Retrospective 10 year Analysis of Implant Survival and Complications in a Suburban Australian General Dental Practice
No major retrospective studies assessing general practice implant failures and complications have been published but many have been published for specialist practice and Universities [1-4]. Problems with implants fall into two major groups; 1) Loss or failure; and 2) complications; and these have two basic mechanisms; 1) mechanical; and 2) biological. The major risk factors for implant failure are predominately biological: 1) poor oral hygiene; 2) history of periodontitis; and 3) history of smoking However complications can occur with a far greater range of events and they entail both mechanical and biological events and include: 1) the design of the implant; 2) variations in the abutment implant interface, both design and iatrogenically introduced variation; 3) timing of implant placement; 4) type of implant surgery; and 5) prosthetic design and placement issues. Problems with mechanical placement and design may lead the clinician to compromise the mechanical components which in turn may facilitate biological complications . These biological complications take longer to result in implant loss and studies using data of ≤5 years implant duration are unlikely to show these losses. These biological complications have been reported in patients with aggressive periodontitis in 5 year studies [1,6] but are clearly more evident in studies with a duration of ≥5 years [2,7-8].
The aim of this study was to retrospectively assess the survival and complication data within patients from a suburban general
practice where the clinician’s education was predominately industry-based CPD training activities.
A retrospective cross-sectional analysis of the data of a general dentist in suburban Melbourne, Australia was undertaken between May and June 2015. All patients had signed consent forms to allow their de-identified data to be used for research and sample size calculations were performed for the various assessments undertaken. The study was approved by the ethics committee at the University of Melbourne Dental School (#1443279.1). The clinician performed all aspects of implant placement including case workup, planning, surgery and prosthesis placement and all patients were on maintenance by either a dentist or dental hygienist.
The data collected included: radiographs at and following placement, clinical records and medical histories, the types of implants and their prosthetic unit types, clinical record evidence of ill fit of abutments, single crowns and bridge work, history of smoking, number of cigarettes, number of teeth, number of teeth with periodontal bone loss of >4mm and maintenance. Implant bone loss was calculated using the percentage of the known length of the implant converted to mm.
Survival was defined as; the implant was still present in the oral cavity, whether in function or not. A prosthetic unit causing implant oral hygiene issues (POHI) was defined as a unit with access issues which reduced adequate cleaning and were classified into 4 major categories: 1) embrasures spaces between implants with joined (bridge-like) prostheses; 2) problems where the implant is placed against an adjacent natural tooth and the margins of the implant are subgingival; 3) Ridge lapped prosthetic units; and 4) separate crowns attached to the abutments with screws from the lingual aspect (cross pinned) prosthetic units. These were determined from the clinical notes and radiographs. An implant alignment issue was defined as: 1) two implants used for bridge work were poorly aligned which required altered prosthetic mechanisms to place the required bridge; 2) the implants were not well aligned to the ridge and left without prosthetic units (sleepers); and 3) where the dentist may have tried to realign the implant at surgical placement. The study and implant construction models for most patients were still available and were used to assess the implant alignment and the spacing issues. The crown length: implant length (CL:IL) ratio was calculated for the Bicon™ and Endopore™ implants and the crown length: implant diameter (CL:ID) ratio was calculated for the Endopore™ implants but not the Bicon™ implants due to their structure.
To eliminate any clinician bias the treating clinician was not allowed to see the data until the manuscript was written. Any comments were assessed for potential bias and any changes suggested to the manuscript were only undertaken if they posed no identifiable bias.
The data was assessed for normal distribution and those not normally distributed were either log converted or arcsine converted if percentage data. Statistica for Windows Ver. 12.0 (StatSoft Inc., Tulsa, USA) was used for t-tests and Pearson correlation coefficients and multivariate analysis to assess the parametric data. The non-parametric data was assessed using Spearman rank correlations and Mann-Whitney U-tests. Odds ratios were calculated by the Logit method. Kaplan Meier product limit analysis was performed. Multiplicity correction was carried out on the data based upon the number of variables assessed in each statistical analysis. Missing data was treated as a null point if found.
