Abstract
OBJECTIVES
(1) quantify 12-month failures of restorations that were repaired or replaced at baseline; (2) test the hypothesis that no significant differences exist in failure percentages between repaired and replaced restorations after 12 months; (3) test the hypothesis that certain dentist’s, patient’s and restoration’s characteristics are significantly associated with the incidence of restoration failure.
METHODS
This prospective cohort study included dentists in the National Dental Practice-Based Research Network. Dentists recorded data for 50 or more consecutive defective restorations. The restorations that were either repaired or replaced were recalled after 12 months and characterized for developing defects.
RESULTS
195 dentists recorded data on 5,889 restorations. 378 restorations required additional treatment (74 repaired, 171 replaced, 84 teeth received endodontic treatment, and 49 were extracted). Multivariable logistic regression analysis indicated that additional treatment was more likely to occur if the original restoration had been repaired (7%) compared to replaced (5%)(OR = 1.6, p < .001; 95% CI: 1.2, 2.1), if a molar tooth was restored (7%) compared to pre-molar or anterior teeth (5%, 6% respectively)(OR = 1.4, p = .010; 95% CI: 1.1, 1.7), and if the primary reason was a fracture (8%) compared to other reasons (6%)(OR = 1.3, p = .033; 95% CI: 1.1, 1.6).
CONCLUSION
An additional treatment was more likely to occur within the first year if the original restoration had been repaired (7%) compared to being replaced (5%). However, repaired restorations were less likely to need an aggressive treatment (replacement, endodontic treatment, or extraction) than replaced restorations.
Keywords: longevity, practice-based research, repair, replacement, decision, defective, restorations, cohort
INTRODUCTION
The longevity of restorations and the cost of replacing restorations are two significant factors determining the long-term cost of restorative therapy. Many factors affect the longevity of restorations, including the restoration quality at the time of insertion, the type and size of the restoration, the restorative material involved, practitioner’s knowledge and experience in secondary caries diagnosis, patient factors like oral hygiene, patient’s age, dentition, and caries risk, and if the patient maintains regular recall appointments in the same dental practice1–6. Most failures occur several years after the restoration was inserted and they are a result of gradual development of secondary caries, some physical defects, such as fracture of restoration or tooth or discoloration of the restoration, or some form of degradation, like marginal breakdown or ‘ditching’7.
Repair of defective restorations rather than replacement of the entire restoration has been a somewhat controversial treatment. The major advantage of repair treatment is that it saves tooth structure8–12 and patient-chair time. It also places minimal stress on the pulp of the tooth. The approach therefore is consistent with the concept of minimally invasive dentistry. However, because it has not been widely accepted as an alternative treatment, not many clinicians have incorporated this practice into routine care13. Therefore, assessing the clinical survival of this treatment, especially in the first twelve months after treatment, is of paramount importance.
Longitudinal studies that assess failure of existing restorations and explore the reasons for failure may provide information to increase the longevity of restorations7. Restorations inserted in practice-based studies provide a unique opportunity to follow-up these restorations in a “real-world setting”, i.e., the ultimate-test of dental restorations, as the clinical conditions are not controlled14. The information gathered from a practice-based setting may improve the longevity of restorations over time, as clinicians can learn the outcome of both types of treatments and hopefully make a decision based on evidence from actually treating existing defective restorations. Therefore, the specific aims of this study were to: (1) quantify the annual failure rate of restorations that were repaired or replaced at baseline; (2) test the hypothesis that there is no significant difference in longevity of restorations that have been either repaired or replaced; and (3) test the hypothesis that some dentist and patient characteristics are significantly associated with the incidence of restoration failure.
METHODS
Selection and recruitment process
This prospective cohort study included 195 dentists of the Dental Practice-Based Research Network (DPBRN) that existed from 2003 to 2012 with a grant from the National Institute of Dental and Craniofacial Research, National Institutes of Health15. DPBRN subsequently evolved into The National Dental PBRN, a consortium of dental practices and dental organizations focused on improving the scientific basis for clinical decision making. The data for this study were collected under the auspices of the DPBRN from 2008 to 2009, and the manuscript of this article was prepared under the aegis of The National Dental PBRN.
At the time of this study, the network was composed primarily of clinicians from five regions: Alabama/Mississippi (AL/MS); Florida/Georgia (FL/GA); dentists in Minnesota, either employed by HealthPartners in Bloomington, MN, or in private practice; Permanente Dental Associates (PDA), in cooperation with Kaiser Permanente’s Center for Health Research in Portland, OR; and dentists from Denmark, Norway and Sweden (SK). Each of the 195 participating dentists recorded data for 50 or more consecutive restorations deemed defective during clinical visits. Practice structures differed some by network region. Dentists from the AL/MS and FL/GA regions were primarily from solo or small group practices (SP), Health Partners and PDA are large group practices (LGP), and SK dentists were from solo or small group private practices (SP) or public health care settings (PHS). Results from previous studies confirm that dentists in practice-based research networks have much in common with dentists at large16–17. The Institutional Review Boards of each participating region approved the study.
Network dentists were recruited through continuing education courses and/or mass mailings to licensed dentists within the participating regions. As part of the eligibility criteria, all dentists completed (1) an enrollment questionnaire describing their demographic and practice characteristics and certain personal characteristics, (2) an assessment of caries diagnosis and caries treatment questionnaire, (3) training in human subject’s protection, and (4) an in-practice network orientation session with the regional coordinator. Copies of the questionnaires and summary data for dentists’ demographic and practice characteristics are also available at http://www.nationaldentalpbrn.org/study-results.php (“Longitudinal study of repaired or replaced dental restorations”).
