An evaluation of peri-implant marginal bone loss according to implant type, surgical technique and prosthetic rehabilitation: a retrospective multicentre and cross-sectional cohort study

1 “This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/s10266-020-00587-9” Lizett Castellanos-Cosano1, Alba Carrasco-García2, José-Ramón Corcuera-Flores3, Javier Silvestre- Rangil4, Daniel Torres-Lagares5, Guillermo Machuca-Portillo5. AN EVALUATION OF PERI-IMPLANT MARGINAL BONE LOSS ACCORDING TO IMPLANT TYPE, SURGICAL TECHNIQUE AND PROSTHETIC REHABILITATION: A RETROSPECTIVE MULTICENTER AND CROSS-SECTIONAL COHORT STUDY. 1 Associate Professor. Oral Surgery, School of Dentistry, University of Seville. 2 PhD student. School of Dentistry, University of Seville. 3 Associate Professor. Master Program, School of Dentistry, University of Seville. 4 Associate Professor. Stomatology and Oral Surgery, Dr. Peset University Hospital, Department of Stomatology, University of Valencia, Spain. 5 Full-time Professor, School of Dentistry, University of Seville. Reprint requests and correspondence to: Guillermo Machuca-Portillo, MD, DDS, PhD, Professor and Chairman of Special Care Dentistry, School of Dentistry, University of Seville, C/Avicena s/n, 41009 Seville, Spain, Phone: +34 954 481128, Fax: +34 954 481128, E-mail: [email protected] Authors contributions: All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Lizett Castellanos-Cosano, Alba Carrasco-García, José-Ramón Corcuera- Flores Javier Silvestre-Rangil, Daniel Torres-Lagares and Guillermo Machuca-Portillo. The first draft of the manuscript was written by Lizett Castellanos-Cosano and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Conceptualization: Guillermo Machuca-Portillo; Methodology: Guillermo Machuca-Portillo, Daniel Torres-Lagares; Formal analysis and investigation: Guillermo Machuca-Portillo, Daniel Torres-Lagares, Lizett Castellanos- Cosano, Alba Carrasco-García, José-Ramón Corcuera-Flores, Javier Silvestre-Rangil; Software: Lizett Castellanos-Cosano, Alba Carrasco-García; Writing - original draft preparation: Lizett Castellanos-Cosano; Writing - review and editing: Guillermo Machuca-Portillo, Daniel Torres-Lagares; Resources: Guillermo Machuca-Portillo; Supervision: Guillermo Machuca-Portillo, Daniel Torres-Lagares; Project administration: Guillermo Machuca-Portillo 2 ABSTRACT Objectives: To evaluate implant loss (IL) and marginal bone loss (MBL); follow-up period of up to 10 years after prosthetic loading. Materials and Methods: Retrospective multi-centre cross-sectional cohort study. Double analysis: 1) all the implants (n=456) were analysed; 2) to allow for possible cluster error, one implant per patient (n=143) was selected randomly. Statistical analysis: Spearman's correlation coefficient; Kruskal-Wallis (post-hoc U-Mann-Whitney); Chi-square (post-hoc Haberman). Results: 1) analysing all the implants (456): IL was observed in patients with past periodontitis (6% vs. 2.2%, p<0.05), short implants (12% vs. 2.8%, p<0.001) and when using regenerative surgery (11.3% vs. 2.9%, p<0.001); greater MBL was observed among smokers (0.39±0.52 vs. 0.2±0.29, p<0.01), maxillary implants (0.28±0.37 vs. 0.1±0.17, p<0.0001), anterior region implants (0.32±0.36 vs. 0.21±0.33, p<0.001), external connection implants (0.2±0.29 vs. 0.63±0.59, p<0.0001), and 2-3 years after loading (p<0.0001). 2) analysing the cluster (143): IL was observed in smokers (18.8% vs. 3.5%, p<0.05), splinted fixed crowns (12.9%, p<0.01), short implants (22.2% vs. 4.0%, p<0.01) and when using regenerative surgery (19.2% vs. 3.4%, p<0.01); greater MBL was observed in maxillary implants (0.25±0.35 vs. 0.11±0.18, p<0.05), in the anterior region (p<0.05), greater IL in the first 3 years (p<0.01), in external connection implants (0.72±0.71 vs. 0.19±0.26, p<0.01) and in short implants (0.38±0.31 vs. 0.2±0.32, p<0.05). Conclusions: There is greater risk in smokers, patients with past periodontal disease, external connection implants, the use of short implants and when regenerative techniques are used. Clinical Relevance: In order to prevent MBL and IL, implantologists should be very meticulous in indicating implants in patients affected by these host factors. KEY WORDS: Marginal bone loss, implant design, bone regeneration, prosthetic functional loading, oral rehabilitation design. 