RCT comparing minimally with moderately rough implants.Part 2: microbial observations
For figures, tables and references we refer the reader to the original paper.
Microbial adhesion to biomaterials has previously been related to factors such as surface-free energy (Quirynen & Bollen1995; Yoshinari et al. 2000) and chemical composition or physical characteristics of the surface including material surface irregularities and roughness (Nakazato et al.1989; Wu-Yuan et al.1995; Rasperini et al. 1998; Teughels et al. 2006). Current knowledge appears to indicate that low surface-free energy materials, with reduced surface roughness, limit plaque accumulation in vivo, and that the influence of surface roughness on plaque accumulation is more dominant than surface-free energy or electrical charge (Quirynen & Bollen1995; Amoroso et al. 2006; Teughels et al. 2006; Burgers et al. 2010). Subgingivally, surfaces with anRa of 0.8 compared with 0.3 μm harbored 25× more bacteria, with a slightly lower density of cocci (Quirynen et al. 1993). Smoothening the surface below anRa = 0.2 μm showed no further significant changes, either in the total amount or in the pathogenicity of adhering bacteria (Bollen et al.1996; Quirynen et al. 1996a).AnRa value of 0.2 μm was therefore suggested as a threshold surface roughness, below which bacterial adhesion cannot be further reduced. These observations were further confirmed by an in vivo study of Rimondini et al. (1997), who examined initial supragingival plaque formation (first 24 h) on titanium disks by scanning electron microscopy. In contrast to these findings, Wennerberg et al. (2003) observed no clinical difference (plaque growth and development of gingivitis) comparing different titanium abutments with varying degrees of roughness. The latter might be explained by the rude parameter for plaque formation (presence/absence), by the oral hygiene of subjects (low plaque scores in general), and by the inclusion of only fully edentulous patients.
An increased plaque formation/maturation on rough surfaces compared with smooth surfaces can be explained by a facilitated initial adhesion of bacteria to a roughened surface, as well as by the impossibility to clean a rough surface. If applicable to dental implants, rough surfaces might be more prone to plaque formation, and as such to peri-implantitis (Roynesdal et al. 1998; Becker et al. 2000; Wennerberg et al. 2003; De Boever et al. 2009). In a 3-year randomized controlled trial (RCT) in partially edentulous patients (split-mouth design), plasma sprayed implants (= rough) showed a significantly higher incidence of peri-implantitis when compared with implants with a turned surface (= minimally rough; Astrand et al. 2004).
Most current implants have a “moderately rough” surface (AlbrektssonWennerberg2004). The Sa value of these implants ranges between 1 and 2 μm. Clinical comparative studies have clearly indicated higher success rates for osseointegration with moderately roughened surfaces compared with minimally rough surfaces, especially in grafted or compromised bone (for review, see AlbrektssonWennerberg2004; Albrektsson2008). The roughness was increased to enhance bone apposition and to favor bone-to-implant contact (for review, see Shalabi et al. 2006).
Besides the roughness, the periodontal status of the remaining dentition also determines the microbial composition at the peri-implant site (Apse et al. 1989; Quirynen & Listgarten1990; Quirynen et al. 2006). Quirynen & Listgarten (1990) detected more motile rots and spirochetes (phase contrast microscope) around implants placed in partially edentulous jaws. Compared with fully edentulous patients and periodontally non-susceptible individuals, patients with a history of periodontitis had a high peri-implant prevalence of anaerobic putative periodontal pathogens such as Porphyromonasgingivalis, Prevotellaintermedia, and Fusobacteriumnucleatum, 3 and 6 months after either one or two stage implants were fitted in the mouth (Mombelli et al. 1995). The presence of periodontopathogens in the peri-implant sulci will not necessarily result in peri-implantitis or implant failure (Leonhardt et al. 1993, 1999; Papaioannou et al. 1995; Sbordone et al. 1999; De Boever & De Boever2006), although P. gingivalis is often detected at failed implants (De BoeverDe Boever2006) or at implants with progressive bone loss (Sbordone et al. 1999; van Winkelhoff et al. 2000; Sumida et al. 2002; Botero et al. 2005). Also, Aggregatibacteractinomycetemcomitans, P. gingivalis, and P. intermedia have been identified as putative periodontal pathogens (Leonhardt et al. 1992; Ong et al. 1992) in the etiology of peri-implantitis. Recently, Staphylococcus aureus and Pseudomonas aeruginosa were added to the list of putative pathogens (Leonhardt et al. 1999; Renvert et al. 2007).