This study initially used a consecutive patient/restoration recruitment design to gather baseline data. Once the study was started, every patient scheduled to have a repair or replacement of a restoration on a permanent tooth was asked to participate until 50 restorations were enrolled by a single practitioner. Patients who returned for additional appointments while data collection was still ongoing were not eligible for further data collection. In order to increase the number of patients, a maximum of four eligible restorations per patient were enrolled during the first appointment. Restorations discovered after the first appointment were not eligible. A consecutive patient/restoration log form was used to record information on eligible restorations whether or not the patient participated in the study. All the data collection forms used for this study are available at http://www.nationaldentalpbrn.org/study-results.php (under the tabs “Longitudinal study of repaired or replaced dental restorations” and “Reasons for replacement or repair of dental restorations”).
The restorations that were repaired or replaced were recalled after 12 months and characterized for quality according to defined criteria. Acceptable: The restoration is of satisfactory quality and is expected to protect the tooth and the surrounding structures or has one or more features that deviate from ideal conditions, but it does not need to be replaced or repaired. Not acceptable: Additional treatment is necessary because future damage to the tooth and/or surrounding tissues is likely to occur or is occurring,
Reasons for restoration failure
“Secondary/recurrent caries” constituted a lesion detected at the margin of an existing restoration. The lesion had the same characteristics as primary caries lesions.
‘Entire restoration was discolored’ included any mismatch between the color of the body of a tooth-colored restoration and the tooth that led to replacement of the restoration.
‘Restoration margins were discolored’ were found at the tooth/restoration interface and led to repair or replacement.
‘Bulk fracture’ of a restoration included isthmus fracture or any fracture through the body of the restoration or the marginal ridge, but with the restoration still in place.
‘Restoration margins were degraded or ditched’, only those restorations with marginal fractures or degraded margins, but without caries were recorded in this category of failure.
‘Restoration was missing’ was recorded when either all of the restoration or a major part was missing due to lack of retention.
“Tooth was fractured’ was any kind of tooth fracture adjacent to a restoration, for example the fracture of a cusp or of an enamel margin.
‘Pain/sensitivity’ of any kind that required repair or replacement of a restoration, or endodontic treatment or extraction of the tooth.
‘Patient request’ included any reason for replacement of a restoration deemed acceptable by the practitioner.
‘Other reason” included any other reason for replacement/repair of restorations and endodontic treatment/extraction of teeth than those listed above.
If a serviceable or intact restoration had been replaced because it was incorporated into a larger restoration, it was not recorded as a failure. An example would be the removal of an intact occlusal Class I restoration that became part of a Class II restoration.
Variable selection
Repair treatment was characterized as the removal of part of the existing restoration and any adjacent pathologically altered as well as esthetically unacceptable tooth tissue followed by placement of restorative material in the prepared site. Repair also included light grinding and polishing, removal of overhangs, polishing discolored tooth-colored restorations, or sealing margins. Restoration replacement was characterized as the entire removal of the existing defective/failed restoration and any adjacent pathologically altered and discolored tooth tissue that was esthetically or functionally unacceptable.
Practitioners collected data for each enrolled restoration that needed repair or replacement on permanent tooth surfaces. Data collected included: (1) the main reason for repair or replacement of the restoration 10; (2) tooth type and tooth surfaces being restored; and (3) the restorative materials used for the old and the new restoration. Dentists diagnosed the need to repair or replace the existing restoration based on the diagnostic methods they typically use in their practice, which consist mainly of visual-tactile in association with radiographic examinations.
Restorative materials were classified as amalgam, direct or indirect resin-based composite (RBC or IRBC), conventional or resin-modified glass-ionomer (GI/RMGI), ceramic or porcelain, cast gold or other metallic-based material, combined metal-ceramic material, and temporary restorative materials. Information about gender, age, race, ethnicity, and insurance coverage of enrolled patients was also recorded in the same recording form.
Statistical analysis
Descriptive statistics were calculated for dentist, patient, and restoration variables. A binary logistic model was used to test the study hypotheses, with Generalized Estimating Equations to adjust for clustering within dentists and restorations within patients. A series of three outcomes was tested using all treated and untreated restorations. First the model was tested in all restorations seen by the one-year follow-up and the dependent variable was coded no treatment = 0 and treatment (repair, replacement, endodontic, and extraction) = 1. The prediction model included treatment for the original defective restoration (repair = 1, replacement = 0); tooth (molar = 1, and premolar or anterior = 0); arch (upper = 1 and lower = 0); original restorative material (amalgam = 1, direct tooth colored/ indirect tooth colored/ gold = 0); repair/replacement restorative material (amalgam = 1, direct tooth colored/ indirect tooth colored/ gold = 0); fractured restoration (fractured restoration = 1 when the primary reason for the defect/failure and coded = 0 when fractured restoration was not the primary reason for the defect/failure); the number of surfaces in the original restoration classified as 1, 2, 3, 4 and 5. Information about patient (gender, age), dentist (gender, years in practice) and practice model were entered in an initial step and included in subsequent testing if they were significant at p ≤ .015. Two-way interactions involving the “treatment for the original defective restoration” variable with tooth, arch, the restoration material variables, the fracture variable, and the number of surfaces were tested individually. For all significant interactions, models were run separately for (1) replacement of the original defective restoration and (2) repair of the original defective restoration.
