Immediate implant placement in molar regions: risk factors for early failure

Transcription

1 Thomas Urban Lambros Kostopoulos Ann Wenzel Immediate implant placement in molar regions: risk factors for early failure Authors affiliations: Thomas Urban, Department of Oral Radiology and Department of Oral & Maxillofacial Surgery, School of Dentistry, Aarhus University, Aarhus C, Denmark Lambros Kostopoulos, Private Practice, Aabyhoej, Denmark Ann Wenzel, Department of Oral Radiology, School of Dentistry, Aarhus University, Aarhus C, Denmark Corresponding author: Thomas Urban Department of Oral Radiology and Department of Oral & Maxillofacial Surgery School of Dentistry Aarhus University Vennelyst Boulevard DK 8000 Aarhus C Denmark Tel.: þ Fax: þ urban@odont.au.dk Key words: dental implant, failure, molar, risk factors Abstract Objectives: To identify risk factors for early failure of immediately placed implants in molar regions associated with three bone regenerative techniques. Methods: Ninety-two patients (44 women and 48 men; mean age 50 years, 35 smokers and 57 nonsmokers) in need of a single implant crown to replace a molar were included. After placing the implant, patients were randomized to one of three treatment groups for bone reconstruction of remaining periimplant defects: Autologous bone (AB) chips, Ossix membrane (OM) or a combination of AB chips and OM. The implant was submerged, and after 4 months of healing a re-entry surgery was made to connect a healing abutment. Implants with a dehiscence on 2 sites (mesial/distal/oral/ buccal) together with 50% visible threads, were judged as failures. A series of simple logistic regression analyses were performed to identify risk factors for failure among the following independent variables: sex, jaw, smoking status, plaque, bleeding on probing, fistula, extraction reason, mean initial periimplant defect size, treatment group, implant length, buccal bone dehiscence (BBD), soft-tissue dehiscence and infection. The identified risk factors entered a multiple logistic regression analysis. Results: Fifteen implants failed before abutment operation (13 explantations/two nonosseointegrated). Treatment group had no impact on failure. Risk factors for failure were: Smoking 410 cigarettes/day (odds ratio [OR] ¼ 9.29, confidence interval [CI] ¼ ), BBD (OR ¼ 11.43, CI ¼ ) and infection (OR ¼ 36.7, CI ¼ ). Conclusions: Implants placed immediately after extraction of a molar were associated with a high risk for failure at abutment operation. There was no difference in failure rate between three bone reconstructive techniques. Date: Accepted 9 January 2011 To cite this article: Urban T, Kostopoulos L, Wenzel A. Immediate implant placement in molar regions: risk factors for early failure. Clin. Oral Impl. Res. 23, 2012; doi: /j x Immediate implant placement following tooth extraction has attracted attention since the first publication on this topic over 30 years ago (Schulte et al. 1978). Placing implants directly in extraction sockets offers considerable advantages over conventional implant treatment, the most obvious ones being reduced surgical intervention in combination with shorter treatment time. Immediate implant installation can be successful in the short term, which has been documented in several reviews (Chen et al. 2004; Quirynen et al. 2007; Schropp & Isidor 2008; Chen & Buser 2009). However, in a systematic Cochrane review of randomized controlled trials (RCT) on dental implants placed in fresh extraction sockets (Esposito et al. 2006), only two studies (Schropp et al. 2003; Lindeboom et al. 2006) fulfilled the inclusion criteria. The clinical evidence of immediate implant placement following extraction of molar teeth is scarce, and consists mostly of case series (Schwartz-Arad et al. 2000; Artzi et al. 2003; Wagenberg & Froum 2006; Cafiero et al. 2008; Fugazzotto 2008a, 2008b; Matarasso et al. 2009; Siciliano et al. 2009). The long-term prognosis (4 60 months) has been documented in a report of clinical case series (Schwartz-Arad et al. 2000) to be higher in the mandible (5-year cumulative survival rate (CSR) of 92%) than in the maxilla (5-year CSR of 82%). The same study also found a difference in the 5-year CSR between smokers (83%) and non-smokers (90%). In the study, a total of 56 implants had been installed immediately in molar regions (17 in the maxilla/39 in the mandible), and six implants had failed (three in each jaw). The authors had a possible explanation for three of the six failures (one failure due to overloading, two failures due to inadequate bone quality). In a retrospective study, 1925 implants had been immediately installed in all regions of the jaws (Wagenberg & Froum 2006). A total failure rate of 4% (77/1925) was found, 71 of the 77 failing implants were pre-restoration, i.e. early failures. Implants in maxillary and mandibular molar regions had a failure rate of 6.5% (13 out of John Wiley & Sons A/S

