Analysis of technical complications and risk factors for failure of combined tooth-implant-supported fixed dental prostheses

Background: The oral rehabilitation with fixed restorations supported by the combination of teeth and dental implants has been advocated in some cases. Purpose: To assess the clinical outcomes of these prostheses. Fixed restorations supported by the combination of teeth and dental implants. Materials and methods: This retrospective study included all patients treated with combined tooth-implant-supported fixed dental prostheses (FDPs) at one specialist clinic. Abutment/prosthesis failure and technical complications were the outcomes analyzed. Results: A total of 85 patients with 96 prostheses were included, with a mean follow-up of 10.5 years. Twenty prostheses failed. The estimated cumulative survival rate was 90.7%, 84.8%, 69.9%, and 66.2% at 5, 10, 15, and 20 years, respectively. The failure of tooth and/or implant abutments in key positions affected the survival of the prostheses. There were seven reasons for prostheses failure, with the loss of abutments exerting a significant influence. Bruxism was possibly associated with failures. Prostheses with cantilevers did not show a statistically significant higher failure rate. No group had a general higher prevalence of technical complications in comparison to the other groups. Conclusions: Although combined tooth-implant-supported FDPs are an alternative treatment option, this study has found that across 20 years of service nearly 35% the prostheses may fail.


| INTRODUCTION
The oral rehabilitation with fixed restorations supported by the combination of teeth and dental implants has been advocated when anatomical restrictions are present together with financial constraints limiting the number of implants, 1,2 especially for patients where there are few remaining teeth or their distribution in the jaw is too unfavorable to permit fixed bridges including long pontic spans and/or cantilever segments. 1Moreover, this type of connection would also be suggested when there is alveolar bone deficiency restricting strategic positioning of implants and refusal of the patient to agree to augmentation procedures. 1,3However, the differences in mobility between natural teeth and implants have raised concerns about the potential risk for biomechanical complications, 1 that is, that an osseointegrated implant may be too rigidly fixed to share the functional loads with connected teeth. 4Therefore, it should be of interest to investigate the outcome of this prosthetic option in routine practice.6][7][8] In this context, this retrospective study aimed to assess clinical outcomes (estimated success and survival; observed complication and failure types) of fixed dental prostheses (FDPs) with combined implant and tooth support (three-unit, multiunit, and full-arch), in relation to implant-, site-, patient-, and prosthetic-risk factors.

| Materials
This retrospective study included patients treated with dental implants during the period 1980 to 2018 at one specialist clinic (Clinic for Prosthodontics, Centre of Dental Specialist Care, Malmö, Sweden).
This study was based on data collection from patients' dental records, and the patients were not recalled for further examination.The implants were placed by specialist dentists in oral surgery, and dentists performing the prosthetic treatment were specialists in prosthodontics.The study was approved by the regional Ethical Committee, Lund, Sweden (Dnr 2014/598; Dnr 2015/72).
The implants were nonrigidly connected to the teeth.The prosthesis was screwed to the implant(s) and cemented to the tooth (teeth), and the latter previously received a cemented inner cap/coping, also known as telescoping double crowns (Figure 1).Neither sliding type attachments nor vertical or horizontal locking screws were used.