The next two models examined only restorations that were treated by the one-year follow-up. The dependent variable for the second model was coded repair = 0 and replacement, endodontic, or extraction = 1. The dependent variable for the third model was coded repair or replacement = 0 and endodontic or extraction = 1. Interactions were not tested for models 2 or 3 because of sample size limitations.
RESULTS
In the original study of defective restorations, data were available for 8,921 restorations with complete data from 6,759 patients. First-year follow-up data were provided from 195 network dentists on 4,648 patients involving 6,059 restorations. Of these, 170 restorations received a temporary restoration and are not included resulting in 5,889 restorations in 4,482 patients in the following analyses. Of these, 378 (6.4%) required additional treatment as follows: 74 (1.3%) were repaired, 171 (2.9%) were replaced, 84 (1.4%) teeth received endodontic treatment, and 49 (0.8%) were extracted.
Dentists were distributed across the network regions as AL/MS=39, FL/GA=43, PDA=40, MN=35, SK=38. Regarding patient characteristics, 58% were female and 42% were males, 25% had dental insurance. Patient race as indicated by the dentist was non-Hispanic White, 82%; Hispanic, 10%; Black or African American, 5%; and Other, 3%. Practice and patient characteristics are presented in Table 1.
Table 1.
Practice and patient variables
| VARIABLE | MEAN (SD) OR % (N) |
|---|---|
| Dentists’ characteristics (n=195) | |
| Gender (male) | 70% (136) |
| Years since dental school graduation | 23 (SD=10.3) |
| Average number of patients seen in a week (missing=9) | 47 (SD=16.5) |
| Practice type | |
| - Solo or small group private practice | 57% (112) |
| - Large group practice | 37% (72) |
| - Public health care setting | 6% (11) |
| Race/ethnicity | |
| - Hispanic White | 5% (10) |
| - Non-Hispanic-White | 84% (164) |
| - Non-Hispanic Black | 3% (5) |
| - Other | 8% (16) |
| Placed the original restoration (missing=9) | 75% (4391) |
| Patients’ characteristics (n=4,482) | |
| Patient gender (male) missing=8 | 42% (1878) |
| Patient age | 53 (SD=15.3) |
| Race – ethnicity (missing=60) | |
| - Non-Hispanic White | 83% (3744) |
| - Hispanic | 8% (356) |
| - Black | 5% (214) |
| - Other | 3% (114) |
| Dental insurance or any third party coverage | 25% (1136) |
Analysis of all restorations
Descriptive statistics for dentist, patient, and restoration variables are listed in Table 2. Table 3 presents regression coefficients for regression models tested. The results of multivariable logistic regression analysis indicated that a treatment at one-year was more likely if the defective restoration had been repaired (7%) compared to replaced (5%; OR = 1.6, p < .001; 95% CI: 1.2, 2.1), if a molar tooth was involved in the treatment (7%) compared to pre-molar or anterior teeth (5%, 6% respectively; OR = 1.4, p = .010; 95% CI: 1.1, 1.7), and if the primary reason for the defect was a fracture (8%) compared to other reasons (6%; OR = 1.3, p = .033; 95% CI: 1.1, 1.6). The number of surfaces involved in the treatment of the defective restoration, whether amalgam material was used in either the original restoration or in treatment of the defective restoration was not associated with needing a treatment by the first year follow-up. The practice model was the only dentist or patient variable that was significant, with treatment more likely to occur for dentists in a large group practice (12%), compared to small group and private practices (5%; OR = 2.2, p< .001; 95% CI: 1.7, 2.9).
Table 2.
Restoration failure and their relationship with dentists’ and tooth’s characteristics
| VARIABLE | Restorations in study % (n) n=5,889 | Restorations that failed % (n) n = 378 |
|---|---|---|
| Practice Model | ||
| SP | 74% (4,358) | 5% (213) |
| LGP | 22% (1,283) | 12% (148) |
| PHS | 4% (248) | 7% (17) |
| Decade of dental school graduation | ||
| - 1979 | 29% (1,680) | 6% (76) |
| - 1980–1989 | 45% (2,630) | 6% (156) |
| - 1990–1999 | 13% (761) | 7% (53) |
| - 2000–2008 | 14% (818) | 8% (68) |
| Treatment | ||
| Repaired | 25% (1,498) | 7% (144) |
| Replaced | 75% (4,391) | 5% (234) |
| Tooth | ||
| Molar | 54% (3,148) | 7% (228) |
| Upper | 25% (1,483) | 8% (118) |
| Lower | 28% (1,665) | 7% (110) |
| Pre Molar | 26% (1,515) | 5% (81) |
| Upper | 14% (816) | 5% (37) |
| Lower | 12% (699) | 6% (44) |
| Anterior | 21% (1,226) | 6% (69) |
| Upper | 16% (938) | 5% (42) |
| Lower | 5% (288) | 9% (27) |
| Number of surfaces in repair/replacement | ||
| One | 27% (1,589) | 6% (87) |
| Two | 32% (1,869) | 7% (114) |
| Three | 22% (1,294) | 7% (81) |
| Four | 9% (512) | 9% (48) |
| Five | 11% (625) | 6% (48) |
| Material of the original restoration | ||
| Amalgam | 53% (3,054) | 6% (194) |
| Direct tooth-colored | 37% (2,154) | 6% (135) |
| Indirect tooth-colored | 11% (606) | 7% (43) |
| Missing information (n=81) | ||
| Repair/replacement material | ||
| Amalgam | 21% (1,215) | 10% (117) |
| Direct tooth-colored | 58% (3,338) | 6% (182) |
| Indirect tooth-colored | 22% (1,252) | 6% (71) |
| Missing information (n=92) | ||
| Reason for repair or replacement | ||
| Secondary/recurrent caries | 43% (2,514) | 7% (167) |
| Fracture/bulk fracture/missing | 36% (2,095) | 8% (157) |
| Degraded/ditched | 8% (461) | 4% (16) |
| Other | 7% (399) | 4% (16) |
| Margins or restoration discolored | 3% (196) | 3% (6) |
| Patient request | 2% (127) | 5% (6) |
| Pain sensitivity | 1% (65) | 9% (6) |
| Missing information (n=36) | ||
SP = solo or small group practices
LGP = large group practices
PHS = public health care settings
Table 3.