2 202 implants) and 5% (13 out of 261 implants), respectively. In nine cases, the reason for failure was infection at the implant site or the adjacent tooth, two implants were removed due to parasthesia, one due to progressive bone loss and 14 were non-osseointegrated. The authors had no explanation for the 14 non-integrated implants. Two recent case series showed more promising results. Of 341 immediately placed implants in the mandibular molar region, one had failed before the abutment operation 3 7 months after implant installation (Fugazzotto 2008a). Of 391 immediately placed implants in the maxillary molar region, two implants failed: one failed 6 weeks post-insertion due to infection, and one was non-integrated at the time of uncovery 6 months after insertion (Fugazzotto 2008b). One early RCT on immediate implant placement in the frontal and pre-molar areas showed that 70% of the periimplant three-walled bone defects healed spontaneously without any use of regenerative techniques. However, periimplant dehiscence-type defects had poor capacity for spontaneous healing (Schropp et al. 2003). Extraction of a molar followed by immediate installation of a cylindrical-shaped implant most often results in large periimplant defects including defects of the dehiscence type, and spontaneous healing cannot be expected. It seems therefore reasonable to assume that this procedure should be combined with a bone reconstructive procedure. No study has been conducted to evaluate various guided bone regeneration (GBR) methods in combination with immediate implant placement in the molar regions of the jaws. Furthermore, no study has evaluated the interrelation between risk factors for failure in immediate implant installation in molar regions. The aim of the present study was therefore to identify risk factors for early failure of immediately placed implants in molar regions associated with three bone regenerative techniques. Material and methods Patient selection Ninety-two patients (44 women and 48 men; mean age 50 years [range of years]) referred to Aarhus Dental School, University of Aarhus, Denmark, for extraction of a lower or upper molar and subsequent implant treatment at the involved site, were consecutively included in this study. The project was approved by the regional Committee of Ethics, registered in a public clinical trials register (ClinicalTrials.gov ID: NCT ), and conducted in accordance with the World Medical Association Declaration of Helsinki II. Each patient was given oral and written information regarding the study, and their written informed consent was obtained. Patients had the opportunity to withdraw from the study at any time. They received the implant free of charge. The first patient in the study was seen for a pre-examination early October 2005, and patients were having the implant placed and reentry surgeries performed in the period from late October 2005 until December Pre-operative examination In addition to the anamnestic information, a clinical examination was performed, and intraoral and panoramic radiographs were taken to perform a pre-operative evaluation of the implant site (Table 2, Table 4). Person-related inclusion criteria were (1) 418 years of age, (2) classified as ASA class 1 or 2 according to the American Society of Anesthesiologists six-category physical status classification system, used to assess patients before surgery. The first two classes are (1) a normal healthy patient and (2) a patient with mild systemic disease (e.g. mild hypertension). Tooth-related inclusion criteria were (1) a molar ripe for extraction due to advanced caries lesion, periodontally compromised tooth, root fracture, periapical pathology or combinations of these (Fig. 1, Table 1) and (2) adequate bone available at the implant site for placing a 10-mm-long implant without violating the mandibular canal or the sinus floor assessed on the panoramic radiograph (magnification factor 1.3). Exclusion criteria were: systemic diseases affecting bone turnover and pregnant or lactating women. Patient operation All patients were treated by one operator, a trained, specialized oral surgeon (T. U.) following a standardized protocol. Pain was controlled with Naproxen (500 mg twice a day for 5 days) and antibiotics were administered systemically (e.g. phenoxymethylpenicillin 1.6 g twice a day for 5 days). Before surgery, local analgesic, 7.2 ml of Xyloplyin Dental Adrenalin (Dentsply, Addlestone, UK) was administered. Post-operatively, the patients rinsed with an antiseptic mouthwash (0.2% chlorhexidine digluconate) for 14 days. After sulcular and carefully placed vertical releasing incisions mesially and distally to the tooth, a full-thickness muco-periostal flap was elevated buccally. Careful tooth extraction was then performed trying to preserve the alveolar bone, followed by meticulous debridement of the socket. After extraction of the upper molars, in no case an identifiable initial sinus perforation was recorded assuming that the membrane was intact before implant placement. A Brånemark System, Groovy, Wide Platform implant (Nobel Biocare, Göteborg, Sweden) with an endosseous diameter of 5 mm was inserted. The implant was placed in level with the surrounding bone when the cover screw was mounted. The osteotomy was performed with drills and osteotomes splitting the remaining septum in order to maximize implant primary stability. All placed implants were clinically stable. In seven cases, a small sinus membrane perforation had occurred, which was verified with the obligatory Valsalva maneuver before implant placement in the upper jaw. In two cases, thetimeofimplantsurgerywasdelayedby14 days due to purulent discharge from the removed tooth. Per-operative examination Before mounting the cover screw, intra-surgical measurements were performed. At four sites (i.e. buccal, distal, oral and mesial), the residual periimplant defects were measured using a manual periodontal probe graduated in 1 mm steps with the measurement rounded to the nearest mm and expressed as follows: Parallel width of the defect (PaW), the horizontal width from the mesial/buccal wall of the defect to the distal/oral wall of the