| Sign In to gain access to subscriptions and/or personal tools. |
The Timing of Subsequent Treatment for Teeth Restored with Large Amalgams and Crowns: Factors Related to the Need for Subsequent Treatment
1 University of Michigan, School of Dentistry, Department of Cariology, Restorative Sciences, and Endodontics, 1011 N. University Ave., Ann Arbor, MI 48109-1078, USA; Correspondence: * corresponding author, jkolker{at}umich.edu
Crowns and large amalgams protect structurally compromised teeth to various degrees in different situations. The aim of this investigation was to evaluate the survival of teeth with these two types of restorations and the factors associated with better outcomes. Retrospective administrative and chart data were used. Survival was defined and modeled as: (1) receipt of no treatment and (2) receipt of no catastrophic treatment over five- and 10-year periods. Analyses included: Kaplan-Meier survival curves, Log-Rank tests, and Cox proportional hazards regression modeling. Crowns survived longer with no treatment and with no catastrophic treatment; however, mandibular large amalgams were least likely to have survived with no treatment, and maxillary large amalgams were least likely to have survived with no catastrophic treatment. Having no adjacent teeth also decreased survival. Crowns survived longer than large amalgams, but factors such as arch type and the presence of adjacent teeth contributed to the survival of large amalgams.
Key Words: dental amalgam • crowns • longitudinal study • survival analysis • treatment outcome
To improve clinical decision-making, dentists and patients must have evidence-based information on factors that produce the best long-term treatment outcomes. Treatment of a tooth that has lost a significant amount of tooth structure is a common clinical situation. Common procedures include the placement of either a large amalgam (replacing 4 or 5 tooth surfaces) or a large amalgam followed by a crown (the large amalgam serves as a foundation for the crown). The latter will be referred to in this article as a ‘crown’. While several studies have evaluated the clinical performance of large amalgams and crowns, there remains uncertainty about which procedure provides the best long-term outcome for a structurally compromised tooth (Leempoel et al., 1985; Walton et al., 1986; Robbins and Summitt, 1988; Letzel et al., 1989; Hawthorne and Smales, 1997; Martin and Bader, 1997; Smales and Hawthorne, 1997; Plasmans et al., 1998; Smales, 1991). Published studies are often difficult to compare, since they rarely compare the two procedures in the same study and can vary in protocol, methods, and the definition of outcome measures. This inadequate evidence base has contributed to variations in the treatment planning for these two procedures (Bailit and Clive, 1981; Elderton and Nuttall, 1983; Grembowski et al., 1990a,b; Shugars and Bader, 1992; Bader et al., 1993; Bader and Shugars, 1993, 1995; Shugars et al., 1997). Several recent studies have used administrative data to evaluate the survival of large amalgam and crown restorations. In one study, teeth with a crown were significantly more likely to ‘survive’ over time compared with teeth restored with a large amalgam (Smales and Hawthorne, 1997). In another study, about 70% of four- and five-surface amalgams were ‘successful’ (needing no further treatment or only an additional one- or two-surface restoration) after 5 yrs, with 10–15% experiencing a catastrophic failure (i.e., endodontic treatment or extraction) (Martin and Bader, 1997). In contrast, 84% of crowns were successful, with about 10% having a catastrophic procedure. In a previous study, we followed teeth restored with a large amalgam or a crown for 5 and 10 yrs in routine patients to gain an understanding of the type of treatment received (Kolker et al., 2004). We created Treatment Outcome Trees as a descriptive approach to evaluate the natural history of teeth treated with a large amalgam or crown. Sixty-four percent of large amalgams and 22% of crowns later received further treatment over the ensuing 10 yrs. Even with these descriptive analyses, however, we were unable to determine how quickly the teeth needed subsequent treatment, the complexity of the procedures received, or the factors relating to the type of treatment received. The purpose of this study was to follow and compare large amalgams and crowns, and determine (a) the survival rates for teeth with these two types of restorations and (b) factors associated with better outcomes.