| Definitions
An implant/a tooth was considered a failure if presenting signs and symptoms that led to implant/tooth removal.A prosthesis was considered as failed in the following cases: (a) loss of one of more teeth and/or implants supporting the prosthesis, and (b) replacement of the prosthesis by a new combined tooth-implant-supported FDPs or another type of prosthesis due to the occurrence of technical complications.
Patients smoking a minimum of one cigarette per day were classified as smokers.Former smokers were the patients who stopped smoking before implant placement, and so remained until the patient's last follow-up.
For this study, the authors followed the definition of bruxism proposed by Lobbezoo et al 9 : "Bruxism is a repetitive jaw-muscle activity characterized by clenching or grinding of the teeth and/or by bracing or thrusting of the mandible.Bruxism has two distinct circadian manifestations: it can occur during sleep (indicated as sleep bruxism) or during wakefulness (indicated as awake bruxism)."The sign and symptoms of bruxism were listed according to the International Classification of Sleep Disorders, 10 following the same guidelines used in a recent study. 11The patients suspected to be bruxers, as diagnosed in the records, were recalled in this previous recent study in order to be clinically reassessed, and were considered as presenting bruxism based on self-report of clenching/grinding during sleep or during wakefulness, plus the inspection part of a clinical examination.Part of these patients-the ones with the type of prosthesis being evaluated here-is included in the present observational study.As electromyography and/or polysomnography were not used, the patients of the present study were classified as presenting the "probable" diagnosis of bruxism, according to the diagnostic grading system of "possible," "probable," and "definite" sleep or awake bruxism suggested for clinical and research purposes. 12According to an international consensus 9 following the recommendations of this diagnostic grading system, One prosthetic unit was considered as one missing tooth or one natural-tooth crown replaced by a prosthesis, in any way or form: a prosthetic crown on an implant (I), on a tooth (T), as a pontic (P), or as a cantilever (C).
The prosthetic structures were classified into three groups, according to the following criteria: • group I: two or three prosthetic units.In the case of 3 units, this could be formed by several configurations: 1I-2T, 2I-1T, 1I-1P-1T, 1I-1T-1C, or 1T-1I-1C; • group II: prosthesis with a minimum of 4 prosthetic units, in any configuration possible, regardless of the number of implants, teeth, The use of telescoping double crowns in a combined tooth-implant-supported fixed dental prosthesis pontics or the presence of cantilever.However, the prosthesis could not be classified as a full-arch prosthesis; • group III: full-arch prosthesis, in any configuration possible.
Complications could have happened to the prosthesis, or more specifically to the implant or abutment tooth: • Complications regarding the prosthesis: loss/fracture of suprastructural porcelain or acrylic teeth; fracture of the prosthesis superstructure (ie, prosthesis with complete transversal buccallingual fracture); looseness of the prosthesis.
• Complications regarding the implant: failure; fracture of the implant itself; loosening, loss or fracture of connecting/ abutment screws; loosening, excessive wear, deformation or fracture of prosthetic abutment; pain.
• Complications regarding the abutment tooth: caries infiltration; post/core fracture or loosening; extraction due to root fracture, loss of periodontal support, or due to extensive caries; endodontic treatment; cement dissolution; intrusion; pain.

| Inclusion and exclusion criteria
Only fixed partial or full-arch restorations supported by the combination of teeth and dental implants were included.Modern threaded cylindrical-or conical-design implants were included, and zygomatic implants were excluded.

| Data collection
The data were directly entered into a SPSS file (SPSS software, version 25, SPSS Inc., Chicago, Illinois) as the dental records of the patients were being read, and it consisted of the following factors: After completion of the prosthetic treatment, the patients were periodically followed up by a dental hygienist at the clinic for a professional dental hygiene program, with attendance frequency to the professional hygiene recall program based on individual needs.

| Statistical analyses
The mean, SD, and percentage were calculated for several variables.
The tests performed were the following: Kolmogorov-Smirnov (to evaluate the normal distribution), Levene's test (to evaluate homoscedasticity), Student's t test or Mann-Whitney (for two independent groups, continuous variables), Kruskal-Wallis test or one-way analysis of variance (for three of more independent groups, continuous variables).The estimated cumulative survival was calculated from lifetable analysis.Comparisons of abutment tooth, implant, and prosthesis survival between some factors were done using the log-rank (Mantel-Cox) test.Univariate Cox regression models were used to evaluate the associations between clinical covariates and prosthesis failure.Hazard ratio (HR) and 95% confidence intervals (95% CI) were estimated from Cox proportional hazard models.In order to verify multicollinearity, a correlation matrix was to be performed in case two or more predictor variables would result in a significant HR (P value cut-off point of .1) in the univariate models, to see whether there were some high correlations among the predictors.In that case, collinearity statistics obtaining variance inflation factor and tolerance statistic were also to be performed to detect more subtle forms of multicollinearity.A final multivariable Cox regression model was to be performed only in case of more than one variable moderately associated (P < .10)with prosthesis failure were identified in the univariate model.Clustering of multiple prostheses within each patient was accounted for using the methods outlined by Lee et al 13 and Lin. 14ta were statistically analyzed using the SPSS version 25 software (SPSS Inc.).The Cox models were performed using SAS software v9.4 (SAS Institute Inc., Cary, North Carolina).The degree of statistical significance was considered P < .05.This observational study followed the STROBE guidelines.