Baseline characteristics associated with whether or not the restoration received treatment of any type during follow-up
| VARIABLE | B (SE) | p value | OR (95% CI) |
|---|---|---|---|
| Model 1: All restorations (n=5,687)a | |||
| Treatment (repair) | .470 (.141) | < .001 | 1.6 (1.2, 2.1) |
| Original material (amalgam) | −.163 (.132) | .217 | 0.8 (0.7, 1.1) |
| Replacement material (amalgam) | .222 (.179) | .215 | 1.2 (0.9, 1.6) |
| Tooth site (Molar) | .310 (.133) | .010 | 1.4 (1.1, 1.7) |
| Reason for failure (fracture) | .260 (.122) | .033 | 1.3 (1.1, 1.6) |
| Surfaces | −.046 (.051) | .374 | 0.9 (0.9, 1.1) |
Model for all restorations included 5,687 restorations because of 202 missing values among the predictor variables. The interactions between the repair/replace variable and both the number of surfaces in the treated restoration (p = .011) and if amalgam was used in the treatment of the defective restoration (p = .021) were significant.
The interactions between the repair/replace variable and both the number of surfaces in the treated restoration (p = .011) and if amalgam was used in the treatment of the defective restoration (p = .021) were significant. Other interactions were not significant. To interpret the significant interactions, the data were then separated into restorations that were replaced and those that were repaired (Table 4). Multivariable logistic regression analysis indicated when the treatment for the defective restoration was a replacement, a treatment at one-year was more likely if the defective restoration had been replaced with amalgam (9%) compared to all other materials (4%; OR = 2.3, p = .029 95% CI: 1.7, 3.2). There were no differences when the treatment for the defective restoration had been a replacement. In a similar analysis of defective restorations that were repaired, the greater the number of surfaces involved in the repair, the more likely it was to receive additional treatment at one-year (OR = 1.3, 95% CI: 1.1, 1.6, p = .004), however the number of surfaces was not associated with subsequent failure for replacements.
Table 4.
Baseline characteristics associated with whether or not the restoration received treatment of any type during follow-up, tested separately for restorations that were replaced and restorations that were repaired.
| VARIABLE | B (SE) | p value | OR (95% CI) |
|---|---|---|---|
| Model 2: Replaced restorations (n=4,226) a | |||
| Original material (amalgam) | −.263 (.143) | .095 | 0.8 (0.6, 1.0) |
| Replacement material (amalgam) | .846 (.170) | <.001 | 2.3 (1.7, 3.2) |
| Tooth site (Molar) | .205 (.097) | .014 | 1.4 (1.1, 1.9) |
| Reason for failure (fracture) | .138 (.162) | .395 | 1.1 (0.8, 1.5) |
| Surfaces | .034 (.051) | .504 | 1.0 (0.9, 1.1) |
| Model 3: Repaired restorations (n=1,461) b | |||
| Original material (amalgam) | −.039 (.211) | .848 | 1.0 (0.7, 1.4) |
| Replacement material (amalgam) | .223 (.225) | .320 | 1.3 (0.9, 1.9) |
| Tooth site (Molar) | .366 (.228) | .109 | 1.2 (0.9, 2.0) |
| Reason for failure (fracture) | −.051 (.119) | .669 | 1.0 (0.8, 1.2) |
| Surfaces | .295 (.103) | .004 | 1.3 (1.1, 1.6) |
Model for replaced restorations included 4,226 restorations because of 165 missing values among the predictor variables.
Model for repaired restorations included 1,467 restorations because of 37 missing values among the predictor variables
Analysis of the treated restorations only
Table 5 presents regression coefficients for predictors of the treatments received for the 378 restorations that were treated during the year 1 follow-up and Table 6 shows the outcomes stratified by whether the defective restoration was originally repaired or replaced. When the restoration required a treatment during the one-year follow-up, it was less likely to need a replacement, endodontic treatment, or extraction if the defective restoration had been repaired (74%) than had it been replaced (85%; OR = 0.5, 95% CI: 0.3, 0.9, p. = .018).
Table 5.
Predicting type or treatment received among restorations that were treated by the one year follow-up.
| B (SE) | p. value | OR (95% CI) | |
|---|---|---|---|
| Treatment was replacement, endo, extraction a | |||
| Treatment (repair) | −.839 (.299) | .005 | 0.4 (0.2, 0.8) |
| Original material (amalgam) | −.382 (.316) | .227 | 0.7 (0.4, 1.3) |
| Replacement material (amalgam) | −.317 (.313) | .311 | 0.7 (0.4, 1.3) |
| Tooth site (Molar) | .306 (.122) | .012 | 1.4 (1.1, 1.7) |
| Reason for failure (fracture) | −.043 (.288) | .958 | 1.0 (0.5, 1.7) |
| Surfaces | −.180 (.117) | .124 | 0.8 (0.7, 1.1) |
| Treatment was endo or extraction b | |||
| Treatment (repair) | −.706 (.261) | .007 | 0.5 (0.3, 0.8) |
| Original material (amalgam) | −.183 (.263) | .487 | 0.8 (0.5, 1.4) |
| Replacement material (amalgam) | .711 (.287) | .013 | 0.5 (0.3, 0.9) |
| Tooth site (Molar) | .346 (.244) | .157 | 1.4 (0.9, 2.3) |
| Reason for failure (fracture) | .292 (.230) | .204 | 1.3 (0.9, 2.1) |
| Surfaces | .069 (.093) | .457 | 1.1 (0.9, 1.3) |
The dependent variable was coded as repair = 0 and replacement, endodontic, or extraction = 1.
