| Sign In to gain access to subscriptions and/or personal tools. |
Stress and Inflammation as a Detrimental Combination for Peri-implant Bone Loss
1 School of Dental Medicine, University of Erlangen-Nuremberg, Glückstr. 11, 91054 Erlangen, Germany; Correspondence: * corresponding author, Siegfried.Heckmann{at}zp.med.uni-erlangen.de
The causes of peri-implant bone loss continue to be controversial. To determine the impact of biomechanical stress and inflammation, we investigated a total of 80 interforaminal implants in situ for more than 10 years. Two stress groups, with 14 patients each, were established: a low-stress situation with single-standing implants, and an increased-stress situation with splinted implants. To categorize inflammation, we introduced a Composite Inflammation Score using 4 inflammatory parameters. Peri-implant bone loss was calculated from digital panoramic radiographs. To differentiate between the effects of stress and inflammation, we compared bone loss in both stress groups at equivalent levels of inflammation. With greater Composite Inflammation Score values, a clear discrepancy between single-standing and splinted implants was evident (p = 0.117/0.000, regression analysis; p = 0.135/0.000, analysis of variance; p = 0.002, t tests). While stress and inflammation alone may not necessarily be detrimental factors, the presence of stress heightens peri-implant bone loss significantly as inflammation increases.
Key Words: peri-implant bone loss • biomechanical stress • inflammation • single-standing implants • splinted implants
The key success criterion in oral implantology is the maintenance of osseointegration, which has been defined as the direct anchorage of implants in the living bone without interposing soft tissue (Brånemark et al., 1983). Uncontrolled ongoing bone loss will thus decrease osseointegration and jeopardize the integrity of the implant. Peri-implantitis and biomechanical stress are the main factors implicated in bone loss around successfully integrated implants (Esposito et al., 1998). Bone loss caused by plaque-induced inflammation has been documented in both animal (Hickey et al., 1991) and human studies (Mombelli et al., 1987). As for biomechanical factors, bone loss has been observed in animal models with dynamic loading (Hoshaw et al., 1994; Isidor, 1997; Miyata et al., 2000; Duyck et al., 2001). Static loading, however, appears to increase bone density (Gotfredsen et al., 2001). In the clinical setting, bone loss has been related to increased stress resulting from the superstructure design (Quirynen et al., 1992), as well as to occlusal factors (Lindquist et al., 1988, 1997). In the presence of peri-implantitis, dynamic loading (Miyata et al., 1998) may, again, be more detrimental than static loading (Gotfredsen et al., 2002). Since clinical data on stress and inflammation are scarce, further insight into the impact of each factor alone, as well as into both factors combined, would be valuable. In the study presented here, stress is categorized by 2 different attachment mechanisms, and inflammation is categorized by means of a newly developed Composite Inflammation Score. It was hypothesized that, at equivalent inflammation levels, bone loss is greater for increased stress levels.
Patients and Implant Characteristics The study group consisted of 28 edentulous patients with interforaminal implants (ITI solid screw; Straumann, Basel, Switzerland) in situ for, on average, over 10 yrs (range, 7.5 to 13.5 yrs). Informed patient consent and approval from the university’s ethical commission were duly obtained (Ref. no. 2780).
Starting from a cohort of 14 patients with 36 single-standing implants with non-rigid telescopic attachments (Heckmann et al., 2004b), a corresponding number of 14 patients with 44 implants splinted over bar attachments was chosen. The criterion for the selection of those for the splinted-implant group was the best possible match with the single-standing-implant group for cofactors of bone loss and inflammation. Almost identical values were obtained in terms of gender, smoking habits, bite force, mean age, and time implants had been in situ. The unique construction design of the non-rigid telescopic attachment and the long in situ time of the implants, coupled with the patients’ elevated ages, account for the sample size. The implants were comprised of one-and two-part types, with the same endosteal design and supra-alveolar implant-abutment connection. Similar healing and loading protocols were applied; thus, implant geometry above bone level was the sole difference (Table 1
Comparison of Stress: Single-standing and Splinted Implants The study began with 2 different biomechanical conditions. In the single-standing group, non-rigid telescopic copings transferred minor stresses to the implants, due to certain construction peculiarities (Heckmann et al., 2001, 2004b). Splinted bar attachments, for example, are expected to induce greater stress, since they cannot be screwed into place without misfit (Jäger and Wirz, 1994; Smedberg et al., 1996; Heckmann et al., 2004a), and counteract the functional deformation of the mandible (Hobkirk and Schwab, 1991).
Classification of Inflammation: Composite Inflammation Score The Modified Plaque Index represents the cause of inflammation, both Sulcus Fluid Flow Rate and Modified Bleeding Index the degree of inflammation, while Keratinized Mucosa is a modulating factor.