3 INTRODUCTION Rehabilitation of tooth loss by means of dental implants is a well-developed and very predictable technique which enables to restore both aesthetics and function in the long term [1]. Radiological marginal bone loss, around dental implants, is a scientifically supported marker for evaluating the long-term success of implant treatment [2]. Radiologically the detection of implant failure is mainly linked to marginal bone loss, hence the importance of early detection [3]. Radiological evaluation of marginal bone loss can be performed in different ways: 1) by evaluating the distance from the implant shoulder to the bone (>1mm); 2) by measuring the number of exposed threads that are not in contact with bone (which can lead to confusion due to the large number of different implant designs in existence); or 3) by evaluating marginal bone loss from implant length [4], but it should always be undertaken using as a reference the bone height around the implants on a radiograph at the time of placing the implants together with several follow-up appointments over time [5]. According to the classic study by Albrektson et al. [6] marginal bone loss of up to 1.5 mm after one year from implant placement and 0.2 mm per year after the first year may be deemed physiological. However, recently it has been noted that marginal bone loss is affected by multiple factors, such as the height of the prosthetic connection from the crown compared to the bone surrounding the implant [7], the loading time [8], the prosthetic connection of implants [9], the fact of having suffered previously from periodontal disease [10], the design of the implant shape [11], social habits such as smoking [12], the insertion of implants at a location subjected to guided bone regeneration techniques [13, 14]; amongst other multiple factors [3]. The aim of this multi-centre study was to evaluate marginal bone loss and failure of implants placed during a follow-up period of up to 10 years after prosthetic loading by assessing the influence of the different variables related to the type of implant (platform -external or internal-, juxta- or supraosseus, conventional length or short, conventional width or narrow), placement site (mandible or maxilla), type of surgery (regenerative or not) and type of prosthetic loading (Kennedy class prior to rehabilitation, type of prosthetic rehabilitation used). MATERIALS AND METHODS A retrospective multi-centre cross-sectional cohort study was carried out, in which the effect of different variables relating to implant treatment on radiological marginal bone loss was evaluated. The population of this study was made up of patients treated with implants in the Master's Course in Oral Surgery at the University of Seville (Spain), the Master’s for Adults and Special Patients at the University of Seville and the Master’s for Adults and Special Patients at the University of Valencia (Spain) between 2009 and 2018. All implants that had had prosthetic rehabilitations with functional loading for more than 2 years were analyzed. To do so, annual check-up orthopantomographs performed were collected from the day the 4 prosthetic rehabilitation was placed on the implant until a period of up to 10 years after functional loading was reached. In this way, a comparative study was carried out between the first orthopantomography taken after placement of the implant’s prosthetic loading and the subsequent orthopantomographs performed from two years after the rehabilitation, and from that moment on, year by year, until the end of each patient's review. Inclusion and exclusion criteria To be included in this study, all the patients’ clinical histories had to be confirmed using a medical report referring to the diseases they had suffered from, as well as the medication taken for the previously identified diseases. It was confirmed that among the patients included in the study no implants had been placed in patients with active periodontal disease and who presented a positive