The dependent variable was coded repair or replacement = 0 and endodontic or extraction = 1.
These models examined only the 378 restorations that were treated by the one-year follow-up. Interactions were not tested because of sample size limitations.
Table 6.
Outcomes for restorations treated after 1 year
| RESTORATION OUTCOME AT ONE YEAR | |||||
|---|---|---|---|---|---|
| TREATMENT AT BASELINE | Repair | Replace | Endodontic | Extraction | Total |
| Repaired restoration | 26% (38) | 49% (70) | 12% (17) | 13% (19) | 144 |
| Replaced restoration | 15% (36) | 43% (101) | 29% (67) | 13% (30) | 234 |
| Total | 20% (74) | 45% (171) | 22% (84) | 13% (49) | 378 |
When the restoration required a treatment during the one-year follow-up, it was less likely to need endodontic treatment or extraction if the defective restoration had been repaired (25%), when compared to restorations that had been replaced (42%; OR = 0.4, 95% CI: 0.3, 0.8, p. = .005). Endodontic treatment or extraction were also less likely for treated restorations if the material used in the repair/replacement was amalgam (23%), compared to other restorative materials (41%; OR = 0.5, 95% CI: 0.3, 0.8, p. = .007). Inspection of Table 5 indicates that the difference between repaired and restored restorations in model 3 results from the greater number of endodontic treatments for the restorations that were originally replaced (29%) compared to those which were repaired (12%) and not from differences in the percentages which were extracted (both 13%).
DISCUSSION
Numerous factors affect the longevity of restorations, and these may be subdivided into operator factors, quality of the restorative material, and patient/restoration factors3. The current study showed that restorative material, number of surfaces involved, tooth type, and primary reason for treatment were associated with the short-term longevity of repaired and replaced restorations in permanent teeth.
Repaired restorations were more likely to receive additional treatment 7% of the time compared to 5% for replaced restorations. However, repaired restorations received a more-conservative treatment during follow-up, such as repair (repaired 26% of the time and replaced/endo/extracted 74%) compared to restorations that had been replaced (repaired only 15% of the time and replaced/endo/extracted 85%). Additionally, replaced restorations were more likely to get endodontic treatment (29%) compared to repaired restorations (12%, see table 5). Although repaired restorations may require a treatment more often during the first year after treatment, they require a more-conservative treatment that does not involve the replacement of the entire restoration or even worse, endodontic and extraction treatments. Several studies have shown that restorations that are replaced lose a significant amount of healthy tooth structure8–9,18, invariably speeding the re-restoration cycle19–21.
Type of tooth was a significant variable. Molar teeth (7%) received additional treatment more often than any other teeth (pre-molar 5% and anterior 6%). A possible explanation may be that molar teeth receive most of the biting forces22. Also related to the same line of thought, restorations involving a higher number of surfaces were repaired and more likely to require additional treatment23–25. A practice-based research study involving thousands of restorations showed that the number of tooth surfaces restored at baseline helped predict subsequent restoration failure; restorations with four or more restored surfaces were four times more likely to fail26. Interestingly, the number of surfaces did not predict a subsequent failure for restorations that were replaced. In repaired restorations, the interface between the old and new material maybe the weakest link and possibly affect its outcome27–28.
A potential weakness in the current study design is the lack of assessment of patient’s caries risk. Both a literature review and a meta-analysis concluded that longevity of restorations were affected by secondary caries (related to the individual caries risk) and fracture of restoration (related to the strength of the material used), as well as patient factors such as bruxism 5–6.
In the current study, when a fracture was the primary reason for the defect, the restoration was more likely to be re-treated within the first year when compared to other reasons (8% - 6%). Previous studies showed that restorations that were repaired due to fracture had a lower survival rate than restorations that were repaired due to other reasons7, 29–30. Fracture as a reason for failure and amalgam as a restorative material may jointly exert negative influences on the survival of restorations30. If the original restoration was treated due to a restoration fracture or a tooth fracture it may be more likely to fail, especially if the restoration is subjected to the same forces that caused the original failure, leading to subsequent or repeated fracture more often.
