, twice for every implant; only a mean value of the two measurements was used.

Statistical analysis

A Linear mixed model was applied to compare the RFA values and the parameters: implant diameter, implant length, presence of a bony dehiscence, or type of cortical anchorage (mono or bi-cortical). Next, the relations between implant stability parameters and HU values as well as cortical thickness at the coronal implant–bone interface were assessed. The patient factor was taken each time as a random factor and residuals were controlled for their normality by a normal quantile plot. The same test was finally used to relate the results of different stability values (PTV vs. RFA; RFA at insertion vs. loading; RFA oro-facial vs. mesio-distal).

Stepwise regression analyses were included to determine the influential variables or variables' interactions on RFA or PTV. The variables included were HU scores, implant length, implant diameter, type of anchorage, cortical thickness and presence or absence of bony dehiscence and their interactions. Again, the patient was taken as the random factor and an improvement of Aikaike's information criterion was used as a criterion to keep factors in or out of the model.

For all tests, only a P-value below 0.05 (P<0.05) was considered to be significant. Corrections for simultaneous hypotheses testing were performed according to Bonferroni's coorection method.

Results


Linear mixed model, implant parameters

No significant differences were recorded for RFA mean scores for different implant lengths, either at implant insertion or at loading (P>0.05). However, PTV values at loading for 8 and 10 mm implants were significantly higher (less stable) than those of 12 and 14 mm (P=0.05). Significantly better stability was recorded at implant loading for wider implants (4.1 mm) compared with smaller implants (3.3 mm) for both PTV (P=0.001) and RFA scores (P=0.02). At insertion, however, RFA scores did not differ significantly between both diameters (P>0.05).

Relations between implant stability parameters, intra–inter device comparison

The RFA scores at insertion and loading correlated significantly (r=0.4, P<0.001). When comparing RFA values at loading with PTV scores, a significantly negative correlation was evidenced (r=−0.52, P<0.001). Finally, the RFA scores recorded in the O-F direction were compared with those recorded in the M-D direction. A significant correlation was found at the insertion stage (r=0.6, P<0.001) as well as at the loading stage (r=0.52, P<0.001).

Linear mixed model, bone parameters (Table 2)

Table 2.  Linear mixed model analysis, Bone parameters

The presence of a bony dehiscence or the type of anchorage (single vs. bi-cortical) had no major impact on the RFA or PTV scores in either directions either at insertion or at loading (P>0.05).

Single regression analysis yielded significant correlations between RFA scores at insertion and most HU values. The best correlation was obtained for the spongious part HU scores (r=0.46, P<0.001) (Fig. 3a). Correlations were also obtained for the shell surrounding the implant (r=0.34, P=0.02) and the area corresponding to the coronal 3 mm of the implant (r=0.35, P=0.02). An exception was noted for the HU values of the coronal cortex and the total volume of the osteotomy site, which did not correlate with RFA or PTV scores (P>0.05).

Figure 3.

Correlation between resonance frequency analysis and different bone parameters.

At loading, similar correlations were found between RFA/PTV scores and HU values of the spongious part [for RFA: r=0.42, P=0.001 (Fig 3b) and for PTV: r=−0.47 (P<0.001)]. With RFA, the HU values of the other measurement sites lost their significance, while the HU values of the osteotomy part regained significance. With PTV, significant correlations were detected for HU measurement at all sites, except for the cortical part and the first 3 mm of the osteotomy site. A detailed description of the different correlations can be found in Table 2.

A strong correlation was observed between cortical bone thickness and primary RFA [r=0.57, P<0.001 at insertion (Fig. 3c), and r=0.27, P=0.01 at loading]. For PTV, a negative correlation (r=–0.34, P=0.01) was obtained.

Stepwise multiple regression analyses (Table 3)

Table 3.  Stepwise multiple regression analyses

The stepwise multiple regression analyzed the relative importance of each variable, including their interactions (Table 3). The correlations between predicted RFA/PTV values and clinically scored values are illustrated in Fig. 4.

Figure 4.

Correlation between model-predicted values and clinically obtained values (r2=0.7, 0.57 and 0.79, respectively, and the sample size 74, 83 and 97 measurements, respectively).

Discussion

Several methods have been proposed in the literature for the assessment of jaw bone quality. The Lekholm & Zarb index (Lekholm & Zarb 1985) has been widely accepted as a practical scale for this purpose, mainly due to its ease of use and interpretation by the regular practitioner. However, due to its nature, this index lacks objectivity and is thus practitioner dependent. HU scores that can be measured on CT images have been proven to be an objective and reliable method to assess bone density as they strongly correlate with bone histomorphometry scores (Todisco & Trisi 2005) and BMD (Stoppie et al. 2006). In the latter, strong correlations were found only when the cortex was egg shell thin.

In this study, implant length or diameter did not seem to influence primary stability when considered as single parameters. However, in a stepwise multiple regression analysis, both parameters became significant, probably due to the elimination of the confounding influence of the cortical thickness and/or the impact of bi-cortical anchorage. The implant diameter, however, was shown to affect the RFA scores significantly at loading and confirms therefore the tendency to use wider implants in zones of poor bone quality or poor anchorage to improve success by increasing the possible bone to implant contact (Winkler et al. 2000).

In the present study, a significant correlation was found between RFA scores from F-O and M-D directions. This result contrasts with the findings of Veltri and co-workers (Veltri et al. 2007), who found significant differences in the scores between both directions. Our results would deny the usefulness of measurements in two directions with the Ostell Mentor®. Our study found a correlation between RFA and PTV scores (r=0.52, P<0.001). The correlation was, however, weaker than that reported by Lachmann et al. (2006) (r2=0.9, P<0.0001). This discrepancy could be the result of the use of a different RFA device in addition to the comparison of the two devices at a different stage.

The results from this study demonstrated significant correlations between HU and RFA scores most notably at insertion, and to a lesser extent, at loading. Constant and numerically the highest correlations were obtained when spongious bone density in the osteotomy site was measured without cortical interference. HU evaluation of other areas showed inconstant and weaker correlations. The coronal cortical part did not correlate at all with RFA or PTV scores. These findings are in agreement with previous papers (Rho et al. 1995; Stoppie et al. 2006) which also noted very poor correlations between cortical bone HU scores and bone structural and mechanical properties. Those can also be explained by the smaller areas of measurements for cortical bone which did not cross the threshold of 1 cm2 set by Taguchi et al. (1991) for valid measurements.

The present study also found a very good correlation between primary stability and cortical thickness, which is in agreement with the conclusions of several clinical papers on the preponderance of cortical thickness as a determining factor in implant primary stability (Miyamoto et al. 2005).

The reduction of the importance of bone parameters in the prediction of RFA at loading (two stages approach) confirms the progressive establishment of bone-to-implant contact (secondary stability) in comparison with mechanical anchorage (primary stability).



Primary implant stability seems to be influenced by mainly bone-related factors, namely bone density of the spongious part of the osteotomy site and the cortical plate's thickness. Other factors such as implant length or diameter, even though not influential on implant stability when considered individually, seem to significantly affect stability when considered in a wider multi-variable model. Further studies on larger populations and using more precise implant localization (Surgical guide, Materialise dental, Leuven, Belgium) are required to confirm our results.

Among the different bone- and implant-related factors, the preoperative evaluation of the cortical thickness and HU of the spongious part of the osteotomy site seems to be the most reliable for implant stability prediction. In this regard, they could serve as a guide for the surgeon for the evaluation of potential osteotomy sites and selection of the regions most susceptible to show high primary stability.


The database is protected by copyright ©dentisty.org 2016
send message

    Main page