reassessment after at least 6 months since periodontal treatment. Patients being treated with medication which could affect bone metabolism (bisphosphonates, monoclonal antibodies and long-term corticosteroids) or with bone metabolism diseases such as Paget´s disease or osteopetrosis, and patients whose implants had less than two years of prosthetic loading, were excluded. Sociodemographic variables The sociodemographic variables, sex and age of the study subjects were analysed (four categories: up to 40 years, from 41 to 55 years, from 56 to 69 years and from 70 years and above). Data relating to smoking was collected. All patients were given a questionnaire to find out whether or not they had smoked 5 or more cigarettes per week during the last 5 years. Any patient who answered affirmatively was classified as "smoker". If a patient reported having been a user of 5 or more cigarettes per week, but had stopped doing so in the last 5 years, they were classified as "former-smokers". Those who reported never using tobacco were classified as "smokers." Clinical variables Concerning implants, for each one the tooth position it occupied was recorded using the FDI nomenclature (excluding third molar positions). It was also specified whether they were implanted in the upper or lower jaw, and anterior region (central incisors, lateral incisors and canines) or posterior region (premolars and molars) were stipulated. As regards dental history, a record was made of whether the patient had suffered any prior periodontal disease. Prior diagnosis of periodontal disease was determined based on a probing depth of ≥4 mm, bleeding from active pockets and loss of clinical insertion of ≥4 mm [15]. Bruxism was assessed by asking: "Have you been told, or have you noticed yourself, that you grind your teeth or clench your jaw when you are asleep or awake?" (Yes, No). In the case that the subject answered affirmatively, he was qualified "Suffering for bruxism: Yes", and in the opposite case as "Suffering for bruxism: No" [16]. 5 According to the implant design, they were classified into cylindrical implants (the implant body has the same diameter at the cervix as at the apex, parallel walls) or tapered (the diameter of the implant decreases towards the apex, converging walls). With regard to the type of implant depending on its placement in relation to the bone crest, two groups were established: 1) subgingival design without a smooth neck and with a juxta-osseous termination and completely covered by the gum; 2) transepithelial design, supragingival with a smooth neck, enabling tissue healing without being submerged. With regard to the abutment-implant connection type, it was specified whether the connection was internal or external, the distinctive factor between the two types being the presence or absence of a geometric figure which extended over the coronal surface of the implant. Regarding length, implants were placed in two groups: short implants (implant length between 6 and less than 8 mm) and conventional implants, whose length is 8 mm or more [17]. Patients with ultrashort implants (<6 mm) were excluded from this study. Concerning implant diameter, patients were grouped into narrow implants (diameter less than 3.5 mm) and conventional implants (diameter equal to or greater than 3.5 mm) [18]. Patients with mini implants (implants with a diameter less than 2.9 mm) were excluded from this study. The medical records of patients recorded whether the placement of each implant needed bone graft incorporation (prior to or during its placement). Equine-derived granulated biomaterial and an equinederived collagen membrane were used in all cases on patients requiring treatment with a bone graft (Bio- Gen® and Bio-Teck®, BIOTECK S.p.A., Arcugnano VI Italy). The type of prosthesis to which each implant was subjected (unitary crown, multiple implant-supported fixed prosthesis or overdenture). All the abutments were screwed. Cement-retained abutments were not employed. The Kennedy class was also recorded. Class I: bilateral free ended partially edentulous; Class II: unilateral free ended partially edentulous; Class III: unilateral bounded