defect. Perpendicular width of the defect (PpW), the horizontal width from the surface of the implant to the socket wall at the level of the alveolar crest. Vertical depth of the defect (VD), the vertical height from the implant shoulder and to the most apical extension of the bony defect. The presence or absence of a buccal dehiscence was recorded. By means of a closed randomization procedure, the subjects were assigned to one of three treatment regimes of the periimplant bone defects: Autologous bone chips (AB), Ossix membrane (OM) (Biomet 3i, Palm Beach Gardens, FL, USA) or a combination of AB chips and OM (ABOM) (Fig. 1, Table 2). The randomization was performed by an assistant, who drew a lot from a non-see-through bag. To eliminate the chance of unfavourable splits between the three treatment modalities with regard to upper or lower molars, the randomization was divided in two: One bag of lotsfortheupperjawandoneforthelowerjaw. Patients in group AB had AB chips harvested using a Safescraper (Meta, Reggio Emilia, Italy). For a detailed description of the procedure, see Urban & Wenzel (2010). The collected bone was meticulously packed into the defect surrounding the implant. In group OM, a cross-linked col John Wiley & Sons A/S 221 Clin. Oral Impl. Res. 23, 2012 /

3 Fig. 1. Patient flow diagram. Table 1. Main reasons for molar tooth extraction (n ¼ 92) Reasons for molar extraction (%) n Caries Apical periodontitis Root fracture Marginal periodontitis Table 2. Patients in the three treatment groups in relation to sex, jaw, smoking and age Sex Molar Smoking Age W M U L Yes No Mean SD Autologous bone (AB) Ossix membrane (OM) ABOM combination Total W, women; M, men; U, upper jaw; L, lower jaw; SD, standard deviation. lagen membrane of bovine origin (Ossix, Biomet 3i) was adapted covering the implant and the circumferential defects. In the combined group ABOM, the periimplant defects were packed with harvested AB and afterwards covered with a collagen membrane as described above. Group AB included 30 patients; the two other groups OM and ABOM had 31 patients (Table 2). Thus, a total of 92 Brånemark System Mk III Groovy WP (Nobel Biocare) were randomized (45 in the maxilla and 47 in the mandible). Of the 45 maxillary implants, 38 were replacing a first molar, and seven were replacing a second molar. Of the 47 mandibular implants, 41 were replacing a first molar, and six were replacing a second molar. The length of the implants varied considerably according to site and jaw (Table 3). To achieve primary closure of the wound, the periosteum was incised at the base of the muco-periostal flap, which was subsequently coronally positioned by means of horizontal mattress sutures. For suturing, a 5 0 polyglactin 910 suture coated with triclosan (Vicryl Plus s )was used. All patients, who during the time before surgical re-entry presented with a soft-tissue dehiscence exposing part of the cover screw, were instructed in meticulous hygiene management at the particular site including application of 1% w/w Chlorhexine Digluconate Gel. If a succeeding infection evolved at the exposed site, it was initially treated by antibiotics before a surgical re-entry was conducted. Table 3. Implant length distribution in the jaws (mm) mm Upper jaw (n) Lower jaw (n) Total Surgical re-entry/-ies At 4 months following implant placement, a surgical re-entry procedure was performed. A crestal incision was made with two vertical releasing incisions mesially and distally, and a mucoperiostal flap was elevated. The dimensions of the marginal bone defects at the mesial, buccal, distal and oral sites of each implant were measured as described previously. 222 Clin. Oral Impl. Res. 23, 2012 / John Wiley & Sons A/S

4 Table 4. Combined pre- and per-operative variables for the treatment site and the adjacent teeth in relation to the three treatment groups Smoking (cigarettes/day) Plaque BOP Fistula Implant length (mm) BBD STD Infection Yes No Yes No Yes No Yes No Yes No Yes No Autologous bone (AB) Ossix membrane (OM) ABOM combination Total BOP, bleeding on probing; BBD, buccal bone dehiscence; STD, soft-tissue dehiscence. The cover screw was removed, a healing abutment was mounted, and the muco-periosteal flap was adapted and sutured around the abutment with 5 0 polyglactin 910 sutures (VicrylPlus s ). In case of minor, incompletely healed periimplant defects or actual dehiscences, three different approaches were taken according to the size of the defect. (1) If VD on any of the periimplant defects or dehiscences at the four sites was 1 mm (o1.5 mm), a healing abutment was mounted on the implant without any regenerative procedure. (2) If VD on any of the periimplant defects or dehiscences at the four sites was 2 3 mm ( mm), AB chips were grafted to the defect before suturing around the mounted healing abutment. (3) If VD on any of the periimplant defects or dehiscences at the four sites was 4mm, a new cover screw was mounted and AB chips were grafted to the defect, and the implant was covered with a collagen membrane (Ossix). After another 4 months, a second re-entry was performed at the sites of the treated defects, and the above criteria were applied again. In all cases, where grafting was performed the visible implant threads were thoroughly cleaned according to the method described by Schou et al. (2004) before the regenerative procedure occurred. If the dehiscence or the intra-bony periimplant defect was assessed to be too large to be treated by GBR with an expected acceptable outcome, the implant was explanted equalling early