To allow for comparability with the radiographic analysis, we referred all measurements taken to the mesial and distal aspects, resulting in 160 sites. The modified Plaque Index and modified Bleeding Index scores were applied as measured. Sulcus Fluid Flow Rate was assessed twice and the mean value calculated. Since Keratinized Mucosa was not to be measured mesially and distally, vestibular and lingual values were averaged. Both Sulcus Fluid Flow Rate and Keratinized Mucosa values were transformed into scores (Table 1
The score values for each inflammatory parameter were divided into 3 classes, representing 3 biologically homogeneous groups (I, II, III). Composite Inflammation Score values were grouped with score values
Radiographic Analysis Each implant was calibrated based on the distance between the uppermost and the lowermost visible thread crests, with 1.25 mm between 2 adjacent crests (Sidexis®, Sirona, Bensheim, Germany). The reference points were the apex and the coronal bone-to-implant contact projected to the midline of the implant. Bone loss was calculated as the length of the corresponding rough surface minus the distance between the reference points. The entire analysis was carried out by one single researcher. Single-standing and splinted implants were evaluated alternately, with 3 measurements averaged for each site.
Statistical Analysis We carried out regression analysis to demonstrate the correlation between bone loss and inflammation. Since data had been gathered from records comprising both bone loss and inflammatory parameters, it was necessary to test in advance whether Composite Inflammation Score, Modified Plaque Index, Sulcus Fluid Flow Rate, Modified Bleeding Index, and Keratinized Mucosa were equally distributed across both stress groups, by the Mann-Whitney-test. As a result, regression analysis was applicable only to the Modified Plaque Index (p = 0.395) and the Modified Bleeding Index (p = 0.870) in the strict sense. The regression lines for Sulcus Fluid Flow Rate (p = 0.000), Keratinized Mucosa (p = 0.000), and Composite Inflammation Score (p = 0.001) were to be extrapolated. Since the regression lines for the factor variable values 0 (single-standing) and 1 (splinted) differed in slope, analysis of covariance was not applicable. Instead, linear regression analysis was used for each stress group and inflammatory parameter separately. Due to the limited applicability of the regression analysis, an additional method, based on the 3 biologically homogeneous groups, was used. The analysis of covariance served to evaluate the general variability of bone loss within both stress groups for each inflammatory parameter. Two-sample t tests were used for pairwise comparisons of the biologically homogeneous groups and the stress groups.
Regression Analysis Unlike the single-standing implants, significant positive associations in the splinted-implant group were observed between bone loss and the inflammatory parameters (Table 2  ): Composite Inflammation Score (slope = 0.565; p = 0.000), Modified Plaque Index (slope = 0.605; p = 0.005), Sulcus Fluid Flow Rate (slope = 1.315; p = 0.000), Modified Bleeding Index (slope = 0.649; p = 0.024), and Keratinized Mucosa (slope = 0.994; p = 0.000).
Analysis of Variance and t Tests The analysis of variance revealed fewer differences in the single-standing than in the splinted groups: Composite Inflammation Score (p = 0.135 vs. 0.000), Modified Plaque Index (p = 0.407 vs. 0.020), Sulcus Fluid Flow Rate (p = 0.015 vs. 0.000), Modified Bleeding Index (p = 0.025 vs. 0.079), and Keratinized Mucosa (p = 0.039 vs. 0.000).
The t tests showed a slight increase in bone loss throughout the biologically homogeneous groups of Composite Inflammation Scores in the single-standing implants, while in the splinted implants, a notable increase was seen. A comparison within the Composite Inflammation Score groups I to III revealed higher bone loss in group I for the single-standing implants and in group II as well as group III for the splinted implants (Fig. 2A
With regard to the 4 underlying parameters, only moderate differences in bone loss were found for the single-standing implants, while a clear increase from groups I to III was found for the splinted implants. This was most evident for Sulcus Fluid Flow Rate and Keratinized Mucosa. When comparing both stress groups within each biologically homogeneous group in terms of Modified Plaque Index, Sulcus Fluid Flow Rate, Modified Bleeding Index, and Keratinized Mucosa, we observed the greatest differences in group III, particularly for Sulcus Fluid Flow Rate and Keratinized Mucosa, and only minor differences in groups I and II (Figs. 2B-2E  ).