This study was approved by the Committee for the Protection of Human Subjects (Institutional Review Board) at the University of Iowa. We used retrospective data from The University of Iowa, College of Dentistry, administrative database and patient records to follow 737 ‘target teeth’ that received a large amalgam restoration (4 or 5 surfaces) in 1987 or 1988. ‘Target teeth’ that received a subsequent crown on or before 365 days from the placement of the large amalgam were classified as crowns; the others were classified as large amalgams. All permanent posterior teeth were included except the following: (1) third molars, (2) teeth identified as having received root canal therapy prior to the placement of the large amalgam or crown restoration, or (3) teeth that served as a bridge abutment. If a patient had more than one target tooth, we randomly selected one tooth for analysis, to avoid related observations. Information from every dental visit related to the ‘target teeth’ during the ten-year period was collected. Survival analyses were completed, and related factors were modeled for two different definitions of survival: (Model 1) receipt of no treatment, and (Model 2) receipt of no catastrophic treatment (not having extraction or root canal procedures). We used the Kaplan-Meier Product Limit method to estimate survival probability curves for each level of a covariate. We used Log-Rank tests to assess the association between potential risk factors and survival time. We used Cox proportional hazards regression to model the association between exposure (restoration type) and time with no treatment and no catastrophic treatment while controlling for confounding risk factors (covariates). We conducted both forward and backward step-wise approaches to create the most parsimonious models. We performed forward step-wise modeling by placing each variable that had a log-rank p < 0.20 individually into the proportional hazards model. The variable with the smallest p-value was first put into a proportional hazards model with restoration type, and then the remaining variables were evaluated individually. The variable that displayed the lowest p-value was then included in the model. This process was repeated until no additional variable maintained a p < 0.05. We performed backward step-wise modeling by including all the variables with a log-rank p < 0.20 in the proportional hazards model. The variable with the largest p-value exceeding the 0.05 level was dropped, and a proportional hazards model with the remaining variables was created. From this second model, the variable with the highest p > 0.50 was dropped, and a reduced proportional hazards model was created. We placed the first dropped variable back into the model to verify its exclusion. This process was followed until all variables with a p > 0.05 were excluded from the model. Possible interaction terms were then tested between the restoration type and the other covariates that remained in the model. The proportional hazards assumption was tested by the inclusion of covariates by log-time interaction terms.
For each level of the covariates, the five- and 10-year survival estimates (Kaplan-Meier) and Log-Rank p-values for both survival Models 1 and 2 are presented (Table 1  ). Crowns were significantly more likely to have survived with no treatment than were large amalgams. Large amalgams had a survival estimate of 0.52 for receipt of no treatment after 5 yrs (48% of large amalgams received some treatment after 5 yrs). The survival estimate for large amalgams receiving no treatment after 10 yrs was 0.46. This compares with a survival estimate of 0.88 for crowns receiving no treatment at 5 yrs and 0.77 at 10 yrs. The survival estimates for large amalgams receiving no catastrophic treatment was 0.87 after 5 yrs and 0.78 after 10 yrs, compared with 0.95 for crowns at 5 yrs and 0.87 at 10 yrs.
The Log-Rank tests comparing restoration type were highly significant, whether the outcome was no treatment or no catastrophic treatment (< 0.0001 and 0.0011, respectively). A Log-Rank p-value < 0.20 was obtained for several other variables (see Table 1  ), which became candidate predictors in survival modeling. Variables that had a p-value > 0.20 (and therefore were not included in the modeling) were: tooth type (molar vs. premolar), diabetes, patient-reported gum disease, grinding or clenching, interest in keeping teeth, number of pins, type of provider, and presence of a prior restoration.
The final models for time to some treatment and time to catastrophic treatment, along with the parameter estimates (β) and Hazard ratios, were determined (Table 2
In Model 1, crowns were more likely than large amalgams to survive with no treatment, and teeth in the maxillary arch were more likely than mandibular teeth to survive with no treatment (Table 3 ). The interaction pattern for restoration type and arch type was consistent in direction. Mandibular large amalgams experienced any treatment at a rate over 6 times (1/0.16 = 6.25) that of maxillary crowns, adjusted for the number of proximal contacts. Additionally, teeth without proximal contacts had any treatment at nearly 2 times the rate of teeth with one or two proximal contacts.
Kaplan-Meier survival curves for the receipt of no treatment were created (Fig.
In Model 2, crowns were more likely to survive with no catastrophic treatment than were large amalgams (see Table 3 ). For no catastrophic treatment, the interaction between restoration type and arch type differed in magnitude and had a reversal of effect. Maxillary crowns were more likely to survive with no catastrophic treatment than were mandibular crowns; however, mandibular large amalgams were more likely to survive with no catastrophic treatment than were maxillary large amalgams. With adjustment for number of proximal contacts and gender, mandibular large amalgams experienced catastrophic treatment at a rate 2 (1/0.48 = 2.08) times that of maxillary crowns. Teeth without proximal contacts received catastrophic treatment at twice the rate of teeth with one or two proximal contacts, and men experienced catastrophic treatment at a rate 1.4 (1/0.7 = 1.42) times that of women. At 10 yrs, 90% of maxillary crown teeth survived with no catastrophic treatment, compared with 83% of mandibular crowns. The catastrophic treatment survival estimates for large amalgams by arch were quite similar at 5 yrs (0.81 for maxillary and 0.83 for mandibular), while at 10 yrs, the survival estimates varied more (0.68 for maxillary, 0.76 for mandibular) (Fig., Panel B).