Demographics and implant data
Two thousand three hundred and fifteen patient records were assessed and 68 patients were found to comply with the requirements. Their demographics and implant details are shown in table 1. The implant type, failure and odds ratio analyses are shown in table 2. Three crowns had excess cement which was removed with one of these implants being lost in less than 4 years. Twenty-three (13.3%) implant prosthetic units became loose. The implants ranged from 8-12 mm in length and 3-5mm in diameter. Sixty-six percent of the implants had cemented crowns and 48.5% of abutments were screw retained. The Bicon™ implants had cemented crowns on impact conical connections.
Effect of Periodontitis
Table 1 shows the demographics of the patients. The patients with periodontitis had only mild to moderate disease. These periodontitis patients lost more early and total implants (Figure 1), increased frequency of peri-implantitis and greater implant associated bone loss compared with the patients with no periodontitis. Table 2 and Figures 1 and 2 show the survival percentages and number of implants with bone loss. The Perio group had no difference in implant loss at 5 years but increased late loss by ten years. Table 3 shows the correlation analysis
Table 1. Patient demographics, and data related to periodontitis, implants and their loss and peri-implantitis. The data is presented as the differences per patient and the implant numbers per group (N(%)).
Table 2. Implant Failure per subject. Loss of implants based upon total, early and late loss. Comparison of implant failure between subjects with and without Periodontitis. The Survival rates are highlighted. Data for implant systems. Correlation analysis and t-tests of subgroups.
of the bone loss, early and late implant loss and the multiple regression analysis, which showed the primary variables associated
with implant failure: peri-implantitis and POHI. Figure 2 and table 2 show the number of implants with bone loss became significantly different by year 2 and continued to increase to 10 years. Early loss was associated principally with misalignment of implants at the time of surgery and having a Bicon™ implant. The late loss of implants was associated with peri-implantitis, POHI and having an external hex Endopore™ implant.
Table 3. Correlation analysis of the associations between the number of sites per patient for Peri-implantitis, Sites with Oral hygiene issues, Sites with Implant Alignment issues and Sites with Implant Abutment issues (entire study population).
Statistical methods: Spearman Rank Correlation Analysis; Multiple regression analysis, Pearson Product Correlation. Odds ratio by Logit method.
Prosthetic issues altering the ability of the patient to perform oral hygiene
Table 3 shows that two major events were related to implant loss: 1) POHI; and 2) implant alignment and abutment/implantinterface issues. Differences between the occurrence of peri-implantitis and late loss of implants and the POHI were assessed within the Perio and NoPerio groups (see Table 2 and Figures 3 and 4). In the NoPerio group there was no association between the POHI and late implant loss but there was a positive association with bone loss. In the Perio group the increasing frequency of POHI correlated with late implant loss, and implants with bone loss.
Figure 3. Comparison of the Association between the Number of Prosthetic units with Oral Hygiene issues and the Number of Implants Lost (Late) in Patients with and without Periodontitis.
Effect of sinus penetration by the implant
The maxillary sinus bone floor was penetrated by 28 implants in 21 individuals, 7 of who had more than one implant penetrating the sinus. Implants that penetrated the sinus were associated with an increase in early loss, bridge prosthetic restorations, more POHI per patient, and implants with bone loss. There was no evidence of sinus clouding or other sinus related issues visible in the radiographs. Some of the sinus penetrating implants showed evidence of increased bone formation when comparing the placement and later reassessment radiographs.
Effects of implant alignment and abutment/implant interface issues
Forty implants in 21 individuals had alignment issues and 11 had more than one implant alignment associated problem.Comparison of these 21 subjects with the remaining 47 study patients (Table 2) revealed that having an implant with analignment issue was associated with an increase in early loss, having a bridge prosthesis restoration, POHI, peri-implantitis and bone loss.