| RESULTS
A number of 86 patients were identified as being rehabilitated with fixed restorations supported by the combination of teeth and dental Example of short-span combined tooth-implantsupported fixed dental prosthesis of the study implants.One patient was not followed up in the clinic after the installation of the prosthesis.Thus, the study included 85 patients (41 men, 44 women) aged a mean ± SD of 61.6 ± 12.2 years (min-max, 23.0-83.3)at the time of prosthesis installation.The patients were followed-up for a mean ± SD of 125.9 ± 78.6 months (min-max, 7.1-315.2).
From a total of 96 combined tooth-implant-supported FDPs (49 in men, 47 in women, 64 in maxillae, 32 in mandibles), 76 patients had one prosthesis, seven patients had two prostheses, and two patients had three prostheses.The prostheses consisted of 2 to 13 prosthetic units (mean ± SD of 5.8 ± 3.1 per prosthesis), of which 32 prostheses pertained to group I, 36 to group II, and 28 to group III.
Examples of short-and long-span prosthesis included in the study are shown in Figures 2 and 3, respectively.The material used was metalacrylic 25 for prostheses and metal-ceramic for 71 prostheses.All prostheses, with the exception of one opposed a full removable denture, had either natural teeth or tooth-or implant-supported fixed prostheses in the opposite jaw.There were a total of 560 prosthetic units, of which 164 teeth, 274 implants, 79 pontics, and 43 cantilevers.
Eight of the 96 prostheses were previously planned as implantsupported prostheses, but as some implants (n = 12; not included among the 274 initial implants) failed before the abutment connection, the prostheses were replanned to combined tooth-implantsupported FDPs, as the option was presented to the patients and/or these refused to be submitted to further implant installation surgery.
Table 1 shows a comparison between the different groups concerning the mean number of prosthetic units, the prevalence of bruxers, smokers, cantilever, sex and prosthesis material distribution, and follow-up time.There was no statistically significant difference in the mean age of the patients between the three groups (P = .826,Kruskal-Wallis test).A total of 18 abutment teeth were extracted, due to extensive caries (n = 5), loss of periodontal support (n = 9), and fracture (n = 4).Information about bruxism habits was available for patients with 83 prostheses.Two out of these four abutment teeth fractures happened in prostheses in bruxers patients (2/20) and the two others in nonbruxers (2/63), (P = .666,log-rank test).Information about smoking habits was available for patients with 74 prostheses.
The smokers smoked between 5 and 30 cigarettes per day.Six out of the nine abutment teeth extracted due to loss of periodontal support happened in prostheses in smokers or former smokers patients (6/23)   and three in nonsmokers (3/45), (P = .135,log-rank test).Thirteen abutment teeth were affected by caries, and four abutment teeth eventually received endodontic treatment.
Of the 274 implants, 31 failed, of which 26 due to loss of osseointegration and 5 implants fractured.For one implant fracture, it was not known whether the patient was a bruxer or not, and the other four implant fractures happened in bruxers.There was a statistically significant higher rate of prostheses with implant fracture in bruxers (4/20) in comparison to nonbruxers (0/63), (P = .004,log-rank test).
F I G U R E 3 Example of long-span combined tooth-implant-supported fixed dental prosthesis of the study Comparison of the mean number of prosthetic units, prevalence of bruxers, smokers, cantilever, sex and prosthesis material distribution, and follow-up time, between the different combined tooth-implant-supported FDP groups Cases with available information are considered.b Besides the six smokers in group I, there were also two former smokers (these two were also bruxers). c In three cases, the prosthesis was installed in patients which were both bruxers and smokers.