partially edentulous; Class IV: partially edentulous crossing the anterior midline. Once the arch was classified, if it presented an intercalate edentulous space, it was called a subdivision, indicating the number of edentulous spaces. A record was kept of the edentulous space in which each of the implants were placed, as well as whether the space was free ended (if there was no natural dentition at the distal end of the implants inserted), intercalate (there was natural dentition, both mesial and distal, from the implant or implants inserted) or completely edentulous. Check-ups After implant placement, review periods were implemented, and these periods were calculated from the date of prosthetic loading. The periods assessed were: 2 to 3 years after prosthetic functional loading; 3 to 5 years after prosthetic functional loading; 5 to 7 years after prosthetic functional loading; 7 to 10 years after prosthetic functional loading. Marginal bone loss (MBL), Mandibular Cortical Index (MCI) and implant loss 6 MBL, measured in millimetres, is a variable widely used in the literature [19, 20]. The head and mouth position of patients in orthopantomography devices was controlled using a positioner. The MCI is used to evaluate and quantify the quality of mandibular bone mass and observe signs of resorption on panoramic radiographs for the identification of changes in bone density such as osteopenia [21]. All measurements for MCI and MBL were made by a single examiner (ACG) with prior calibration for measuring both indices (MCI and MBL) and obtaining correct specificity. Calibration consisted in the measurement of all implants on all radiographs by a second examiner (JRCF). The assessor calculated a correlation coefficient of 0.925, which implies a coincidence percentage above 90%. To perform the MBL measurements, the National Institute of Health “Image J” program for Microsoft Windows was used for processing digital images previously used by other authors [22], for measuring the distance in millimetres between the bone-implant contact point after placement with the current boneimplant contact point and using the Lagervall & Jansson [4] classification, modified by Corcuera-Flores, et al., [13] to include a fourth category. This classification divides implants into different categories according to MBL around the implant: 0) No MBL; 1) MBL 1/3 of the implant length or less; 2) MBL more than 1/3, but less than 2/3; 3) MBL 2/3 or more than 2/3 of the implant length; 4) Implant loss, for lost implants. Statistical Analysis. All the statistical procedures were undertaken using the SPSS 19.0 program for Windows (IBM, Chicago, IL). A double analysis was carried out: a) On the one hand all the implants placed (456) were analysed. b) To take into account any possible cluster error, one implant per patient was selected (143) in accordance with a random procedure. Frequency and percentage were used for describing the qualitative variables. The correlation coefficient was measured using the Spearman test. The qualitative data comparison was performed using the Chisquare test and applying Haberman’s post-hoc test. The Kolmogorv-Smirnov test was performed, and it was observed that the crossing of variables in relation to marginal bone loss (MBL) followed an abnormal distribution, therefore non-parametric tests were employed to corroborate statistical significance. The Mann-Whitney U test was used for dichotomous variables and the Kruskal-Wallis test for variables with more than two categories, later applying the Mann- Whitney U test to see which groups showed marked differences between each other. To compare the crosses of all variables for implant loss the Chi-square test has been employed, later applying the Haberman test to assess in which groups differences were marked. Statistical significance was established when p <0.05. 