Effect of sinus penetration by the implant diameter
The CL:IL ratios of the Endopore™ implants were compared with the conical connect Bicon™ implants. The Endopore™implants had a shorter CL:IL than the Bicon™ implants (table 3) but higher bone loss rates. The Endopore™ CL:IL correlatedwith bone loss, particularly in those with periodontitis (All r=.38, p<.02, Perio r=.48, p<.007). There was no correlationbetween the CL:IL and bone loss in the Bicon™ group (All r=.25, p=.08, Perio r=.02, p=.93). The CL:ID for the Endopore™implants was also associated with bone loss (All r=.39, p<.01, Perio r=.49, p<.006).
Figure 4. Comparison of the Association between the Number of Prosthetic units with Oral Hygiene issues and the Number of Implants with Bone Loss in Patients with and without Periodontitis.
Effect of Implant surface and placement
In the whole study population the rough surfaced Endopore™ implants had a higher late loss rate than those with smoothsurfaces when the data between five and ten years was used (Smooth 0.20±0.57 v Rough 1.40±1.50, p<0.01). The Bicon™implants had a higher early loss rate compared with the other implants, whilst the Endopore™ implants had a higher late lossrate (Table 2).
The rate of survival of implants in the general dentist practice were 91.5% and 72.1% at 5 and 10 years, respectively, and lower than studies of the same time period for university-based 9 and specialist clinics . In our study implants in subjectswith periodontal disease were no different at 5 years, but by 10 years there was a lower survival rate (61.8% v 82.4%).
Ten year implant survival in periodontitis patients in a specialist clinic showed increased loss in a periodontitis group (Perio 90.7% v NoPerio 93%, at 10 years)  consistent with our study. Importantly this study has revealed that peri-implantitis highly correlated with bone loss, the duration, the measures of periodontal attachment loss and the number of POHI per patient. Therefore it is likely that the combination of implant type and POHI, the high level of cementation and the subsequent lack of retrievability of the prosthetic units may be major contributors to the implant failures.
Regression modelling revealed two major factors that predicted implant loss: 1) peri-implantitis; and 2) POHI. These two factors have been previously identified . In that study 91% of the subjects with peri-implantitis had poor oral hygiene access compared with 55% without peri-implantitis. Poor cleanability occurred in 75% in that study, compared with 43% in this study. Serino had the patients demonstrate oral hygiene to the examiner, whilst in our study it was determined by reports in the notes and obvious anomalies identified in the radiographs suggesting our study may be underreporting POHI. Whilst POHI was not different between our Perio and NoPerio groups, POHI correlated with the number of implants with bone loss in all subjects. These data are consistent with the comparison of well- or ill-fitting crown margins on cemented implant crowns where crestal bone loss was greater in the ill-fitting group compared to the well-fitting group . The results are also consistent with plaque related bone loss in periodontal patients with overhanging/faulty margins . Clearly these studies have identified a critical issue which can be corrected, which in turn may enhance implant survival and reduce implant complications.
Cementation of crowns on implants was originally introduced for aesthetic reasons and to compensate for screw loosening complications on the early external hex implants . The two major disadvantages of cementation are: 1) The lack of retrievability; and 2) the difficulty associated with visualizing and removing excess cement at the crown implant interface. In this study cementation of crowns was the most common method of fixation of the prosthetic units. A study assessing residual cement retention with subjects with peri-implantitis found that 81% of implants had residual cement compared with none in controls without peri-implantitis . Thus the use of cement retention is a significant risk factor for the development of peri-implantitis, however this may be avoided with the use of the newer conical connection implants without cementation.