failure. The criteria determining an early implant failure were: (1) a non-osseointegrated implant, (2) an osseointegrated implant with extensively compromised periimplant supporting bone in terms of: (a) (b) presence of a dehiscence with a minimum PaW of 43 mm on at least two of the four sites: buccal, distal, mesial or oral and visible implant threads 50 % of the total implant length on one of the four sites buccal, distal, mesial or oral. Statistical analyses Sample size was calculated to include 26 patients in each randomization group. To allow for patient drop-out, a 20% surplus of patients was added. The sample size was calculated in a power analysis to be able to find a statistically significant difference in bone defect reduction of 1 mm or more between the groups. Bone defect reduction and not implant loss was the variable used forsamplesizecalculation. The analysis of the data was performed using SPSS v (SPSS Inc., Chicago, IL, USA). All data were first analyzed by descriptive methods. Differences between the three treatment groups AB, OM and ABOM regarding sex and smoking were tested by the w 2 test. Difference in age between the three treatment groups was tested by analysis of variance. Furthermore, the w 2 test was used to test differences in sex, soft-tissue dehiscences and infection between smokers and non-smokers as well as differences in soft-tissue dehiscences and infection between implants having and not having a buccal bone dehiscence (BBD). Also, the w 2 test was used to test differences between number of failures at sites with and without an initial sinus perforation on the implants placed in the upper jaw. If the minimum expected counts in the contingency table in the w 2 test was below five, or below 10 when there is only one degree of freedom, Fischer s exact test was chosen. The outcome was implant failure. Both nonosseointegrated implants and explanted implants were defined as failures. In order to identify risk factors for implant failure before or at the surgical re-entry, a multiple logistic regression model was applied. Initial analyses Implant failure was the dependent variable in a series of logistic regression analyses to identify a minimal set of independent risk factors. The following independent variables were entered separately as categorical variables: sex, age (divided into two age groups at the mean value o50 and 50 years), jaw, smoking status, plaque on adjacent teeth and the tooth to be extracted, bleeding on probing (BOP) on adjacent teeth and the tooth to be extracted, fistula in relation to the extracted tooth, reasons for extraction of the molar, mean initial periimplant defect size (entered into the model for PaW, VD and PpW not only as continuous-scale independent variables but also divided into two groups by the median value), treatment group, implant length, BBD, soft-tissue dehiscence and infection. Plaque and BOP were entered in the model as yes/no. When dividing the initial mean PpW into two groups according to the median of the group, all the cases with buccal dehiscences entered the model with the assigned value of 4, which was the maximum defect size buccally in the non-dehiscence cases. Therefore, the distribution of the two groups is different in PpW, PaW and VD. The level of statistical significance was set at Po0.1 in the initial analyses. Intermediate analysis A second logistic regression model included all the independent variables that were statistically significant in the initial analyses. Smoking status entered the model in the ramified version with four categories. Final analysis In the final logistic regression model, a recoding of the variable smoking status was performed because initial and intermediate analyses did not yield a statistically significant impact of smoking 10 cigarettes/day. Non-smokers and light smokers ( 10 cigarettes/day) were therefore grouped and so were medium ( cigarettes/day) and heavy ( cigarettes/ day) smokers. After dichotomization of the variables PaW, VD and PpW, only VD was significant and entered the final analysis. No further detailed analyses of interactions and residuals in the regression model were performed. The level of statistical significance was set at Po0.05. Results The patient population at baseline and its detailed characteristics are presented in Tables 2 6. No statistically significant differences were observed with respect to sex (P ¼ 0.87) or age (P ¼ 0.3) between the three treatment groups. In the AB group, in one case replacing a second upper molar extracted due to extensive perio John Wiley & Sons A/S 223 Clin. Oral Impl. Res. 23, 2012 /

5 dontitis, the implant was displaced into the maxillary sinus after an attempt to correct the line of insertion and regain primary stability. This particular case did not enrol in the randomization, and therefore group AB included 30 patients. In 41 of the 92 randomized patients, a softtissue dehiscence developed before the surgical re-entry leaving part of the cover screw exposed to the oral cavity. In 13 of the 41 patients having a soft-tissue dehiscence, a post-operative infection evolved at the implant site. The number of smokers differed between the three groups because more smokers had been included in groups AB and ABOM than in group OM. The difference between groups OM and AB was statistically significant (P ¼ 0.037). For softtissue dehiscences, a significant difference was seen between group AB and ABOM (P ¼ 0.029), but not between OM and ABOM (P ¼ 0.075). For infection, no significant difference was observed between the three