On the basis of 15 prospective investigations, Karoussis et al. (2007) conducted a systematic review with respect to the survival rate of dental implants in patients with a history of periodontitis. They found no differences in survival rate between patients with and without a history of periodontitis, but significantly greater pocket probing depths, more peri-implant marginal bone loss, and a higher incidence of peri-implantitis in the former group. This is confirmed by two reviews (Quirynen et al. 2007; Heitz-Mayfield 2008), which identified strong evidence for poor oral hygiene, a history of periodontitis and cigarette smoking as indicators for peri-implant disease.
Information on implant survival or success in fully edentulous patients with a history of periodontitis is scarce. Recent reviews indicate that subjects, both partially and fully edentulous, with a history of periodontitis may be at a greater risk for peri-implant infections (Al Zahrani2008; Renvert & Persson2009).
Animal studies seem to suggest that the progression of ligature-induced peri-implantitis, if left untreated, is more pronounced at implants with a moderately rough surface than at implants with a turned surface (Albouy et al. 2008; Albouy et al. 2009). The latter could point to a difference in subgingivalmicrobiota.
The initial colonization of peri-implant pockets has been studied especially on machined (abutment) surfaces in partially edentulous patients (van Winkelhoff et al. 2000; Quirynen et al. 2005; Furst et al. 2007). These studies all show rapid colonization of the pristine pocket even after 20 min (Furst et al. 2007) by periodontopathogens. A recent study assessed the initial colonization in patients with a history of aggressive periodontitis (De Boever et al. 2009) and found that after 10 days, a microbiota was formed compared with teeth in 5/22 patients, and a fairly similar microbiota in 14/22 patients. Surprisingly, the microbial composition remained almost unchanged after 10 days over a 6-month period.
The present study therefore compared the subgingivalmicrobiota around minimally and moderately rough implant and abutment surfaces in both partially and fully edentulous patients susceptible to periodontitis, via a split-mouth study design. Samples were taken at different time points up to 1 year, and analyzed using three different microbiological techniques (qPCR, checkerboard, and culture).
Material and methods
Patients
The descriptive statistics of the patients and subgroups, as well as the surgical protocol, implant characteristics, clinical parameters, and radiographic data are summarized in the accompanying part I (Van Assche et al. forthcoming). Per patient, ≥2 minimally (turned, Tur) and ≥2 moderately rough (TiUnite, TiU) implants (MKIII; Nobel Biocare, Goteborg, Sweden) were randomly alternated (computer randomization program) within the jaws of these patients.
Samples
At each visit (abutment connection, 3 days, 1 and 2 weeks, 3 months and 1 year after abutment connection), samples were taken from the subgingival implant microbiota (two pooled implant sites per surface). If possible, implants in the upper jaw were preferred, and implants with early soft tissue perforation were excluded. After removal of the supragingival plaque using curettes and cotton, the abutments were isolated and dried. Per site, eight paper points were inserted for 20 s and dispersed in 2 ml of reduced transport fluid (for details, see Syed & Loesche1972). In partially edentulous patients, also the subgingivalmicrobiota around teeth in the antagonistic jaw was sampled (before abutment installation and at 1 year). For the latter, the two teeth with the deepest pockets (≥4 mm) were selected (eight paperpoints each, pooled sample). Each sample was homogenized by vortexing for 30 s and processed within 12 h.
Microbiological processing
Culture
The samples were cultured under aerobic (3 days) and anaerobic conditions on selective and non-selective agar plates to quantify the colony forming units (CFU/ml) under aerobic and anaerobic conditions. For details concerning the growth conditions, please refer (Quirynen et al. 1999).
qPCR
Samples for qPCR were frozen at −80°C until the DNA was extracted with InstaGene matrix (Bio-Rad Life Science Research, Hercules, CA, USA) according to the instructions of the manufacturer. Five microliter of the purified DNA was used for the quantification of Tanerella forsythia (Shelburne et al. 2000), P. gingivalis (Boutaga et al. 2003), A. actinomycetemcomitans, and P. intermedia (Boutaga et al. 2005). As a standard for the qPCR, a fragment of the 16S rRNA gene of T. forsythia ATCC 43037, P. gingivalis ATCC 33277, A. actinomycetemcomitans ATCC 43718, and P. intermedia ATCC 25611 was amplified with primers flanking the annealing site of the qPCR primers. This fragment was ligated into the pGEM-T easy vector system (Promega, Madison, WI, USA) and used to transform Escherichia coli DH5α. Plasmids were isolated from the clones using the High Pure Plasmid Isolation Kit (Roche Diagnostics GmbH, Mannheim, Germany). The concentration of the plasmid was determined using the GeneQuant RNA/DNA calculator (Amersham Pharmacia Biotech, Roosendaal, The Netherlands) at a wavelength of 260 nm. A 10-fold dilution series of the plasmid was used in each qPCR run to construct the standard curve. Primers, probes and qPCRmastermix were synthesized using Eurogentec (Seraing, Belgium). qPCR was performed on the ABI 7700 Sequence Detection System platform (Applied Biosystems, Foster City, CA, USA). Data were collected during each annealing phase. In each run, template controls were included. Results were expressed in log10 Genome Equivalents (Geq)/ml or number of bacterial genome/ml.