For when the information about smoking habits was available, eight of the prostheses with any implant failure happened in smokers or former smokers (8/29), in comparison to seven of nonsmokers (7/45), (P = .112,log-rank test).When only loss of osseointegration (and not fracture) was considered as cause of implant failure, seven of the prostheses with implant failure happened in smokers or former smokers (7/28), in comparison to four of nonsmokers (4/42), (P = .059,log-rank test).Sixteen patients lost 18 implants that were directly connected to teeth (the implants were directly adjacent to a tooth or teeth), and 10 patients lost 13 implants that were not directly connected to teeth, that is, the implants that had another implant between itself and a tooth, but that were part of the combined tooth- Then, 20 of the 96 prostheses failed.There was no statistically significant difference between failures of prostheses installed in women and men (P = .753,log-rank test).Patients with prostheses that failed (mean ± SD 57.3 ± 9.9 years, min-max 35.9-57.The reasons for failure are presented in Table 3.The main reason was the installation of new fixed partial prostheses, one (or more) prosthesis only with implants and the other (or more) prosthesis only with teeth, no longer connecting implants to teeth, due to mechanical/biological complications (n = 6), followed by failure of the implant(s) included in the prosthesis (n = 5).
Prostheses with cantilevers did not show a statistically significant higher failure rate in comparison to prostheses without cantilevers (P = .906,log-rank test).Bruxers presented a statistically significant higher prosthesis failure rate in comparison to nonbruxers (P = .022,log-rank test).Of the eight cases of prosthesis failure in bruxers, four cases involved the failure of all implants included in the prosthesis, in two cases new prostheses no longer connecting implant and teeth were installed due to constant minor complications, one case involved the loss of the abutment tooth of the prosthesis due to root fracture, and in one case, a bar overdenture was installed after the failure of some implants of the implant-tooth prosthesis.
Smokers did not present a statistically significant higher prosthesis failure rate in comparison to nonsmokers (P = .575,log-rank test).
Of the seven cases of prosthesis failure in smokers, three cases involved the failure of all implants included in the prosthesis, two cases involved loss of both implants and teeth due to loss of periodontal support, in one case, there was loss of the abutment tooth due to loss of periodontal support, and in one case, new prostheses no longer connecting implant and teeth were installed due to constant minor complications.
Table 4 shows the total number of failures (teeth, implants, prostheses) and complications between the groups.Fifty-three prostheses (55.2%) presented some type of technical complication and/or loss of at least one supporting tooth and/or implant (only failures after prosthesis installation are considered), but when pain in teeth/implants was not considered, as this is a subjective symptom.Twenty-eight (29.2%) out of 96 prostheses presented loss of at least one supporting tooth and/or implant, and 44 (45.8%) presented at least one technical complication (excluding pain in teeth/implants).Table 5 shows a comparison of the prevalence of complications between the different prosthetic groups.Some complications were most commonly observed in long-span prostheses (group III), such as cross-sectional fracture of the prosthesis and the occurrence of fracture screws, and others in middle-span prostheses (group II), such as loose screws and porcelain fractures.However, no group had a general higher prevalence of complications in comparison to the other groups.
Table 6 shows a comparison between failed and nonfailed prostheses according to different factors, besides the results of the Cox model.Bruxism was suggested to be the only factor to statistically significantly exert some influence on the occurrence of prosthesis failure.
As there was only one variable that was moderately associated (P < .10)with prosthesis failure, no multivariable Cox regression model was performed.