groups (P ¼ 0.067). No significant difference was shown in smoking status between men and women (P ¼ 0.911). A significant proportion of the patients who developed a soft-tissue dehiscence (P ¼ 0.006) and a succeeding infection (P ¼ 0.004) were smokers (Table 5). The number of infections (n ¼ 8) was the same in the group with BBD and the group without a BBD. The difference in the percentage of implants with infection in the two groups was highly significant (Po0.001). Also, a significantly larger number of smokers (P ¼ 0.045) was present in the BBDgroup,butthenumberofsoft-tissuedehiscences was not significantly larger in the BBD group (P ¼ 0.051). Implant fate at first re-entry During the period from implant placement to the surgical re-entry, one implant in the maxilla exfoliated, and this belonged to the group AB; thus, a total of 91 implants underwent a surgical re-entry, and 66 implants received a healing abutment at first re-entry. Twenty-six of the 66 implants had bone chips grafted to the persisting defects before suturing around the mounted healing abutment. Twelve implants fulfilled the criteria for explantation at first re-entry. All, but one of the 12 implants were partly osseointegrated and had to be removed surgically. Thirteen implants received a GBR procedure using AB chips and an OM due to defects/dehiscences of large size. Table 5. Smoking status in relation to buccal bone dehiscence, soft-tissue dehiscence and infection Buccal bone dehiscence Soft-tissue dehiscence Infection No Yes Total No Yes Total No Yes Total Smoker No Yes Total Table 6. Mean initial defect size standard deviation (SD) of intrasurgical periimplant defects assessed at four sites per implant (n ¼ 92) Mesial Buccal Distal Oral Mean PaW (mm SD) VD (mm SD) PpW (mm SD) PaW, parallel width; VD, vertical depth; PpW, perpendicular width. Implant fate at second re-entry On 13 implants, a second surgical re-entry was made after application of a GBR procedure. Two partly osseointegrated implants had to be removed at this stage due to large remaining defects fulfilling the criteria for explantation. The remaining 11 implants received healing abutments. Seven of these had bone chips grafted to the persisting defects before suturing around the mounted healing abutment. These defects had no site with a VD of 42 3 mm. Thus, a total of 15 implants were lost/removed; nine in men and six in women, seven in the maxilla and eight in the mandible. The majority of the failed implants (11 out of 15) had been placed at sites, where the molar tooth was extracted due to apical periodontitis. In 10 of the 15 failed implants, an infection at the implant site had been present. Eight of the 10 failed implants with prior infection were in smokers. 31.4% of the smokers lost their implant compared with 7% in the non-smoker group. Two of the seven implants with a recordable sinus perforation failed, which was not a significantly higher rate compared with failures that had had no sinus perforation (P40.36). These two implants replaced a molar extracted due to apical periodontitis, respectively marginal periodontitis. The initial logistic regression analyses (Table 7) showed that the variables smoking status, PaW (only the non-dichotomized variable), VD, bone dehiscence and infection at the implant site had a statistically significant impact on implant failure. The other variables tested had no significant impact. To assess the interrelationship between smoking, VD, bone dehiscences and infection, these variables were entered into the final regression model. The risk factors for implant failure in the final model were: Smoking (odds ratio [OR] ¼ 9.29, confidence interval [CI] ¼ ), VD (OR ¼ 36.94, CI ¼ ), BBD (OR ¼ 11.43, CI ¼ ) and infection (OR 36.7, CI ¼ ) (Table 7). Discussion Extracting a molar and immediately placing an implant results in periimplant defects of such a size that spontaneously healing cannot be expected (Schropp et al. 2003; Chen et al. 2004). On the existence of this evidence our study was designed without a negative control group (i.e. no GBR procedure) in order to comply with the demands of the Danish Committee of Research Ethics. Three GBR procedures were used in this study: AB, OM and the concomitant use of AB and an OM. AB has for a long time been considered the gold standard grafting material for bone augmentation and was therefore chosen (Palmer & Palmer 1999; Esposito et al. 2009). A collagen membrane was chosen because bio-resorbable membranes have been successfully used for boneaugmentationinanimalaswellasinhuman studies (Kostopoulos & Karring 1994; Chiapasco & Zaniboni 2009). Previous research has shown that it is advantageous to use cross-linked collagen membranes because mucosal healing is less disturbed when compared with the use of non-resorbable membranes, even in case of premature exposures of the membrane (Moses et al. 2005; Tal et al. 2008). The combined use of AB and OMs was used in order to test whether this treatment modality would result in improved periimplant bone/soft tissue healing by improving membrane support and by reducing the risk for displacement of the bone particles during healing. A recent systematic review (Atieh et al. 2010) achieved one RCT on survival of implants placed into fresh extraction sockets partly involving molar regions (Bianchi & Sanfilippo 2004). However, in this particular study randomization was limited to the use or non-use of connective tissue graft to achieve wound closure, and the focus was 224 Clin. Oral Impl. Res. 23, 2012 / John Wiley & Sons A/S