Checkerboard
Briefly, the samples were lysed and the DNA was placed in lanes on a nylon membrane using a Minislot device (Immunetics, Cambridge, MA, USA). After fixation of the DNA to the membrane, the membrane was placed in a Miniblotter 45 (Immunetics) with the lanes of DNA at 90° to the lanes of the device. Digoxygenin-labeled whole genomic DNA probes to 38 bacterial species were hybridized in individual lanes of the Miniblotter. After hybridization, the DNA probes presented the target DNA using chemifluorescence detection. A computer-linked instrument read the intensity of the fluorescent signals resulting from the probe–target hybridization. Two lanes in each run contained standards at concentrations of 105 and 106 cells of each species. The sensitivity of the assay was adjusted to permit the detection of 104 cells of a given species by adjusting the concentration of each DNA probe. Signals were converted into absolute counts by comparison with the standards on the same membrane. Failure to detect a signal was recorded as zero. All samples of one patient were analyzed on the same checker board, to optimize their reading.
All microbiological evaluations were performed blind. The mean values were calculated on positive values.
Statistics
For culture and qPCR, a linear mixed model was fit with time, edentulism, and implant type as three fixed, crossed factors, and the patient as a random factor. Mean values were calculated based on the positive concentrations, and a threshold of 1.4 log10 was considered for qPCR. A normal QQ-plot was used to assess the normal distribution of the error terms and to confirm the validity of the model's assumptions. Multiple comparisons were set up to compare the implant surfaces per edentulism subgroup, and edentulism groups per implant surface. One-sided P-values were calculated, and corrected according to Sidak's method for multiple hypothesis testing. Checkerboard data were analyzed via a Mann–Whitney test. A statistically significant change, difference, or correlation was considered if P < 0.05.
Estimates of the differences between implant types and their corresponding 95% confidence intervals were calculated based on the results of the linear mixed effects. Differences were considered as non-relevant, when a clinically relevant difference was not included in the confidence interval.
Results
Due to the significant differences between fully edentulous patients (n = 10) and partially (n = 8) edentulous patients, the data will be discussed for each group, separately.
TiU vs. Tur in fully edentulous patients
Culture
For both surfaces, a mean concentration of ±5 log10 CFU/ml was found after 3 days, both for aerobic and for anaerobic species (Fig. 1). This number increased up to ±6 log10 after 2 weeks, and remained stable afterwards.
Figure 1.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in fully edentulous patients analyzed by culture techniques. Ae, aerobe; Anae, anaerobe; Tu, Turned surface; TiU, TiUnite surface; d, days; w, week; m, months; y, year after abutment connection.
In general, the mean number of aerobic and anaerobic CFU/ml was comparable for both implant surfaces (Tur as minimally rough, TiU as moderately rough), for each time point (P > 0.05); with differences ≤0.4 log10). Over the entire period, a mean increase was seen of 0.5 log10 (aer) and 0.3 log10 (anaer) for Tur, and respectively 0.2 log10 (aer) and 0.8 log10 (anaer) for TiU. During the whole follow-up period, the proportion of anaerobic bacteria was systematically higher, for both the TiU and Tur surfaces.
qPCR
Only very few significant differences could be found between both surfaces (TiU vs. Tur) in the amount of key pathogens. In general, the composition and concentration between both surfaces was similar at each time point (P > 0.05). The concentrations also remained relatively stable over time, with an exception of T. forsythia that showed an increase (≥1 log10) after 3 months. A. actinomycetemcomitans was present over the entire time period in a low concentration, with significant difference at 3 days and 2 weeks between both surfaces (Fig. 2).
Figure 2.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in fully edentulous patients analyzed by qPCR. *Significant difference P < 0.05. Aa, Aggregatibacteractinomycetemcomitans; Pg, Porphyromonasgingivalis; Pi, Prevotellaintermedia; Tf, Tanerella forsythia; d, days; w, week(s); m, months; y, year after abutment connection.
The detection frequency of these key pathogens is summarized in Table 1. Most pathogens were often detected (>70%). No major differences could be observed between both implant surfaces (TiU vs. Tur). P. gingivalis was almost always present in all samples of all patients. For the other species, the detection frequency slightly decreased over time, especially for T. forsythia (70–35%).