With regard to the material used for treatment, if the restoration was replaced, amalgam failed more often (9%) compared to all other materials (4%), more than twice the rate, especially if the tooth was a molar (6%) compared to pre-molar (4%) or anterior (5%). And that makes sense, as molar teeth are often more likely to receive amalgam restorations than any other tooth. When the restoration was repaired though, the subsequent use of amalgam was not associated with one-year failures. In the majority of the restorations that were repaired a resin-based composite material was used as the restorative material. RBC materials are known to bond to the existing tooth surfaces potentially reinforcing the existing tooth structure particularly in large restorations of molar teeth31–32. Furthermore, amalgam restorations are also most likely to be replaced due to patient requests for a more-esthetic material33. It is possible that patients’ expectations for a more-esthetic appearance than that of the existing amalgam material may have influenced the clinician’s decision to replace the existing restoration with a tooth-colored material. However, in the present study very few restorations were originally repaired or replaced because of patient request (2%). Studies have confirmed the decline of amalgam as a restorative material in recent years34–36. While the reasons for replacement of amalgam restorations have remained fairly constant, the reasons for failure of resin-based composite restorations and their longevity have changed markedly during recent years. When resin-based Class II composite restoration came into common use about 15–20 years ago, the difference in longevity between amalgam and composite restorations was quite marked37. However, more recent long-term data indicate that the longevity of composite restorations has increased3. Indeed, in the current study six percent (Table 2) of direct tooth-colored restorations failed within a year. The reason for this change is manifold, including improved material quality and clinical experience of clinicians in handling the materials. Long survival rates for posterior composite restorations can be expected provided that patient, operator and materials factors are taken into account when the restorations are performed5–6. Additionally, other unobserved variables such as practitioners’ preferences and skills may also have played a role in the results.
The longevity of restorations is closely linked to their replacement rate. The diversity of opinions among clinicians about what constitutes a failed restoration is a major problem when reporting on the longevity of restorations. No generally accepted, objective criteria have been established for what degree of failure constitutes a condition that will cause future damage to the tooth or to the patient. In the current study, practice model was the only dentist variable that was significant, with treatment more likely to occur (OR 1.4) for dentists in a large group practice (12%) compared to small group and private practices (5%). In the large group practice, it is possible that patients are seen by different dentists within the same practice. As previously demonstrated, subjective decision-making still prevails38–39 and marked variations between clinicians in diagnosing failures have been demonstrated1–2, 4, 40–42. Studies have shown that dentists are more likely to intervene in a restoration that they had not placed11, 43–45. Another study46, based on insurance claims, suggests that patients who change dentists are far more likely to have restorations replaced. Conclusions about longevity of restorations might also be strongly influenced by clinicians’ thresholds for replacing restorations that they consider “defective”. If dentists have a low threshold for replacing a restoration, then naturally the age of restorations will be shorter.
CONCLUSION
Six percent of restorations received additional treatment such as repair or replacement after one-year of treatment. An additional treatment was more likely to occur if the defective restoration had been repaired (7%) compared to replaced (5%), if a molar tooth was involved in the treatment, if a greater number of surfaces were involved in the original treatment, and if the primary reason for the defect was a fracture. However, when the restoration required a treatment at the one-year follow-up, it was less likely to need an aggressive treatment (replacement, endodontic treatment, or extraction) if the defective restoration had been repaired than if it had been replaced.
The practice model was the only dentist or patient variable that was significant; treatment was more likely to occur for dentists in a large group practice compared to small group private practices.
Acknowledgments
This work was supported by National Institutes of Health grants DE-16746, DE-16747 and DE-22516. Opinions and assertions contained herein are those of the authors and are not to be construed as necessarily representing the views of the respective organizations or the National Institutes of Health. The informed consent of all human subjects who participated in this investigation was obtained after the nature of the procedures had been fully explained.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Valeria V. Gordan, Email: vgordan@dental.ufl.edu.
Joseph L. Riley, III, Email: jriley@dental.ufl.edu.
D. Brad Rindal, Email: Donald.B.Rindal@HealthPartners.Com.
Vibeke Qvist, Email: viq@sund.ku.dk.
Jeffrey L. Fellows, Email: Jeffrey.Fellows@kpchr.org.