6 Table 7. Odds ratio (OR), 95% confidence intervals (CI) and P-values (P) for various risk factors for implant loss. Variables Category (n) OR (initial), CI (P) OR (final), CI (P) Sex Women (44) 1 Men (48) 1.46, (0.51) Age 50 (48) 1 o50 (44) 2.53, (0.12) Jaw Maxilla (45) 1 Mandible (47) 1.11, (0.85) Smoking No (57)/recoded in final OR (67) 1 1 Yes (35)/recoded in final OR (25) 6.07, (0) 9.29, (0.03) 40 to 10 cigarettes/day (10) to 20 cigarettes/day (15) 4.82, (0.04) 420 to 30 cigarettes/day (10) 30.92, (0) Plaque No (70) 1 Yes (22) 2.54, (0.12) Bleeding on probing No (26) 1 Yes (66) 2.94, (0.18) Fistula No (83) 1 Yes (9) 1.54, (0.62) Extraction reasons Caries (11) 1 Apical periodontitis (54) 1.15, (0.87) Root fracture (5) 0, 0 (0.99) Marginal periodontitis (22) 0.45, (0.46) Initial mean defect size PaW 1.47, (0.01) o4.5 (45) (47) 2.16, (0.19) VD 1.57, (0) o4.5 (45) (47) 18.67, (0) 36.94, (0.048) PpW 0.63, (0.55) o2 (39) 1 2 (53) 2.29, (0.19) Treatment group Autologous bone (AB) (30) 1 Ossix membrane (OM) (31) 0.7, (0.66) ABOM combination (31) 2.26, (0.23) Implant length 10 mm (25) mm (25) 1.57, (0.64) 13 mm (35) 3.41, (0.14) 15 mm (7) 4.6, (0.17) Buccal bone dehiscence No (76) 1 1 Yes (16) 23.67, (0) 11.43, (0.03) Soft-tissue dehiscence No (51) 1 Yes (41) 0, 0 (0.99) Infection No (79) 1 1 Yes (13) 49.33, (0) 36.7, (0) not on molar regions (42 molar implants out of 116 placed implants). Thus, it may be assumed that no study has been conducted to evaluate various GBR methods in combination with immediate implant placement in molar regions. Themainfocusinthepresentstudywason risk factors for early failure and their interrelation. Failure of an implant was defined as either a non-osseointegrated implant (n ¼ 2) or a partly osseointegrated implant with extensively compromised periimplant supporting bone (n ¼ 13). Fulfilling these criteria, the implant was removed or explanted. Using the criteria, a total of 15 implants failed at the abutment operation: four in group AB, three in group OM and eight in group ABOM. The same surgeon who inserted the implants also decided at re-entry which implants were failures. It was not possible to maintain a total blinding towards the treatment group because the surgeon was often able to remember which patients belonged to which group. The criteria for failure were well defined, so we believe that this procedure did not cause recall bias, which is supported by the fact that there were no significant differences in failure rate between the groups. The explantation criteria may be regarded as rather strict, and therefore the short-term survival rate, 83% for the lower molar region and 84.4% for the upper molar region, may seem low in comparison with other studies on immediate implant placement in molar regions. When comparing success rates between studies, it is therefore important to be aware of the differences between success/failure criteria that exist between studies. Some studies have implemented the criteria for implant survival defined by Albrektsson et al. (1986), focusing mainly on the osseointegration and function of the implant. Other studies, however, have focused not only on whether the implant was osseointegrated and in function at a given time point but also on periimplant bone and soft tissue conditions (Cafiero et al. 2008; Matarasso et al. 2009; Siciliano et al. 2009), which may predict its longer-term prognosis. It may be judged unacceptable to compromise the patients oral health in terms of risking early onset periimplantitis, involving patient discomfort and increased costs in terms of controlling periimplant pathogens by means of periimplant supportive therapy and surgery. In a previous systematic review (Hämmerle et al. 2002) from the Fourth European Workshop on Periodontology in 2002, it was noted that the criteria for success/failure were not uniform in the reviewed papers, and it may be appropriate to discuss the previously stated criteria by Albrektsson et al. (1986). The logistic regression analysis showed that the main factor determining implant failure was smoking. The risk of early failure was nearly 10 times higher (OR ¼ 9.3) if the patient smoked 410 cigarettes/day than if the patient smoked 2011 John Wiley & Sons A/S 225 Clin. Oral Impl. Res. 23, 2012 /

7 lessornotatall.smokingwasnotarandomization item in the present study. Thus, the higher number of smokers allocated to group ABOM could explain the higher number of failures in this group. That smoking may be linked to lower implant survival in conventional implant placement has been established in recent studies including meta-analyses (Bain et al. 2002; Hinode et al. 2006; Strietzel et al. 2007). The observation of early implant failure associated with smoking at abutment connection has also been reported in a previous study (Debruyn & Collaert 1994). However, no quantity graduation was made because smoking was a dichotomous variable. A recent study graduated smoking habits into number of smoked cigarettes/day and found an increased risk of early implant failure in smokers, increasing with the number of smoked cigarettes (Alsaadi et al. 2007). The former study found an OR for failure of 1.87 for smoking cigarettes and 2.72 for smoking 420 cigarettes. None of these studies were based on immediately placed implants, and in comparison, the OR in our study was much higher. On the other hand, a retrospective study on immediately placed implants in all sites of the jaws found no significant difference in implant failure between smokers (410 cigarettes/day) and non-smokers (Wagenberg & Froum 2006). Thus, the findings in the literature may still be controversial. Although smoking 10 cigarettes/day was not associated with a short-term increased risk of implant failure in our study, the long-term outcome of smoking even a low number of cigarettes after immediate implant placement awaits documentation. Also, the presence of a BBD was associated with implant failure. This observation was also reported in a recent study, which found that complete osseointegration did not occur in molar site implants with buccal self-contained dehiscences (Siciliano et al. 2009). However, in that study the implants were placed transmucosally and not submerged, and the periimplant defects were treated with deproteinized bovine bone particles in conjunction with a resorbable collagen membrane. In our study, 16 implants had a BBD with no difference between the three regenerative groups. Soft-tissue dehiscences were often present together with BBDs. The majority of the BBDs were seen in smokers, and the deleterious effect of smoking on hard as well as soft tissue has been described in implant dentistry (Debruyn & Collaert 1994; Quirynen et al. 2002; Strietzel et al. 2007). Moreover, infection constituted a risk factor for implant failure. To get an infection in the newly augmented periimplant bone, soft-tissue dehiscence had to be present. Soft-tissue dehiscence (n ¼ 41) did not always lead to clinically identifiable infection followed by failure of the implant, but in the majority of infections (nine out of the 13) this was the case. A significant proportion of smokers developed soft-tissue dehiscence, and 10 out the 13 infections were recorded in smokers. Even though both BBDs and infection had a significant impact on failure in the final regression analysis, the CIs were huge, limiting the credence of these variables. It would be meaningful to the clinician if preimplant risk factors could be identified; i.e. just after tooth extraction, it would be possible to estimate the fate of the implant if installed immediately.thesizeofthebonedefectsremaining after implant placement might be estimated even before the implant is inserted; however, only the mean VD of the defect was associated with implant failure. In the present study, implants with a mean VD 4.5 mm had a higher risk for failure. The CIs were, however, very wide, and therefore the confidence in this variable may also be low, and the result should be interpreted with caution. Immediate implant placement may imply besides comfort also savings in time and money for the patient. The attempt for time saving should, however, be balanced against a possible higher risk for implant failure or a higher risk for compromised periimplant bone and soft tissue status. The longterm fate of the surviving implants in molar regions in the present study is yet to be documented. Conclusion Immediate placement of implants in molar regions was associated with a relatively high risk of failures at abutment operation regardless of the type of bone regeneration method used. No differences in failure rate were seen among three periimplant bone reconstructive techniques. Smoking 410 cigarettes/day was the overall dominant risk factor for implant failure, while the presence of a buccal dehiscence and/or infection further increased the risk. Acknowledgements: The authors wish to thank G. Egebaek, B. Jensen and R. S. Pedersen, for assisting in the surgeries. We also wish to thank E. Gotfredsen for valuable help with SPSS. The study was supported by a grant from the Danish Research Agency, Ministry of Science, Technology and Innovation, # References Albrektsson, T., Zarb, G., Worthington, P. & Eriksson, A.R. (1986) The long-term efficacy of currently used dental implants: a review and proposed criteria of success. The International Journal of Oral & Maxillofacial Implants 1: Alsaadi, G., Quirynen, M., Komarek, A. & van Steenberghe, D. (2007) Impact of local and systemic factors on the incidence of oral implant failures, up to abutment connection. Journal of Clinical Periodontology 34: Artzi, Z., Parson, A. & Nemcovsky, C.E. (2003) Widediameter implant placement and internal sinus membrane elevation in the immediate postextraction phase: clinical and radiographic observations in 12 consecutive molar sites. The International Journal of Oral & Maxillofacial Implants 18: Atieh, M.A., Payne, A.G., Duncan, W.J., de Silva, R.K. & Cullinan, M.P. (2010) Immediate placement or immediate restoration/loading of single implants for molar tooth replacement: a systematic review and meta-analysis. The International Journal of Oral & Maxillofacial Implants 25: Bain, C.A., Weng, D., Meltzer, A., Kohles, S.S. & Stach, R.M. (2002) A meta-analysis evaluating the risk for implant failure in patients who smoke. Compendium of Continuing Education in Dentistry 23: , 702, 704 passim; quiz 708. Bianchi, A.E. & Sanfilippo, F. (2004) Singletooth replacement by immediate implant and connective tissue graft: a 1-9-year clinical evaluation. Clinical Oral Implants Research 15: Cafiero, C., Annibali, S., Gherlone, E., Grassi, F.R., Gualini, F., Magliano, A., Romeo, E., Tonelli, P., Lang, N.P. & Salvi, G.E., Italia I.T.I.Study Group. (2008) Immediate transmucosal implant placement in molar extraction sites: a 12-month prospective multicenter cohort study. Clinical Oral Implants Research 19: Chen, S.T. & Buser, D. (2009) Clinical and esthetic outcomes of implants placed in postextraction sites. The International Journal of Oral & Maxillofacial Implants 24 (Suppl.): Chen, S.T., Wilson, T.G. Jr & Hammerle, C.H. (2004) Immediate or early placement of implants following tooth extraction: review of biologic basis, clinical procedures, and outcomes. The International Journal of Oral & Maxillofacial Implants 19 (Suppl.): Chiapasco, M. & Zaniboni, M. (2009) Clinical outcomes of GBR procedures to correct peri-implant dehiscences and fenestrations: a systematic review. Clinical Oral Implants Research 20 (Suppl. 4): Debruyn, H. & Collaert, B. (1994) The effect of smoking on early implant failure. Clinical Oral Implants Research 5: Esposito, M., Grusovin, M.G., Felice, P., Karatzopoulos, G., Worthington, H.V. & Coulthard, P. (2009) Interventions for replacing missing teeth: horizontal 226 Clin. Oral Impl. Res. 23, 2012 / John Wiley & Sons A/S