7 RESULTS A total of 143 patients and 456 implants were analysed. The variables analysed relating to patients can be seen in Table 1. Analysing all the implants (456), it can be observed that, although there was a greater percentage of women (52.4%) than of men (47.6%), and more implants were placed in the former than in the latter (50.2% vs. 49.8%), there were no significant differences either regarding these data or the percentage of failures (3.5% vs. 4.8%). When analysing all the implants placed, a statistically significant association was noted between implants placed in patients who had been successfully treated earlier for periodontal disease (p<0.05) and the frequency of failed implants. A greater marginal bone loss (MBL) was detected between patients who were smokers and non-smokers or ex-smokers (p<0.01), or in patients who were not taking medications (p<0.05). Analysing one implant per patient (n=143), it can be observed that smokers are more likely to lose implants (p<0.05). Table 2 analyses clinical variables relating to diagnosis and prosthetic rehabilitation. When analysing all implants, a statistically significant difference was discerned between mean annual marginal bone loss and the region where the implants were placed, maxilla (77.2%; MBL=0.28 ± 0.37 mm), mandible (22.8%; MBL=0.10 ± 0.17 mm) (p<0.0001); anterior region (24.8%; MBL=0.32 ± 0.36mm) and posterior region (75.2%; MBL=0.21 ± 0.33 mm) (p<0.001). On examining Kennedy class, the greatest mean annual bone loss was detected in class IV (21.1%; MBL=0.48 ± 0.50 mm) (p< 0.001) and in completely edentulous patients (34.0%; MBL= 0.39 ± 0.46 mm) (p< 0.001). A statistically significant association was noted between type of prosthesis and frequency of failed implants (p< 0.05), as well as with mean annual marginal bone loss (p<0.001). When review periods were studied, the greatest annual mean marginal bone loss was observed in the period between 2-3 years after placement of prosthetic loading (12.7%; MBL=0.49 ± 0.53 mm) (p<0.0001). When analysing implant per patient, greater marginal bone loss was seen in implants placed in the maxilla than in the mandible (p<0.05), as well as in the anterior maxilla region compared to other regions (p<0.05), the difference being even more significant when comparing it with the posterior mandibular region (p<0.01). Furthermore, greater marginal bone loss was detected between 2-3 years after prosthetic loading and 7-10 years (p<0.01), as well as between 3-5 years and 7-10 years (p<0.01). Table 3 shows the variables analysed relating to the implant. When the implants were analysed, a statistically significant association was noted between annual marginal bone loss, implant design (cylindrical 78.9%, MBL=0.2 ± 0.28 m; tapered 21.1%, MBL=0.39 ± 0.40 mm -p< 0.001-) and abutment/implant connection type (internal connection 92.3%, MBL=0.2 ± 0.29 mm; external connection 7.7%, MBL= 0.63 ± 0.59 mm) -p<0.0001-). A statistically significant association was observed between implant length and implant failure (failed short implants 12.5%, failed conventional implants 2.8% - p<0.001-), as well as between regeneration and implant failure (failed implants requiring bone regeneration 11.3%; compared to failed implants in which it was not necessary to perform bone regeneration techniques 2.9% -p<0.001-). After analysing implant per patient, results similar to the ones previously commented on in relation to implant connection were observed with greater marginal bone loss in external connection implants (p<0.01). It was noted that short implants also presented greater marginal bone loss than conventional implants (p<0.05). 8 DISCUSSION Proper evaluation of marginal bone loss for dental implants is needed to assess prognosis and longevity of the treatment carried out. To do so it is essential to resort to complementary diagnostic tests, where panoramic radiography or orthopantomography and periapical radiography are amongst the most outstanding two-dimensional radiological tests [23], and to a lesser extent Cone Beam Computed Tomography (CBCT), despite it offering the possibility of obtaining a three-dimensional evaluation of MBL, also measuring marginal bone loss of the vestibular and lingual region of the implant, besides being able to obtain horizontal values [24]. Notwithstanding, the majority of published studies analysing marginal bone loss employ periapical radiography with a parallelisation technique [25]. In this study orthopantomographs