In this study 16.9% of implants penetrated through the maxillary sinus floor and no sinus lift or grafting procedures were performed. None of the patients revealed any associated sinus pathology, but some had increases in bone around some of the implants within the maxillary sinus. This is consistent with a study of implants, which penetrated the maxillary sinus, which also found no clinical signs of sinusitis, However, they did find sinus mucosal thickening in 60.8% . A previous animal experiment found that sinusitis did not develop when the implant penetrated the sinus . In fact maxillary sinus complications are higher in subjects with sinus grafting procedures than in those without [16-17] and grafting was identified as a risk factor for implant failure in the posterior maxilla. A study which evaluated the procedure of lifting the sinus mucosal floor without graft placement  revealed that all sites healed normally with significant bone formation (mean 6.5mm). The additional bone observed in the sinus floor may be related to this event. These data suggest that penetration of the sinus by the tip of the implant does not appear to have a negative effect on long term implant survival or the development of sinusitis. Further studies are required to assess these interesting findings.
In this study 17.6% had issues with alignment of the implants or an issue with their abutments which was associated with increases in: early loss; bridgework as a restorative method; POHI and peri-implantitis. These alignment issues and their associated factors appear to indicate problems with obtaining good implant alignment at surgical placement, which could have been prevented by the use of surgical guides. The major problems associated with alignment consisted of four basic events: 1) the initial surgical placement and trying to alter the implant alignment; 2) exposure of the implants following 1st stage surgery which may have allowed bacterial colonization of the implant; 3) the use of bridges with prosthetic design changes initiated by the need to adapt to the misaligned implants; and 4) the use of the Bicon™ implant which may have placement stability issues. Two of these issues (1 and 3) may have been reduced with the use of surgical guides, which in turn should eliminate the alignment issue and the clinical need to adjust the prosthetic componentry to fit the implant positions. With the Bicon™ implants the implant is placed by tapping and this may not achieve the highest level of primary stability and secondly the method of cutting the plastic conical implant insertmay result in sharp edges which may penetrate the surgical flaps facilitating bacterial colonization. There are no studies which assess the use of surgical guides and their potential influence upon long term implant survival.
Early implant loss was higher in the periodontitis patients (Table 2) and most were Bicon™ implants. The majority of the implants in this study were placed with mid-crestal surgical incisions which allowed potential colonization of the implant and its cover screw with bacteria in the post-operative period . Bacterial colonization of exposed and covered implants  is common with exposed implants but not found in those with primary surgical closure and complete coverage. The exposed implants were colonized with Prevotella sp., β haemolytic Streptococcus and Fusobacterium sp., and showed greater crestal bone loss than the covered implants. Crestal bone remodeling after the initial placement of an implant is a normal event in all implants and the bone level seems to accommodate to the position of the implants roughened surface/ smooth surface interface whether submerged or not . The increase in early failure noted with misalignment of the implants and the Bicon™ implants suggests implant exposure may have occurred along with bacterial colonization and that open bone/ implant margins may have facilitated the loss. Therefore modification of the surgical incisions used at 1st stage placement of implants resulting in primary surgical closure of the implant sites is likely to reduce early implant failure but studies to demonstrate this are required.
Whilst there were minimal data to assess the outcomes of the individual implant systems, an assessment of implant loss between 5 and 10 years for rough surface and smooth surface implants was made. Those implants with very rough surfaces(Endopore™) were lost more frequently than the smoother surface implants by ten years and this occurred predominatelywithin the periodontitis patients. The Endopore™ system has an external hex attachment method and the crown length: implant length (CL:IL) was found to have a strong positive correlation with bone loss  as in this study. Malchiodi claimed the increased loading of the implant lead to mechanical failure, however we found that the Endopore™ implantsactually had a lower CL:IL than the Bicon™ implants but had greater bone loss. For the Endopore™ implants the CL:IL andCL:ID ratios were significantly correlated with the bone loss but this was not seen in the conical connection Bicon™ implantsfor the CL:IL ratio. These data suggest that the external hex attachment mechanism and the increasing CL:IL and CL:ID ratios were associated with increasing leverage/flexing of the implant/abutment interface facilitated bacterial colonization of the implant/crown interface . These data would strongly support the use of the newer conical implant connections above the older external hex implant systems.