8 and vertical bone augmentation techniques for dental implant treatment. Cochrane Database of Systematic Reviews CD Esposito, M.A., Koukoulopoulou, A., Coulthard, P. & Worthington, H.V. (2006) Interventions for replacing missing teeth: dental implants in fresh extraction sockets (immediate, immediate-delayed and delayed implants). Cochrane Database of Systematic Reviews CD Fugazzotto, P.A. (2008a) Implant placement at the time of mandibular molar extraction: description of technique and preliminary results of 341 cases. Journal of Periodontology 79: Fugazzotto, P.A. (2008b) Implant placement at the time of maxillary molar extraction: treatment protocols and report of results. Journal of Periodontology 79: Hämmerle, C.H., Jung, R.E. & Feloutzis, A. (2002) A systematic review of the survival of implants in bone sites augmented with barrier membranes (guided bone regeneration) in partially edentulous patients. Journal of Clinical Periodontology 29 (Suppl. 3): ; discussion Hinode, D., Tanabe, S., Yokoyama, M., Fujisawa, K., Yamauchi, E. & Miyamoto, Y. (2006) Influence of smoking on osseointegrated implant failure: A metaanalysis. Clinical Oral Implants Research 17: Kostopoulos, L. & Karring, T. (1994) Augmentation of the rat mandible using guided tissue regeneration. Clinical Oral Implants Research 5: Lindeboom, J.A., Tjiook, Y. & Kroon, F.H. (2006) Immediate placement of implants in periapical infected sites: a prospective randomized study in 50 patients. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology & Endodontics 101: Matarasso, S., Salvi, G.E., Iorio Siciliano, V., Cafiero, C., Blasi, A. & Lang, N.P. (2009) Dimensional ridge alterations following immediate implant placement in molar extraction sites: a six-month prospective cohort study with surgical re-entry. Clinical Oral Implants Research 20: Moses, O., Pitaru, S., Artzi, Z. & Nemcovsky, C.E. (2005) Healing of dehiscence-type defects in implants placed together with different barrier membranes: a comparative clinical study. Clinical Oral Implants Research 16: Palmer, P. & Palmer, R. (1999) Dental implants. 8. Implant surgery to overcome anatomical difficulties. British Dental Journal 187: Quirynen, M., De Soete, M. & van Steenberghe, D. (2002) Infectious risks for oral implants: a review of the literature. Clinical Oral Implants Research 13: Quirynen, M., Van Assche, N., Botticelli, D. & Berglundh, T. (2007) How does the timing of implant placement to extraction affect outcome? The International Journal of Oral & Maxillofacial Implants 22 (Suppl.): Schou, S., Berglundh, T. & Lang, N.P (2004) Surgical treatment of peri-implantitis. The International Journal of Oral & Maxillofacial Implants 19(Suppl.): Schropp, L. & Isidor, F. (2008) Timing of implant placement relative to tooth extraction. Journal of Oral Rehabilitation 35 (Suppl. 1): Schropp, L., Kostopoulos, L. & Wenzel, A. (2003) Bone healing following immediate versus delayed placement of titanium implants into extraction sockets: a prospective clinical study. The International Journal of Oral & Maxillofacial Implants 18: Schulte, W., Kleineikenscheidt, H., Lindner, K. & Schareyka, R. (1978) The Tübingen immediate implant in clinical studies. Deutsche Zahnarztliche Zeitschrift 33: Schwartz-Arad, D., Grossman, Y. & Chaushu, G. (2000) The clinical effectiveness of implants placed immediately into fresh extraction sites of molar teeth. Journal of Periodontology 71: Siciliano, V.I., Salvi, G.E., Matarasso, S., Cafiero, C., Blasi, A. & Lang, N.P. (2009) Soft tissues healing at immediate transmucosal implants placed into molar extraction sites with buccal self-contained dehiscences. A 12-month controlled clinical trial. Clinical Oral Implants Research 20: Strietzel, F.P., Reichart, P.A., Kale, A., Kulkarni, M., Wegner, B. & Kuchler, I. (2007) Smoking interferes with the prognosis of dental implant treatment: a systematic review and meta-analysis. Journal of Clinical Periodontology 34: Tal, H., Kozlovsky, A., Artzi, Z., Nemcovsky, C.E. & Moses, O. (2008) Long-term bio-degradation of crosslinked and non-cross-linked collagen barriers in human guided bone regeneration. Clinical Oral Implants Research 19: Urban, T. & Wenzel, A. (2010) Discomfort experienced after immediate implant placement associated with three different regenerative techniques. Clinical Oral Implants Research 21: Wagenberg, B. & Froum, S.J. (2006) A retrospective study of 1,925 consecutively placed immediate implants from 1988 to International Journal of Oral & Maxillofacial Implants 21: Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Supporting information in accordance with the CONSORT 2010 checklist of information to include when reporting a randomised trial. Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article John Wiley & Sons A/S 227 Clin. Oral Impl. Res. 23, 2012 /