Background
Within the last decades, dental implant treatment has become more and more indispensable in modern dentistry to restore missing teeth and preserve the bone from natural atrophy.
To improve osseointegration of dental implants, many different implant configurations have been designed over time. Among the current 1300 different implant systems there is little discussion about the general design of dental implants [
1]. The macro-design describes the broad shape (conical or cylindrical), length, diameter, arrangement and number of threads and numerous other characteristics/configurations of a dental implant. Common implant shapes are cylindrical or tapered [
2]. The micro-design describes the surface characterization of the implant. Design features of the implant–prosthesis–abutment complex, such as the cervical margin, emergence angle and emergence profile, as well as the design of the implant–abutment and abutment–prosthesis junctions and their locations in relation to the oral tissues, can significantly impact peri-implant tissues in the long term [
3]. There is an ongoing discussion in literature whether the broad part of the implant, which is supposed to be covered by bone, should be designed in a more or less rough way in order to directly increase cell accumulation and therefore favor osseointegration [
4]. The roughness of an implant’s surface is achieved either by reductive or additive procedures. The majority of implants hold moderately rough surfaces (Ra = roughness between 1.0 and 2.0/3.0 micron/µm) as these show stronger bone responses than smoother or rougher implant surfaces [
4‐
6]. Unfortunately, the effect of cell accumulation on rough implant surfaces does not only favor bone cell adhesion. Rough implant surfaces increase cell accumulation in general, attracting fibrous tissue and bacteria and therefore favor peri-implant infections of the surrounding soft and hard tissue [
7,
8]. A greater surface roughness comes with increased biofilm formation around the implant [
9,
10]. A review including 22 articles showed that implant roughness is clearly associated with peri-implantitis. The higher the surface roughness, the higher the mean per-implantitis rate. Up to an arithmetic mean surface roughness of Ra 1 µm, a significant reduction of peri-implantitis was found. Higher risk of peri-implantitis appeared for Ra values greater than 1.2 µm [
11]. Other studies revealed that surface roughness below a Ra value of 0.2 microns had no further influence on the quantity and composition of the biofilm formation, bacterial adhesion and colonization [
12,
13]. Regarding this, Meier et al. recommended a surface roughness on implant parts in contact to the oral cavity below that value of 0.2 µm, as this value seems to indicate a threshold for increased biofilm accumulation [
14].
In the collar region of implants, which is exposed to the oral cavity, it is crucial to avoid plaque accumulation. A balance between osseointegration on one hand and the absence of plaque accumulation on the other hand has been found to achieve optimal long-term clinical outcomes. Thus, it is to clarify the resulting question: which implant collar design is best to stabilize the crestal bone level, avoiding peri-implantitis and subsequent risk of implant loss? Until now, there is no consensus in literature favoring one over the other. Concerning the impact of the machined collar neck width, there is a tendency to keep the machined collar short, as implants with shorter machined collar regions appeared to be more effective in decreasing marginal bone loss [
15,
16].
In the present study, BEGO Sedamos RS- and RSX-implants were used to directly compare the effect of machined versus rough-surfaced shoulders/collar regions on the surrounding tissue. Primarily, the peri-implant bone formation was evaluated, the secondary objective was to assess indicators for soft tissue inflammation.
Discussion
This prospective randomized-controlled clinical study was conducted to evaluate the impact of implant shoulder design on marginal bone resorption. No significant difference in the influence of machined or micro-structured implant collar region on peri-implant crestal bone levels or on peri-implant soft tissue health could be found. The difference in marginal bone resorption of 0.04 mm was statistically not significant (p = 0.7; Wilcoxon test).
Comparable results were observed in literature. Bassetti et al. and Karlsson et al. described that implants with machined and implants with rough neck design both fulfilled implant success criteria and that the modification in implant neck texture had no significant influence on peri-implant bone, resulting in the same amount of bone loss [
20,
21]. Other authors, accordingly, could not find a significant association between a special implant design or implant surface on crestal bone changes [
22,
23]. Furthermore, no implants characteristics were found to be related to the presence of peri-implantitis [
24,
25].
The average peri-implant bone resorption in the present study was in accordance with current literature. The RS implants showed a mean bone resorption of 0.61 mm (SD ± 0.28 mm) and the RSX implants 0.57 mm (SD ± 0.24 mm) 2 years after implant insertion. Other studies showed marginal bone loss of 1.07 mm after 10 years [
26] and 0.36 mm (SD; 0.55) after 2 years [
27]. Buser et al. detected radiologically a mean distance from the implant shoulder to the first bone-to-implant contact of 3.32 mm (SD ± 0.73) 10 years after surgery [
28]. Nicolau et al. noticed a bone loss of max. 2.00 mm ± 1.19 mm within 10 years [
29]. After an initial remodeling phase (5 to 6 months) where the depth of implant placement had an influence on the initial bone remodeling, no significant differences were detected [
29]. Several studies report bone resorption up to 1.5/2 mm during the 1st year after insertion as a normal remodeling process and consider a further marginal bone loss of 0–0.2 mm in following years as a physiological progression [
30‐
33].