were used to analyse marginal bone loss, similarly to other previous authors [10, 26, 27, 28, 29, 30, 31, 32]. When analysing the variables relating to patients, it was observed that smokers presented higher annual mean marginal bone loss scores than former smokers and non-smokers, coinciding with the results of Mumcu & Beklen [33], where smokers presented greater marginal bone loss irrespective of the type of rehabilitation on implants, this marginal bone loss doubling on occasions according to the study by Stoker et al. [34]. Sayardoust et al. [35] pointed to greater marginal bone loss in smokers with mechanised implants. Various publications have reported that implant failure in patients with type 2 diabetes is a rare event [36]. However, other studies show how patients with this disease had worse peri-implant inflammatory parameters [37]. Other authors found that, in patients with poorer glycemic control, bone factors during healing were negatively modulated, but did not affect implant stability [38]. In the present study, implants were only placed in type 2 diabetic patients with adequate control of the disease, measured by glycosylated hemoglobin. Patients with this condition were grouped among patients with prior relevant systemic pathology. In any case, as can be seen in Table 1, these conditions did not influence either the loss of the implants or their marginal bone loss. In this study, after having analysed all the implants, a significant association was obtained between having suffered prior periodontal disease and implant loss, results coinciding with those previously reported by Corcuera-Flores et al. [13], who found a statistically significant association between marginal bone loss and a history of periodontitis. When studying implant design, it was in cylindrical implants where the lowest scores for marginal bone loss were observed compared to tapered implants, which coincided with the results of Kadkhodazadeh et al. [11], who also obtained lower marginal bone loss one year after placement in cylindrical implants with a highly polished surface. In relation to the type of connection between the implant and the prosthetic rehabilitation, a statistically significant association was detected between annual mean marginal bone loss and the type of connection. Internal connection implants suffered less annual mean marginal bone loss than external connection implants. These results coincide with those observed by Cooper et al. [39] and Pessoa et al. [40]. Cooper et 9 al. [39] found marginal bone loss after 3 years of 0.31 ± 0.41 mm and 0.04 ± 0.51 mm for internal connection implants and external connection implants, respectively. Kim et al. [41] carried out a follow-up after 4 to 12 years from implantation, observing greater mean MBL in external connection implants (0.47 mm) compared to those with internal connection (0.15 mm). However, Esposito et al. [42]) and Esposito et al. [43] highlighted that after a follow-up of one year significant marginal bone loss occurred both in internal connection implants (0.85 mm) and in external connection ones (0.98 mm), without any statistically significant differences existing with regard to connection type. When analysing implant design in terms of the behaviour of marginal bone crest, several studies have shown a certain association between this variable and marginal bone loss. Kütan et al. [44] reported significantly less marginal bone loss in juxta-osseous implants compared to infracrestal ones. Subsequently Lago et al. [45] reported less marginal bone loss in juxta-osseous implants compared to transepithelial implants, the difference between both implant designs being 0.31 mm at one year from the start of the study, 0.53 mm between the first year and after 5 years, and 0.85 mm from the start of the study up to 5 years. No statistically significant differences were obtained in this study when comparing transepithelial and juxtaosseous implants. On studying implant length, greater statistically significant marginal bone loss was observed in short implants compared to conventional implants. However, Hadzik et al. [46] focused their study on the crownimplant ratio, where