Teeth in which a crown was placed over a large amalgam survived longer than teeth with a large amalgam alone. The presence of a proximal tooth was protective, and women were more likely to have survived without receiving catastrophic treatment. The interaction effect of arch with the type of restoration was significant but differed for the two models. Many of our findings were very similar to those of Martin and Bader (1997); however, our survival time for teeth with large amalgams was shorter. At 5 yrs, we found that 52% of large amalgam teeth survived with no treatment, while they found that 62% of four-surface amalgams and 55% of five-surface amalgams survived (Martin and Bader, 1997). Our findings for crowns were higher, with 88% of crowns surviving with no treatment, compared with 80% of crowns in their study. Some of these differences may be due to differences in practice settings (i.e., managed care group vs. dental school). In the literature on outcomes for large amalgams and crowns, the number of proximal contacts has not been reported. However, in a study of endodontically treated teeth, teeth with two proximal contacts had the best survival estimates (Aquilino and Caplan, 2002). The same investigative team found that teeth with 0 or 1 proximal contact at endodontic access were lost three times as often as teeth with two proximal contacts (Caplan and Weintraub, 1997; Caplan et al., 2002). Caplan and co-workers suggested that neighboring teeth may help distribute occlusal forces (Caplan et al., 2002). Teeth that have 0 or 1 proximal contact may be more likely to be present in individuals with fewer teeth and a poorer oral health status (and perhaps fewer financial resources), but this is speculative. Few studies have examined the differences in restoration survival by arch. Some authors reported no differences in outcomes between the maxillary and mandibular arches (Drake et al., 1990; Köhler et al., 2000). One study reported lower restoration survival rates in the mandibular arch (McDaniel et al., 2000). Findings of restoration performance by patient gender have been mixed, with some showing no difference (Dawson and Smales, 1992; Dunne and Millar, 1993; Mjör et al., 2000), and one study finding that survival was superior in female patients (Mahmood and Smales, 1994). Women access dental care differently and react to health promotion in a more positive manner than do men (Zakrzewska, 1996) and are known to access more medical and dental health care and to seek care more frequently (Woolfolk et al., 1999). This may act as a preventive measure, increasing the longevity of restorations through more frequent routine maintenance. The receipt of less treatment after 10 yrs in crowns could be related to decisions by dentists and patients regarding which teeth should receive crowns. Dentists in this study may have been more likely to ‘treatment plan’ simply a large amalgam for teeth with a questionable prognosis, from a perceived cost-benefit perspective. The difference in outcomes between the two treatment options could thus represent appropriate treatment planning and case selection, so that teeth with a better prognosis receive the more expensive and longer-lasting restoration. The longer survival of crowns should be interpreted carefully, since it relates to less ‘need’ for treatment. Since this was an observational study, the findings describe the actual care provided, which differs conceptually from the care that was ‘needed’, as measured against a clinical gold standard. Dental practice pattern variations exist and are also expected to exist among practitioners in a dental school. Treatment planning is still driven by the individual dentist (and, in this case, instructor) and patient preferences. Continued research to develop an evidence base for the appropriate provision of routine services is essential. Since these data were collected from a group of patients seen at a dental school, generalization of the findings beyond this population should be made cautiously. Other information that would have strengthened the findings, but was not available, included the reason for the initial placement of a large amalgam or crown, the practitioner’s long-term prognosis for the tooth/restoration, the level of insurance coverage for crowns and/or personal income information, periodontal status, and additional information about the oral health status of the patient. A limitation with the use of only administrative and chart data was our inability to identify treatment provided outside the dental school. By survival analysis, the data were censored at the patient’s last visit to the school. Treatment received outside the College when they were also receiving treatment at the College is missing. The findings do suggest that it is appropriate to consider patient and oral health characteristics in the clinical decision-making process when faced with a tooth that has lost a significant amount of structure. This study provides information that can be used to facilitate dentist and patient education regarding factors to consider when choosing between a large amalgam and a crown. To assist the clinical decision-making process, we plan future research to consider the cost-effectiveness of the two different restorative options.
Research was completed while Dr. Kolker was a doctoral student at The University of Iowa, during which she was a Research Fellow at the Iowa City Veterans Affairs Medical Center (2001–2003) and a NIDCR-K16 Dentist-Scientist Award Recipient-Project # DEOO175 (1996–2001). This report was based on a dissertation submitted to the graduate school, University of Iowa, in partial fulfillment of the requirements for the PhD degree. Received for publication November 11, 2003. Revision received July 1, 2004. Accepted for publication August 23, 2004.
Journal of Dental Research, Vol. 83, No. 11,
854-858 (2004) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