Recent studies defined the impact of platform switching [
34] and the effect of bone loading [
35] as critical factors in maintaining peri-implant bone levels. Today’s literature on implant shoulder design and the effect on marginal bone remodeling is still very inconsistent. Some authors describe a rather positive influence using roughed neck implants. Koodaryan et al. showed in their systematic review including 12 articles that insertion of implants with rough and rough-surfaced micro-threaded neck design influenced the rate of bone loss and favored lesser marginal bone resorption compared to machined-neck implants [
30].
In contrast, Sánchez-Siles et al. observed in their retrospective study including 1244 implants that the amount of marginal bone loss was significantly lower in smooth-necked implants (1.18 ± 1.39 mm) compared to rough-surfaced implants (2.41 ± 1.35 mm) after 10 years of function (
p < 0.001) [
36]. Valderrama et al. even reported a mean bone gain of 0.11 mm for non-machined collar implants compared to a mean bone loss of 1.00 mm for implants with a machined collar of 2.8 mm [
37]. A review in 2018 found more crestal bone resorption when rough collars were used in comparison to the use of machined [
38].
Comparing implants with differently wide machined collar surfaces also led to unequal conclusions in literature [
25]. Hänggi et al. on one hand observed that bone remodeling did not significantly differ between two types of implants with differently wide machined-neck implants (2.8 mm vs. 1.8 mm) [
39]. While on the other hand a histomorphometric analysis performed by Schwarz et al. obtained lower mean implant shoulder-to-crestal bone contact values for implants with shorter/thinner machined neck; 0.4 mm vs. 1.6 mm [
16]. Similar findings were described by Calvo-Guirado et al. [
15]. The machined collar neck of the RS implant used in the present study, with a width of 0.5 mm, therefore lies in the recommended dimension.
Radiographic evaluation is the most common method in literature to assess bone loss. Unfortunately, conventional radiographs only monitor the mesial or distal aspect of bone loss around the implant body [
40]. There are in general limitations of two-dimensional radiological measurements provided in most studies [
41]. In this study in particular, as panoramic images were used for follow-ups. Periapical radiographs should have been used. Those have proven reliable in determining the bone level changes at different follow-ups [
42].
According to the manufacturer’s instructions and to clinical and scientific proposals the machined-neck implants were inserted positioning the machined neck 0.29 mm (SD ± 0.28) above the crestal bone. Although the non-machined neck implants were proposed to be inserted equi- or subcrestal, they ended up in a slightly epicrestal position with an average value of 0.09 mm (SD ± 0.25 mm). Nevertheless, there was a statistically significant difference in insertion depths as expected (Wilcoxon test:
p < 0.05). These different insertion depths limit results of the study’s scientific significance of the comparison especially in light of the short follow-up period of only 2 years. Schwarz et al. described in his review including 13 publications different outcomes of crestal bone changes after insertion of implants with the same design in different positions relative to the crestal bone [
43]. It remains questionable, if the effect of implants with different designed collars and additionally various insertion depth are comparable. Furthermore, although data were assessed for possible confounders (implant lengths and diameter, insertion site, augmentation, type of overdenture) other not evaluated factors such as angulation, three different surgeons preforming the operations, patient-specific factors, one examiner measuring the bone loss, and the type of examination (measurement methods) might limit conclusions of this study.
Hüzeler et al. described a statistical difference for base-line mean values of crestal bone height and for mean bone level changes after 1 year comparing peri-implant bone levels around implants with and without platform switching. The concept of a platform switching appeared to limit crestal resorption [
32]. Also, other studies showed a relationship between reduced marginal bone level changes and the presence of platform switching [
44]. All implants used in the present study are equipped with platform switching. There is evidence that the characteristics of a platform switch are favorable to keep biological width and therefore reduce marginal bone resorption [
45].
In the presented work, there were no differences in evaluation of peri-implant soft tissue health between machined and non-machined neck implants. One patient developed a peri-implantitis in both inserted implants with probing depth higher than 6 mm and bleeding on probing. This peri-implantitis rate is in correlation with the true “peri-implantitis” frequency of 1–2% suggested by Albrektsson et al. during a follow-up period of 10 years or more [
46].
It is suggested that peri-implantitis progresses in a non-linear, accelerating pattern and that, for the majority of cases, the onset occurs early during follow-up within 3 years of function [
47,
48]. Besides, peri-implantitis is considered to be directly linked to surface roughness [
11]. The surface roughness of the machined collar of implants used in this study was with a value of 0.4 µm within the recommended range. This value also lies within the surface roughness of natural enamel which is reported to be between 0.07–0.5 µm [
19,
49,
50].
However, most studies on the impact of implant collar design on marginal bone levels are performed in animal studies [
41]. There is a lack of evidence from clinical trials reflecting real-life situations, as animal models may not completely recreate the anatomical, physiological, biochemical/functional, or pathological environment of clinical conditions in humans [
51]. According to the present body of literature, there is no implant surface or material, reducing the risk for peri-implantitis [
25]. In terms of biofilm adsorption recent studies showed that roughness does not influence the diversity of the microbiome [
52].
Furthermore, it was observed that several studies in literature did not consider the initial position of the implant neck relative to the crest, resulting in distorted values of mean bone losses [
41]. This fact, besides others, makes it very difficult to compare any study evaluating peri-implant bone changes.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.