MBL was similarly low both in short and conventional implants, 0.34 ± 0.24 mm and 0.22 ± 0.46 mm, respectively. Mendoza-Azpur et al. [17] argue that rehabilitation with implants 5.5 to 7 mm in length is equally predictable as with implants that are 10 to 12 mm long. However, this is not confirmed in this study. In the present study, implants placed using bone regeneration presented greater statistically significant MBL than implants placed without bone regeneration. Therefore, this coincides with the results obtained by Koutouzis el al. [47], who observed greater marginal bone loss in the group of patients who had undergone a xenograft and implant compared to the group without regenerated bone, although their results were not statistically significant. Similar to this study, Corcuera-Flores et al. [13] found a statistically significant association between marginal bone loss and the placement of implants in regenerated regions. The majority of scientific studies coincide in the existence of greater marginal bone remodelling during the first year, although numerical scores differ across the different studies. Dimensional stability of tissue is usually studied one year after prosthetic loading of the implant since during the first year. Various authors establish limits for marginal bone loss of 1.5–2 mm [48], establishing a single value of 1.8 mm [49] compared to the 1.5 mm subsequently established [50]. Some authors posit that after the first year a marginal bone loss greater than or equal to 0.2 mm per year is established [6]. In this study, after analysing all the implants, the highest values for marginal bone loss were found in the years closest to prosthetic loading of the implant, the period being from 2 to 3 years from prosthetic loading when the highest levels were described. This was followed by the period from 3 to 5 years, whilst from 7 to 10 years less marginal bone loss was observed. The greatest statistically significant difference was reached between the period from 2 to 3 years when comparing it with the period from 7 to 10 years; and between the period from 3 to 5 years and from 7 to 10 years. 16 TABLES Variables Categories Frequency Placed Implants (%) Frequency Failed Implants (%) MBL per year Mean S.D. Sign. All implants (n=456) Gender Female 229(50,2) 8(3,5) 0,23 0,31 Male 227(49,8) 11(4,8) 0,25 0,37 Age at time of surgery Up to 40 years 15(3,3) 0 0,21 0,23 From 41 to 55 years 131(28,7) 8(6,1) 0,20 0,27 From 56 to 69 years 251(55,0) 10(4,0) 0,23 0,36 70 years or more 59(12,9) 1(1,7) 0,37 0,40 History of periodontal disease Yes 232(50,9) 14(6,0*1) 0,26 0,39 No 224(49,1) 5(2,2*1) 0,21 0,28 Suffering from bruxism Yes 113(24,8) 5(4,4) 0,24 0,42 No 343(75,2) 14(4,1) 0,24 0,31 Relevant previous systemic pathologies Yes 122(26,8) 4(3,3) 0,20 0,27 No 334(73,2) 15(4,5) 0,25 0,36 Medication consumption Yes 143(31,4) 4(2,8) 0,18 0,26 <0,05 No 313(68,6) 15(4,8) 0,26 0,37 Smoker Yes 67(14,7) 3(4,5) 0,39 0,52 <0,01 No 346(75,9) 12(3,5) 0,20 0,29 Former smoker 43(9,4) 4(9,3) 0,29 0,30 Mandibular cortical index (MCI) MCI 1 238(52,2) 10(4,2) 0,24 0,34 MCI 2 193(42,3) 9(4,7) 0,25 0,35 MCI 3 25(5,5) 0 0,15 0,23 One implant per patient (n=143) Gender Female 75(52,4) 3(4,0) 0,20 0,26 Male 68(47,6) 6(8,8) 0,24 0,38 Age at time of surgery Up to 40 years 9(6,3) 0 0,18 0,18 From 41 to 55 years 48(33,6) 6(12,5) 0,22 0,24 From 56 to 69 years 65(45,5) 2(3,1) 0,22 0,37 70 years or more 21(14,7) 1(4,8) 0,24 0,38 History of periodontal disease Yes 66(46,2) 6(9,1) 0,26 0,40 No 77(53,8) 3(3,9) 0,18 0,24 Suffering from bruxism Yes 37(25,9) 2(5,4) 0,23 0,40 No 106(74,1) 7(6,6) 0,21 0,29 Relevant previous systemic pathologies Yes 44(30,8) 2(4,5) 0,24 0,28 No 99(69,2) 7(7,1) 0,21 0,34 Medication consumption Yes 52(36,4) 2(3,8) 0,23 0,28 No 91(63,6) 7(7,7) 0,21 0,35 Smoker Yes 16(11,2) 3(18,8*1) 0,36 0,59 No 115(80,4) 4(3,5*2) 0,20 0,27 Former smoker 12(8,4) 2(16,7) 0,28 0,33 Mandibular cortical index (MCI) MCI 1 76(53,1) 6(7,9) 0,20 0,29 MCI 2 56(39,2) 3(5,4) 0,26 0,37 MCI 3 11(7,7) 0 0,13 0,19 Table 1: Variables analyzed relative to the patient. MBL: marginal bone loss; S.D.: standard deviation. Significance: *1: p<0,05; *2: p<0,01. 