Checkerboard
Fig.3shows a nearly identical subgingivalmicrobiota for both TiU and Tur implants, for each time point (exception Veillonellaparvulae at 1 year). The changes over time are also very small. Concentrations for A. actinomycetemcomitans, P. gingivalis, P. Intermedia, and T. forsythia at 1 year were, respectively, 0.2, 3.6, 1.2, and 0.6 × 105 counts for Tur, and respectively, 0.2, 0.3, 1.1, and 1 × 105 counts for TiU. The higher concentration of P. gingivalis for Tur was due to one higher concentration; without this value, the mean was 1.8 × 105 counts.
Figure 3.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in fully edentulous patients analyzed by checkerboard. Tur, Turned surface; TiU, TiUnite surface; d, days; w, week(s); m, months; y, year after abutment connection.
TiU vs. Tur in partially edentulous group
Culture
For both surfaces, a mean concentration of ±5 log10 was found after 3 days, both for aerobic and for anaerobic species (Fig. 4). The mean number of anaerobic species progressively increased up to ±6 log10 after 2 weeks, and up to 6.5 log10 after 1 year. For the aerobic species, an increase is seen as well, but after 1 year, their number just reached 6 log10. The mean number of aerobic and anaerobic CFU/ml was comparable for both implant surfaces, for each time interval (P > 0.9); with differences ≤0.3 log10. Over the entire period, a mean increase was seen of 1 log10 (aer) and 1.6 log10 (anaer) for Tur and 0.9 log10 (ae and anaer) for TiU, respectively.
Figure 4.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in partially edentulous patients analyzed by culture techniques. Ae, aerobe; Anae, anaerobe; Tu, Turned surface; TiU, TiUnite surface; d, days; w, week; m, months; y, year after abutment connection.
At 1 year, the mean number of subgingival CFU of the remaining teeth was still 0.6 log10 higher for both aerobic and anaerobic microbiota when compared with Tur and respectively 1 log10 when compared with TiU.
qPCR
No significant differences could be found between both surfaces (TiU vs. Tur) in the amount of key pathogens, either at a certain time point, or over the entire period, except for A. Actinomycetemcomitans at 1 week (Fig. 5), a difference of less than 0.1 log10.
Figure 5.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in partially edentulous patients analyzed by qPCR. *Significant difference P < 0.05. Aa, Aggregatibacteractinomycetemcomitans; Pg, Porphyromonasgingivalis; Pi, Prevotellaintermedia; Tf, Tanerella forsythia; d, days; w, week(s); m, months; y, year after abutment connection.
Over time, however, a consistent increase in the number of key pathogens could be observed, again at a similar rate for both implant surfaces. In comparison with the teeth, both surfaces scored quite similar for P. gingivalis, T. forsythia, slightly lower for P. intermedia and higher (>1 log10) for A. actinomycetemcomitans at 1 year.
The detection frequency is summarized in Table 1. Most pathogens were often detected (±70%). No major differences could be observed between both implant surfaces (TiU vs. Tur). P. gingivalis was almost always present, while A. actinomycetemcomitans was less frequently detected on both surfaces. At 1 year, only 2/8 were positive for the latter bacteria.
Checkerboard
Also, this technique failed to show significant differences in the biofilms of both implant surfaces, either at a specific time point, or over the entire period (Fig. 6). A continuous change over time is obvious for both surfaces, especially for the group of Actinomyces and the orange and red complex.
Figure 6.
Subgingivalmicrobiota along Turned (Tur) and TiUnite (TiU) implants in partially edentulous patients analyzed by checkerboard. Tur, Turned surface; TiU, TiUnite surface; d, days; w, week(s); m, months; y, year after abutment connection.
Concentrations for A. actinomycetemcomitans, P. gingivalis, P. intermedia, and T. forsythia at 1 year were respectively 0.5, 4.7, 6, and 5.2 × 105 counts for Tur and 0.3, 6.3, 11.1, and 9 × 105 counts for TiU.
The impact of the presence of teeth on the overall microbial load (fully [F] vs. partially [P] edentulous subgroup)
Culture
For both subgroups, the number CFU/ml at day 3 were comparable (aerobe as well as anaerobe), but with time, the partial subgroup harbored higher counts (Fig. 1 vs. Fig.4). At 1 year, the partially edentulous group scored 0.6 log10 (Tur) and 0.5 log10 (TiU) higher for aerobes, and 1 log10 (Tur) and 0.2 log10 (TiU) for anaerobes (P > 0.05).
qPCR
Although the concentration of pathogens was quite similar at day 3 for both the full and partial groups (Fig.2 vs. Fig.5), it changed significantly after 1 year, with higher numbers in the partial group, especially for P. gingivalis (>2 log10 difference, P = 0.01) and P. intermedia (nearly 1 log10 difference). In the P-group, A. actinomycetemcomitans was detected less frequently (also confirmed by checkerboard), but at higher concentrations (>1 log10), whereas T. forsythia was detected more frequently at comparable concentrations.