This study was designed to differentiate between the impact of inflammation and stress on bone loss in two different implant-stress groups. The only inflammatory parameters to be included were those that clearly reflected inflammation-induced bone loss, not bone loss caused by biomechanical factors. With regard to the Composite Inflammation Score, all statistical tests revealed an increasing discrepancy in bone loss between single-standing and splinted implants as score values increased. One explanation for bone loss may therefore be seen in the different biomechanical conditions, since possible cofactors were controlled. Thus, increased stress in splinted implants might enhance the susceptibility to inflammatory bone loss, or vice versa. If one considers single-standing implants as being at high, and splinted implants at low, Composite Inflammation Score levels, stress and inflammation alone do not necessarily provoke bone loss. Since the Composite Inflammation Score was established on the basis of 4 different inflammatory parameters, its strong point lies in the combination of them. Similarly, the individual parameters revealed a tendency toward greater bone loss in the splinted implants. These parameters underline the fact that higher plaque formation, bleeding, exudation of sulcus fluid, and lack of keratinized mucosa may be associated with more bone loss around splinted implants. Site-specific inflammation is more likely to occur around splinted bar constructions, in addition to the assumed stress situation. The differences in bone loss found, however, cannot be explained by a niche effect, since sites of equivalent inflammation were compared in all cases. The reason for the negative correlation between Keratinized Mucosa and bone loss in the single-standing implants remains unclear. However, the missing positive correlation is concurrent with relevant literature, suggesting that keratinized mucosa might be dispensable around implants which are easily accessible to oral hygiene (Wennström et al., 1994). The causes of peri-implant bone loss have been investigated in various studies, most of which were based on animal models (Hoshaw et al., 1994; Miyata et al., 2000; Duyck et al., 2001; Hürzeler et al., 1998). Regardless of their usefulness for particular scientific questions, animal studies often do not produce long-term data (Gotfredsen et al., 2001) and are not easily applicable to the situation in humans, i.e., occlusal overload is difficult to monitor, or intra-oral force devices may not match the real-life situation (Carr et al., 1996). For this reason, clinical studies would be of considerable value, which, for ethical reasons, excludes invasive methods. Moreover, one has to seek valid assumptions to determine stress. In the few studies on this subject, long cantilevers (Lindquist et al., 1988) and parafunctions (Lindquist et al., 1988; Quirynen et al., 1992) were reported to be associated with bone loss. The multifactorial etiology of poor oral hygiene and extensive loading in combination (Lindquist et al., 1988) is also concurrent with the study presented. But, in a long-term follow-up of the same patients, smoking was underlined as being even more critical. The authors also suggested that signs of parafunction might not truly reflect the loading situation (Lindquist et al., 1997). Taking this as a basis, the approach applied in the study distinguishes between the effects of stress and those of inflammation on bone loss. Stress was categorized by 2 different attachment mechanisms (single-standing, splinted), and the inflammatory situation was standardized based on a Composite Inflammation Score. Clinical and radiographic parameters were recorded in patients whose implants had been in situ for more than 10 yrs. Slight differences in implant geometry were not considered to affect the results. Though the number of implants per patient could not be exactly matched for all single-standing and splinted implants, the basic character of the attachment systems, with their implications for biomechanical stress, remained unchanged. No attempt was made to measure the stress situation in splinted implants over time, since this would not produce information of great value to this study. This approach would have required the removal of the bars, with the risk of refixation (Jäger and Wirz, 1994). Furthermore, the obvious limitations of in vivo tests (Lindquist et al., 1997; Gotfredsen et al., 2001) must be pointed out. Nevertheless, there is considerable evidence of increased stress in splinted implants from in vitro tests (Jäger and Wirz, 1994; Heckmann et al., 2001) and combined in vitro and in vivo studies (Hobkirk and Schwab, 1991; Smedberg et al., 1996; Heckmann et al., 2004a). While one- and two-part implants were used in this study, the microgap between implant and abutment was located above bone level in both groups and should therefore not notably have contributed to the total degree of inflammation (Broggini et al., 2003). Although the reliability of the clinical parameters presented has been controversial (Verhoeven et al., 2000), these parameters nevertheless appear to be appropriate for monitoring the peri-implant soft-tissue conditions (Mombelli and Lang, 1994). The Composite Inflammation Score is therefore intended as a plausible combination of well-established inflammatory parameters. An evaluation of the Composite Inflammation Score on the basis of longitudinal data would be valuable. With the design at hand, however, only cross-sectional ten-year data were available. Due to the impracticality of taking peri-apical radiographs in elderly patients with severely atrophied mandibles, we used digital panoramic radiographs (De Smet et al., 2002) to calculate peri-implant bone loss. The radiographic analysis rests upon the assumption that the implants were placed as recommended, with the coronal border of the rough surface and the adjacent bone at the same level (Weber et al., 1992). Ultimately, the hypothesis of this study was supported in relation to moderate and increased levels of inflammation: In splinted implants, a proportional relationship between inflammation and bone loss appears to exist. In single-standing implants, however, almost no correlation was found. The increasing discrepancy between the two implant groups can therefore be ascribed to the splinted attachment type. Thus, on the basis of the assumed stress situations, neither stress nor inflammation alone would induce notable bone loss, whereas stress coupled with inflammation constitutes a detrimental combination.
The authors thank Dr. J. Röckl (Teningen, Germany) and Dr. R. Schrott (Nuremberg, Germany) for their assistance with the patient enrollment. This project was funded by the School of Dental Medicine, University of Erlangen-Nuremberg, Germany. Received for publication May 19, 2005. Revision received April 5, 2006. Accepted for publication May 11, 2006.
Journal of Dental Research, Vol. 85, No. 8,
711-716 (2006)
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
1 (I), 2–3 (II), and 
4 (III). For the individual parameters Modified Plaque Index, Sulcus Fluid Flow Rate, Modified Bleeding Index, and Keratinized Mucosa, the groups’ score values were 0 (I), 1 (II), and 