| DISCUSSION
The use of implant therapy in special populations requires consideration of potential benefits to be gained from the therapy.To better appreciate this potential, the present study aimed to assess the clinical outcome of fixed restorations supported by the combination of teeth and dental implants.
The installation of implant-tooth prosthesis was, in eight cases of the present study, the choice of treatment after some implants failed before the prosthetic abutment connection, for prostheses that were initially planned to be solely implant supported.The connection of the remaining implants with teeth, together with the refusal of the patients to be submitted to further implants surgeries, was a way to better take advantage of the abutments available in the jaws.
Although the combination of teeth and implants as support for fixed Reasons for prosthesis failure, according to the combined tooth-implant-supported FDP group Installation of new fixed partial prostheses: one (or more) prosthesis only with implants and the other (or more) prosthesis only with teeth, no longer connecting implants to teeth, due to constant mechanical/biological complications Failure of the implant(s) included in the prosthesis Loss of both implants and teeth of the prosthesis, due to loss of periodontal support Loss of the tooth (teeth) included in the prosthesis, due to caries Loss of the tooth (teeth) included in the prosthesis, due to loss of periodontal support Loss of the tooth (teeth) included in the prosthesis, due to root fracture Failure of some implants included in the prosthesis, and change from a fixed prosthesis to an overdenture prostheses remains under debate, because it is not a frequent choice among many practitioners, 2 dentists and patients could benefit from being open to this alternative of treatment, especially in patients that received multiple implants and one or more implants fail before the installation of a prosthesis, as early failure may affect a considerable rate of implants. 15,16review suggested that a greater number of problems occur to teeth than to implants with combined tooth-implant-supported FDPs, 17 but this was not the case in the present study.The loss of a tooth (or teeth) or an implant (or implants) in implant-tooth prostheses may be the reason for the failure of the whole prosthesis, and this was the reason for the failure of most prostheses in the present study.
It is, therefore, important to try to identify the factors that have led to the failure of these abutments.Bruxism was a factor suggested to exert a significant influence on the failure of the prostheses of the present study, although a "definite" diagnosis of bruxism was not established for the patients in the present study, and this was a limitation.Despite that, bruxism has been shown to significantly affect the implant failure 11,18,19 or implant fracture rates negatively 20 .The condition is suggested to generate overload of prosthetic rehabilitations on implants, which could possibly cause implant fracture or peri-implant marginal bone loss (MBL), ultimately resulting in implant failure. 21Some reports have also noted a possible association between bruxism and survival of combined tooth-implant-supported FDPs. 22,23ss of abutment teeth or implants was observed to have a higher prevalence in smokers/former smokers than in nonsmokers, 26.1% vs 6.7% for teeth, respectively, and 35% vs 9.5% for implants, respectively.Even if not reaching statistically significant difference, which could be due to the relative small number of patients, these figures may very well be clinically relevant, although there are no standards for calculating clinically important changes in outcomes. 24Smoking is a factor that has the potential to negatively affect healing and the outcome of implant treatment, 25 and has a detrimental effect on the incidence and progression of periodontitis on natural teeth. 26ostheses that are solely implant or tooth supported are subjected to the same deleterious influence of these habits as combined tooth-implant-supported FDPs.8][29][30][31][32] Moreover, studies comparing solely implant-supported and tooth-implant prostheses demonstrated comparable results regarding the technical and biological complications between these two treatments.with different results.A study observed that there was a clear tendency of more implant failures and tooth complications in combined tooth-implant-supported FDPs. 35Another study observed that there was significantly more MBL in combined tooth-implant-supported FDPs in comparison to solely implant-supported prostheses, 36 contrasting with the results of Gunne et al. 28 Twenty prostheses failed, and some clinicians might be tempted to calculate the general failure rate of the study.However, this would not be appropriate, as this outcome was observed over time and not all participants were observed for the same time therefore censoring has occurred.Therefore, all statistics should include time to event methods, 37 namely, the methods of survival analysis.Still, the results of the life table analysis should be interpreted with caution.The estimated CSR after 26 years was 58.4%.
However, there was no change in the CSR from the year 21, as there were no recorded failures.Numbers entering the interval are low and the censored numbers are proportionally high, reducing the confidence of the outcomes. 37The most recent observations are the least reliable because of the decreasing number of patients at risk for the event of interest.The information about bruxism and smoking habits was not available for every patient.That is why the number of failed and nonfailed prostheses do not add up to the total number of prostheses of the study.b Smokers and former smokers included.
It is a matter of debate whether combined tooth-implantsupported FDPs should be connected rigidly or nonrigidly.Some authors support the rigid connection while others the nonrigid one (see Tsaousoglou et al 17 for details).Although it seems that a higher rate of prosthetic technical complications occur in nonrigid screwretained prostheses, 39 there is limited evidence that rigid connection between teeth and implants present better results when compared with the nonrigid one, according to a recent systematic review on the subject. 17Unfortunately, the cases of the present study were all manufactured with nonrigid connections, preventing any comparison.
Some studies have tested the use of precision attachments to connect the tooth-and implant-supported segments, and cases of intrusion of the natural-tooth portion were noted. 1,3,40,41Cases of intrusion were also reported with the use of copings on a naturaltooth abutment. 42,43The lack of reports of teeth intrusion in the present study may be related to the limitations of its retrospective nature.However, from a prosthetic design perspective, it has been suggested that the use of semi-precision attachments and telescopic systems-the system used for the present cases-may overcome the different biomechanical behaviors of implants and teeth, [40][41][42]44 decreasing the chance of intrusions.
The limitations of the present study include, first, the fact that this is a retrospective study, which inherently results in flaws, manifested by gaps in information and incomplete records.Second, the lack of information on biological factors such as oral hygiene status, bleeding on probing, and probing pocket depth, is also related to the retrospective nature of the present study.Third, a "definite" diagnosis of bruxism was not established for the patients in the present study.Unfortunately, there is a scarcity of reliable and valid diagnostic tools for the condition. 9Fourth, since this was not a prospective study, treatment was not standardized.Having this in mind, the study population was heterogeneous regarding prosthesis design, such as the presence and the extension of cantilevers, or the presence and number of pontics, forming small groups of the precise type of prosthesis configuration.Reduced sample sizes increase the chance of assuming as true a false premise.Moreover, several professionals were involved in the treatment of these patients for the long time of observation of the study, which could had some influence on the failure of implants. 45,46All these confounding factors may have affected the outcomes and not just the fact that the FDPs were supported by the combination of teeth and dental implants.