Deborah A. Dilbone, Email: ddilbone@dental.ufl.edu.
Solomon G. Brotman, Email: sgbrotman@prodigy.net.
Gregg H. Gilbert, Email: ghg@uab.edu.
References
- 1.Tveit AB, Espelid I. Class II amalgams: interobserver variations in replacement decisions and diagnosis of caries and crevices. Int Dent J. 1992;42(1):12–18. [PubMed] [Google Scholar]
- 2.Bader JD, Shugars DA. Agreement among dentists’ recommendations for restorative treatment. J Dent Res. 1993;72(5):891–896. doi: 10.1177/00220345930720051001. [DOI] [PubMed] [Google Scholar]
- 3.Manhart J, Chen H, Hamm G, Hickel R. Buonocore Memorial Lecture. Review of the clinical survival of direct and indirect restorations in posterior teeth of the permanent dentition. Oper Dent. 2004;29:481–508. [PubMed] [Google Scholar]
- 4.Setcos JC, Khosravi R, Wilson NH, Shen C, Yang M, Mjör IA. Repair or replacement of amalgam restorations: decisions at a USA and a UK dental school. Oper Dent. 2004;29(4):392–397. [PubMed] [Google Scholar]
- 5.Demarco FF, Corrêa MB, Cenci MS, Moraes RR, Opdam NJ. Longevity of posterior composite restorations: not only a matter of materials. Dent Mater. 2012;28:87–101. doi: 10.1016/j.dental.2011.09.003. [DOI] [PubMed] [Google Scholar]
- 6.Heintze SD, Rousson V. Clinical effectiveness of direct class II restorations - a meta-analysis. J Adhes Dent. 2012;14:407–431. doi: 10.3290/j.jad.a28390. [DOI] [PubMed] [Google Scholar]
- 7.Opdam NJ, van de Sande FH, Bronkhorst E, Cenci MS, Bottenberg P, Pallesen U, Gaengler P, Lindberg A, Huysmans MC, van Dijken JW. Longevity of posterior composite restorations: a systematic review and meta-analysis. J Dent Res. 2014 Oct;93(10):943–949. doi: 10.1177/0022034514544217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gordan VV. Clinical Evaluation of Replacement of Class V Resin Based Composite Restorations. Journal of Dentistry. 2001;29:485–488. doi: 10.1016/s0300-5712(01)00030-6. [DOI] [PubMed] [Google Scholar]
- 9.Gordan VV, Mondragon E, Shen C. Evaluation of the Cavity Design, Cavity Depth, and Shade Matching in the Replacement of Resin Based Composite Restorations. Quintessence International. 2002;32:273–278. [PubMed] [Google Scholar]
- 10.Gordan VV, Riley JL, III, Blaser PK, Mondragon E, Garvan CW, Mjor IA. Alternative Treatments to Replacement of Defective Amalgam Restorations: Results of a 7-year Clinical Study. Journal of American Dental Association. 2011;142( 7):842–849. doi: 10.14219/jada.archive.2011.0274. [DOI] [PubMed] [Google Scholar]
- 11.Gordan VV, Riley JL, III, Geraldeli S, Rindal DB, Qvist V, Fellows JL, Kellum HP, Gilbert GH for The DPBRN Collaborative Group. Repair or replacement of defective restorations by dentists in the Dental PBRN. JADA. 2012;143( 6):593–601. doi: 10.14219/jada.archive.2012.0238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Moncada G, Vildósola P, Fernandez E, Estay J, de Oliveira O, Júnior, de Andrade M, Martin J, Mjor IA, Gordan VV. Longitudinal results of a 10-Year Clinical Trial of Repair of Amalgam Restorations. [accepted on 4/22/2014];Operative Dentistry. 2015 doi: 10.2341/14-045-C. [DOI] [PubMed] [Google Scholar]
- 13.Gordan VV, Garvan CW, Richman J, Fellows JL, Rindal DB, Qvist V, Heft MW, Williams OD, Gilbert GH for The DPBRN Collaborative Group. How Dentists Diagnose and Treat Defective Restorations: Evidence from The Dental PBRN. Operative Dentistry. 2009;34:664–673. doi: 10.2341/08-131-C. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Norton WE, Funkhouser E, Makhija SK, Gordan VV, Bader JD, Rindal DB, Pihlstrom DJ, Hilton TJ, Frantsve-Hawley J, Gilbert GH National Dental Practice-Based Research Network Collaborative Group. Concordance between clinical practice and published evidence: findings from The National Dental Practice-Based Research Network. J Am Dent Assoc. 2014 Jan;145(1):22–31. doi: 10.14219/jada.2013.21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gilbert GH, Williams OD, Korelitz JJ, Fellows JL, Gordan VV, Makhija SK, Meyerowitz C, Oates TW, Rindal DB, Benjamin PL, Foy PJ National Dental PBRN Collaborative Group. Purpose, structure, and function of the United States National Dental Practice-Based Research Network. J Dent. 2013 Nov;41(11):1051–1059. doi: 10.1016/j.jdent.2013.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Makhija SK, Gilbert GH, Rindal DB, Benjamin P, Richman JS, Pihlstrom DJ, Qvist V DPBRN Collaborative Group. Practices participating in a dental PBRN have substantial and advantageous diversity even though as a group they have much in common with dentists at large. BMC Oral Health. 2009 Oct 15;9:26. doi: 10.1186/1472-6831-9-26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Makhija SK, Gilbert GH, Rindal DB, Benjamin PL, Richman JS, Pihlstrom DJ DPBRN Collaborative Group. Dentists in practice-based research networks have much in common with dentists at large: evidence from the Dental Practice-Based Research Network. Gen Dent. 2009 May-Jun;57(3):270–5. [PMC free article] [PubMed] [Google Scholar]
- 18.Gordan VV. In vitro evaluation of margins of replaced resin based composite restorations. J Esthet Dent. 2000;12:217–223. doi: 10.1111/j.1708-8240.2000.tb00223.x. [DOI] [PubMed] [Google Scholar]
- 19.Brantley CF, Bader JD, Shugars DA, Nesbit SP. Does the cycle of rerestoration lead to larger restorations? J Am Dent Assoc. 1995 Oct;126(10):1407–13. doi: 10.14219/jada.archive.1995.0052. [DOI] [PubMed] [Google Scholar]
- 20.Mjör IA, Gordan VV. Failure, repair, refurbishing and longevity of restorations. Oper Dent. 2002;27(5):528–534. [PubMed] [Google Scholar]