17 Variables Categories Frequency Placed implants (%) Frequency Failed implants (%) MBL per year Mean S.D. Sign. All implants (n=456) Position: maxillamandibular Maxilla 352(77,2) 17(4,8) 0,28 0,37 <0,0001 Mandibular 104(22,8) 2(1,9) 0,10 0,17 Position: anteriorposterior Anterior 113(24,8) 3(2,7) 0,32 0,36 <0,001 Posterior 343(75,2) 16(4,7) 0,21 0,33 Position: By arcade/location Maxilla/Anterior 86(18,9) 3(3,5) 0,39 0,38 <0,0001 Maxilla/Posterior 266(58,3) 14(5,3) 0,24 0,36 Mandibular/Anterior 27(5,9) 0 0,13 0,19 Mandibular/Posterior 77(16,9) 2(2,6) 0,09 0,16 Toothless class (Kennedy) Class I 281(61,6) 8(2,8) 0,19 0,26 <0,0001 Class II 16(3,5) 1(6,3) 0,12 0,15 Class III 63(13,8) 3(4,8) 0,11 0,20 Class IV 96(21,1) 7(7,3) 0,48 0,50 Toothless area Toothless free end 142(31,1) 6(4,2) 0,18 0,23 <0,0001 Toothed interlayer section 159(34,9) 5(3,1) 0,15 0,23 Complete edentulism 155(34,0) 8(5,2) 0,39 0,46 Prosthesis type Unitary fixed crown 112(24,6) 2(1,8) 0,14 0,21 <0,001 Splinted fixed crowns 270(59,2) 12(4,4) 0,26 0,37 Overdenture 71(15,6) 2(2,8) 0,32 0,36 Unloaded prosthesis 3(0,7) 3(100,0)*1 --- --- Interval from first review From 2 to 3 years 58(12,7) 3(5,2) 0,49 0,53 <0,0001 From 3 to 5 years 146(32,0) 4(2,7) 0,32 0,35 From 5 to 7 years 166(36,4) 8(4,8) 0,15 0,24 From 7 to 10 years 86(18,9) 4(4,7) 0,09 0,14 One implant per patient (n=143) Position: maxillamandibular Maxilla 109 (76,2) 9 (8,3) 0,25 0,35 <0,05 Mandibular 34 (23,8) 0 0,11 0,18 Position: anteriorposterior Anterior 28 (19,6) 1 (3,6) 0,29 0,35 Posterior 115 (80,4) 8 (7,0) 0,20 0,31 Position: By arcade/location Maxilla/Anterior 16 (11,2) 1 (6,3) 0,42 0,40 <0,05 Maxilla/Posterior 93 (65,0) 8 (8,6) 0,22 0,33 Mandibular/Anterior 12 (8,4) 0 0,11 0,15 Mandibular/Posterior 22 (15,4) 0 0,12 0,20 Toothless class (Kennedy) Class I 96 (67,1) 5(5,2) 0,20 0,24 Class II 4 (2,8) 1(25,0) 0,00 0,00 Class III 27( 18,9) 1(3,7) 0,16 0,21 Class IV 16 (11,2) 2(12,5) 0,53 0,66 Toothless area Toothless free end 45 (31,5) 4(8,9) 0,18 0,22 Toothed interlayer section 64 (44,8) 3(4,7) 0,18 0,25 Complete edentulism 34 (23,8) 2(5,9) 0,34 0,49 Prosthesis type Unitary fixed crown 59 (41,3) 1(1,7) 0,15 0,23 Splinted fixed crowns 62 (43,4) 8(12,9)*2 0,27 0,38 Overdenture 22 (15,4) 0 0,27 0,35 Unloaded prosthesis 0 0 0 0 Interval from first review From 2 to 3 years 21 (14,7) 3(14,3) 0,45 0,55 <0,01 From 3 to 5 years 48 (33,6) 3(6,3) 0,28 0,31 From 5 to 7 years 49 (34,3) 2(4,1) 0,15 0,21 From 7 to 10 years 25 (17,5) 1(4,0) 0,07 0,13 Table 2: Variables analyzed related to prosthetic diagnosis and rehabilitation. MBL: marginal bone loss; S.D: standard deviation; Sign.: Statistical significance: *1: p<0,05; *2: p<0,01. 18 Variables Categories Frequency Placed implants (%) Frequency Failed implants(%) MBL per year Mean S.D. Sign. All implants (n=456) Implant design Cylindrical 360(78,9) 15(4,2) 0,20 0,28 <0,001 Conical 96(21,1) 4(4,2) 0,39 0,49 Implant positioning relative to bone crest At tissue level 356(78,1) 14(3,9) 0,24 0,35 At bone level 100(21,9) 5(5,0) 0,22 0,31 Abutment-implant connection Internal 421(92,3) 19(4,5) 0,20 0,29 <0,0001 External 35(7,7) 0 0,63 0,59 Implant length Short 64(14,0) 8(12,5)*3 0,26 0,26 Conventional 392(86,0) 11(2,8)*3 0,23 0,35 Implant width Narrow 59(12,9) 2(3,4) 0,19 0,25 Conventional 397(87,1) 17(4,3) 0,25 0,35 Regenerative surgery Yes 71(15,6) 8(11,3)*3 0,25 0,28 No 385(84,4) 11(2,9)*3 0,24 0,35 One implant per patient (n=143) Implant design Cylindrical 116 (81,1) 7(6,0) 0,19 0,27 Conical 27(18,9) 2(7,4) 0,33 0,48 Implant positioning relative to bone crest At tissue level 109(76,2) 7(6,4) 0,24 0,35 At bone level 34(23,8) 2(5,9) 0,15 0,20 Abutment-implant connection Internal 136(95,1) 9(6,6) 0,19 0,26 <0,01 External 7(4,9) 0 0,72 0,71 Implant length Short 18(12,6) 4(22,2)*2 0,38 0,31 <0,05 Conventional 125(87,4) 5(4,0)* 2 0,20 0,32 Implant width Narrow 15(10,5) 0 0,19 0,28 Conventional 128(89,5) 9(7,0) 0,22 0,33 Regenerative surgery Yes 26(18,2) 5(19,2)*2 0,23 0,27 No 117(81,8) 4(3,4)*2 0,22 0,33 Table 3: Variables analyzed relative to the implant. MBL: marginal bone loss; S.D: standard deviation; Sign.: Statistical significance: *2: p<0,01; *3: p<0,001.