| CONCLUSIONS
The estimated CSR of the combined tooth-implant-supported FDPs was 90.7%, 84.8%, 69.9%, and 66.2% at 5, 10, 15, and 20 years, respectively.This kind of prosthesis is an alternative to the solely implant-supported prostheses, and can be a choice of treatment after the occurrence of some early implant failures, for prostheses that were initially planned to be solely implant-supported, especially when patients refuse to be submitted to additional implant installation surgeries.

"
possible" sleep or awake bruxism should be based on self-report, by means of questionnaires and/or the anamnestic part of a clinical examination."Probable" sleep or awake bruxism should be based on self-report plus the inspection part of a clinical examination."Definite" sleep bruxism should be based on self-report, a clinical examination, and a polysomnographic recording, preferably along with audio/video recordings.
(a) implant-related factors: implant surface (turned/machined or moderately rough surfaces), implant system; (b) site-related factor: implant jaw location (maxilla/mandible); (c) patient-related factors: patient's sex, age of the patient at the day of the prosthesis installation, smoking habits, bruxism; (d) prosthetic factors: prosthesis extension (in prosthetic units), type and number of units (tooth, implant, pontic, cantilever), complications (listed above); and (e) other factors: failure, reason for failure, and time until failure (for teeth, implants, and prostheses), and follow-up time.
22,28,[31][32][33][34]There are, however, studiesT A B L E 4Comparison of the prevalence of failures (teeth, implants, prostheses) and of complications between groups At least one occurrence-the same prosthesis may have had more than one event throughout the follow-up.
a Unit of analysis: prosthesis.b c n-total number of events.d Unit of analysis: implant.e Unit of analysis: tooth.f For the metalloacrylic prostheses.g For the metalloceramic prostheses. 38 Comparison between failed and nonfailed prostheses according to different factors, and HR estimated from univariate Cox proportional hazard models a