- 21.Tyas MJ. Placement and replacement of restorations by selected practitioners. Aust Dent J. 2005 Jun;50(2):81–9. doi: 10.1111/j.1834-7819.2005.tb00345.x. [DOI] [PubMed] [Google Scholar]
- 22.Kikuchi M, Korioth TW, Hannam AG. The association among occlusal contacts, clenching effort, and bite force distribution in man. J Dent Res. 1997 Jun;76(6):1316–25. doi: 10.1177/00220345970760061201. [DOI] [PubMed] [Google Scholar]
- 23.Bohaty BS1, Ye Q, Misra A, Sene F, Spencer P. Posterior composite restoration update: focus on factors influencing form and function. Clin Cosmet Investig Dent. 2013;5:33–42. doi: 10.2147/CCIDE.S42044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Balevi B. Caries risk and number of restored surfaces have impact on the survival of posterior composite restorations. Evid Based Dent. 2014;15(4):118–119. doi: 10.1038/sj.ebd.6401065. [DOI] [PubMed] [Google Scholar]
- 25.Pallesen U1, van Dijken JW, Halken J, Hallonsten AL, Höigaard R. Longevity of posterior resin composite restorations in permanent teeth in Public Dental Health Service: a prospective 8 years follow up. J Dent. 2013;41(4):297–306. doi: 10.1016/j.jdent.2012.11.021. [DOI] [PubMed] [Google Scholar]
- 26.McCracken MS1, Gordan VV, Litaker MS, Funkhouser E, Fellows JL, Shamp DG, Qvist V, Meral JS, Gilbert GH National Dental Practice-Based Research Network Collaborative Group. A 24-month evaluation of amalgam and resin-based composite restorations: Findings from The National Dental Practice-Based Research Network. J Am Dent Assoc. 2013;144(6):583–593. doi: 10.14219/jada.archive.2013.0169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gordan VV, Shen C, Mjor IA. Marginal gap repair with flowable resin-based composites. Gen Dent. 2004;52(5):390–394. [PubMed] [Google Scholar]
- 28.Shen C, Mondragon E, Gordan VV, Mjör IA. The effect of mechanical undercuts on the strength of composite repair. J Am Dent Assoc. 2004;135(10):1406–1412. doi: 10.14219/jada.archive.2004.0050. [DOI] [PubMed] [Google Scholar]
- 29.Gordan VV, Riley J, 3rd, Geraldeli S, Williams OD, Spoto JC, 3rd, Gilbert GH National Dental PBRN Collaborative Group. The decision to repair or replace a defective restoration is affected by who placed the original restoration: findings from the National Dental PBRN. J Dent. 2014;42(12):1528–1534. doi: 10.1016/j.jdent.2014.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Opdam NJ1, Bronkhorst EM, Loomans BA, Huysmans MC. Longevity of repaired restorations: a practice based study. J Dent. 2012;40(10):829–835. doi: 10.1016/j.jdent.2012.06.007. [DOI] [PubMed] [Google Scholar]
- 31.Abu-Hanna AA, Mjör IA. Resin composite reinforcement of undermined enamel. Oper Dent. 2004;29(2):234–237. [PubMed] [Google Scholar]
- 32.Abu-Hanna AA, Mjör IA. Combined amalgam and composite restorations. Oper Dent. 2004;29(3):342–344. [PubMed] [Google Scholar]
- 33.Gordan VV1, Riley JL, 3rd, Worley DC, Gilbert GH DPBRN Collaborative Group. Restorative material and other tooth-specific variables associated with the decision to repair or replace defective restorations: findings from The Dental PBRN. J Dent. 2012;40(5):397–405. doi: 10.1016/j.jdent.2012.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sunnegårdh-Grönberg K, van Dijken JW, Funegård U, Lindberg A, Nilsson M. Selection of dental materials and longevity of replaced restorations in Public Dental Health clinics in northern Sweden. J Dent. 2009;37:673–678. doi: 10.1016/j.jdent.2009.04.010. [DOI] [PubMed] [Google Scholar]
- 35.Christensen GJ. Amalgam vs. composite resin. J Am Dent Assoc. 1998;129(12):1757–1759. doi: 10.14219/jada.archive.1998.0148. [DOI] [PubMed] [Google Scholar]
- 36.Ottenga ME, Mjor I. Amalgam and composite posterior restorations: curriculum versus practice in operative dentistry at a US dental school. Oper Dent. 2007;32(5):524–528. doi: 10.2341/06-134. [DOI] [PubMed] [Google Scholar]
- 37.Mjör IA, Jokstad A, Qvist V. Longetivity of posterior restorations. Int Dent J. 1990 Feb;40(1):11–7. [PubMed] [Google Scholar]
- 38.Nuttall NM, Elderton RJ. The nature of restorative dental treatment decisions. Br Dent J. 1983 Jun;154(11):363–5. doi: 10.1038/sj.bdj.4805093. [DOI] [PubMed] [Google Scholar]
- 39.Gordan VV. Diagnosing restorations, and when repair or replacement is required. How successful are repaired restorations? Int Dent J. 2012;62:223–243. [Google Scholar]
- 40.Merrett MC, Elderton RJ. An in vitro study of restorative dental treatment decisions and dental caris. Br Dent J. 1984 Aug;147(4):128–33. doi: 10.1038/sj.bdj.4805448. [DOI] [PubMed] [Google Scholar]
- 41.Maryniuk GA. Quintessence Int. 1990 Apr;21(4):311–9. [PubMed] [Google Scholar]
- 42.Bader JD, Shugars DA. Understanding dentists’ restorative treatment decisions. J Public Health Dent. 1992;52:102–110. doi: 10.1111/j.1752-7325.1992.tb02251.x. [DOI] [PubMed] [Google Scholar]
- 43.Davies JA. The relationship between change in dentist and treatment received in the general dental service. Br Dent J. 1984;157:322–324. doi: 10.1038/sj.bdj.4805482. [DOI] [PubMed] [Google Scholar]
- 44.Gordan VV, Garvan CW, Richman J, Fellows JL, Rindal DB, Qvist V, Heft MW, Williams OD, Gilbert GH for The DPBRN Collaborative Group. How Dentists Diagnose and Treat Defective Restorations: Evidence from The Dental PBRN. Operative Dentistry. 2009;34:664–673. doi: 10.2341/08-131-C. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Bogacki RE, Hunt RJ, del Agila M, Smith WR. Survival analysis of posterior restorations using an insurance claims database. Oper Dent. 2002;27:488–492. [PubMed] [Google Scholar]