Source of Funding
Conflicts of interest
3.Karoussis IK, Salvi GE, Heitz-Mayfield LJ, Brägger U, Hämmerle CH et al. Long-term implant prognosis in patients with and without a history of chronic periodontitis: a 10-year prospective cohort study of the ITI Dental Implant System. Clin Oral Implants Res. 2003, 14(3): 329-339.
4.Karoussis IK, Brägger U, Salvi GE, Bürgin W, Lang NP. Effect of implant design on survival and success rates of titanium oral implants: a 10-year prospective cohort study of the ITI Dental Implant System. Clin Oral Implants Res. 2004, 15(1): 8-17.
8.Rosenberg ES, Cho SC, Elian N, Jalbout ZN, Froum S et al. A comparison of characteristics of implant failure and survival in periodontally compromised and periodontally healthy patients: a clinical report. Int J Oral Maxillofac Implants. 2004, 19(6): 873-879..
9.Ferrigno N, Laureti M, Fanali S. Dental implants placement in conjunction with osteotome sinus floor elevation: a 12-year life-table analysis from a prospective study on 588 ITI implants. Clin Oral Implants Res. 2006, 17 (2): 194-205.
11.Jansson L, Blomster S, Forsgårdh A, Bergman E, Berglund E et al. Interactory effect between marginal plaque and subgingival proximal restorations on periodontal pocket depth. Swed Dent J. 1997, 21(3): 77-83.
14.Jung JH, Choi BH, Jeong SM, Li J, Lee SH et al. A retrospective study of the effects on sinus complications of exposing dental implants to the maxillary sinus cavity. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 2007, 103(5): 623-625.
15.Jung JH, Choi BH, Zhu SJ, Lee SH, Huh JY et al. The effects of exposing dental implants to the maxillary sinus cavity on sinus complications. Oral Surg Oral Med Oral PatholOral Radiol Endod. 2006, 102: 602-605.
16.Nolan PJ, Freeman K, Kraut RA. Correlation between Schneiderian membrane perforation and sinus lift graft outcome: a retrospective evaluation of 359 augmented sinus. J Oral Maxillofac Surg. 2014, 72(1): 47-52.
19.Thor A, Sennerby L, Hirsch JM, Rasmusson L. Bone formationat the maxillary sinus floor following simultaneous elevation of the mucosal lining and implant installation without graft material: an evaluation of 20 patients treated with 44 Astra Tech implants. J Oral Maxillofac Surg. 2007, 65 (7 Suppl 1): 64-72.
21.Cardaropoli G, Leonhardt AS. Enamel matrix proteins in the treatment of deep intrabony defects. J Periodontol. 2002, 73(5): 501-504Barboza EP, Caúla AL, Carvalho WR. Crestal bone loss around submerged and exposed unloaded dental implants: a radiographic and microbiological descriptive study. Implant Dent. 2002, 11(2): 162-169.
22.Hermann JS, Buser D, Schenk RK, Cochran DL. Crestal bone changes around titanium implants. A histometric evaluation of unloaded non-submerged and submerged implants in the canine mandible. J Periodontol. 2000, 71(9): 1412-1424.
23.Malchiodi L, Cucchi A, Ghensi P, Consonni D, Nocini PF. Influence of crown-implant ratio on implant success rates and crestal bone levels: a 36-month follow-up prospective study. Clin Oral Implants Res. 2014, 25(2): 240-251.
24.do Nascimento C, Miani PK, Pedrazzi V, Gonçalves RB, Ribeiro RF et al. Leakage of saliva through the implant-abutment interface: in vitro evaluation of three different implant connections under unloaded and loaded conditions. Int J Oral Maxillofac Implants. 2012, 27(3): 551-560.