INTRODUCTION

Margin discoloration has generally been considered a sign of early failure in composite resin restorations and is often used as a criterion for restoration replacement.¹ The assumption is that discoloration is associated with debonding and fluid penetration into the interfacial space.^2,3 The ingress of cariogenic fluid associated with plaque retention has been specified as a potential cause of recurrent caries.⁴ There is little clinical evidence in the recent literature to support this cause-and-effect relationship in vivo.⁵

The seal of a cavity margin has been investigated for many years, especially with the development of newer bonding agents and adhesive materials. Much of the published research has been based on in vitro microleakage studies, which involve assessing the penetration of tracer solutions into the interface. After a restored extracted tooth is exposed to a tracer solution, usually in conjunction with an external stimulus, it is sectioned through at least one margin interface. Dark-colored dyes, such as methylene blue^6,7 and basic fuchsin,^8,9 as well as silver nitrate,^9,10 are tracers that can readily be identified on tooth cross sections. The degree of penetration along the interface is assessed using either a rating scale^9-12 or measuring software.^13,14 To simulate clinical performance, external stimuli are often applied, such as thermal^8,10,15,16 and/or mechanical cycling.^17,18

There have been attempts to correlate these in vitro tests with clinical performance, but they have been largely unsuccessful at establishing a direct relationship.^19,20 It is possible that a relationship could be established if studies were restricted to a highly caries active population and margin discoloration was considered as one contributing factor. However, in such a population, there are also many factors that contribute to caries development, such as genetics, diet, oral hygiene compliance, education, socioeconomics, and chronic medical conditions that require medication.

This lack of cause-and-effect association has led to the concept of restoration repair as an alternative treatment to restoration replacement.²¹ A number of studies have documented short-term clinical success with restoration repair procedures, which is less invasive, more conservative of adjacent tooth structure, and less costly to provide.^22-24 Margin debonding and discoloration usually occur in a localized area of the exposed margin and are often associated with a small defect or discontinuity at the interface, frequently created through wear, occlusal forces, or inadequate initial adaptation of the restorative material. With adequate access and isolation, margin repair should be a treatment of choice, especially in moderate- to low-risk populations. It has also been shown by Hamilton and others²⁵ in a randomized trial that the progression of diagnosed pit and fissure incipient caries can be very slow and that observation may be another viable alternative.

The purpose of this study was to morphologically characterize the defective margin area associated with discoloration in teeth where there is no evidence of active recurrent caries and to conduct a randomized clinical trial to assess the effectiveness of resealing the margin with a flowable composite resin.

METHODS AND MATERIALS

The first phase of the study was a single-variable observational design, and the second phase was a randomized clinical trial (Figure 1). A total of 152 patients were recruited from the clinics at the School of Dentistry as they appeared for routine recall and/or restorative treatment. The baseline clinical and morphological data collected at the initiation of the study included 360 restorations that met the specified inclusion criteria. The study protocol and consent form were approved by the Health Sciences Institutional Review Board at the university. Each patient who participated was given a full description of the study and voluntarily signed a patient consent form.

View Figure

• Download as Powerpoint
• Download Figure

The inclusion criteria required that each patient have at least one and no more than three existing composite restorations with visual evidence of discoloration on an accessible margin but no diagnosis of recurrent caries. The criteria for recurrent caries were one of the following: 1) a margin crevice with tactile softness to an explorer, 2) white surface demineralization with cavitation, or 3) undermining opacity in enamel or color change within dentin. All margins had to be located at the gingival crest or above; the patient had to be able to perform normal hygiene procedures and be available for five yearly recalls.

Evaluation Procedures

For each restored tooth that was enrolled in the study with a defective margin, the following clinical measurements and evaluations were made at the baseline appointment:

• The observed color of the defect was recorded, and the length and width of the discolored area was measured to the nearest 0.5 mm with a Goldman-Fox periodontal probe (Hu-Friedy Mfg Co, Chicago, IL, USA) marked in 1-mm increments.

• Each restoration was evaluated independently by two trained examiners using modified USPHS criteria^26,27 for color match, margin discoloration, margin adaptation, and recurrent caries, the primary outcome for the study (Table 1).

• A microscopic evaluation of margin defects was made at 20× magnification using a surgical microscope (Urban, Global Surgical Corp, St Louis, MO, USA). A visual description of the location of the discoloration within the margin and the microstructure of the defective margin were recorded, as viewed directly through the microscope and evaluated according to the criteria in Table 2.

Table 1 Modified Ryge Criteria for Evaluation of Tooth-Colored Restorations

View Fullscreen

Table 2 Criteria Used to Describe the Magnified Margins

View Fullscreen

RESULTS

Morphologic Characterization (Clinical Microscope at 20×)

Complete baseline, three-year, and five-year data for stain location at the margin interface, as viewed directly through a surgical microscope, are shown in Table 3. All evaluations were made by direct vision through the microscope eyepiece. Randomization of the baseline assignments was very effective in producing almost identical location distributions in both treatment and control groups. At one year, 153 controls continued to exhibit staining in one location (92.7%), while the resealed group was skewed toward “no discoloration” (107 restorations, 65.6%). Similar patterns were demonstrated in the resealed group at both three years (69 restorations, 50.4%) and five years (52 restorations, 47.3%). At five years, the stain was more in the composite than in the interface (32.7% vs 18.2%) for the resealed group, while the reverse was true of the control group (30% vs 49%). Only 12 (12%) of the untreated control restorations and two (1.8%) of the resealed restorations evaluated showed evidence of stain in the tooth structure, which may explain the lower incidence of recurrent caries over time in these defective margins. After five years, 79% of the control restorations and 50.9% of the resealed restorations had stain accumulation in either the interface or in the composite, which did not appear to have a demineralizing effect on adjacent tooth structure.

Table 3 Microscopic Stain Location Within the Restoration Margins at Each Recall Period (Number of Restorations)^a

View Fullscreen

For margin contour, as evaluated microscopically (Table 4), 33% of the restorations in each group had overlapped resin margins at baseline, while another 50% to 53% had crevices at the margin interfaces. The general distribution remained the same for the control group (33% overlap; 49% crevice) after one year; however, 75 (46%) of the resealed restorations evaluated had continuous margins and only 22 (13%) exhibited crevices. At three years and even more so at five years, margin defects were more evident and the distribution of the defects in treated teeth approached those of the untreated controls (31% vs 40% overlap; 49% vs 33% crevices at five years). Shallow crevices did not appear to progress to deeper crevices over time and were similar in both treated (23%) and untreated (37%) groups at five years. Significant patient loss occurred with time, as the population was more focused in the elderly. During the five years, there were 36 teeth in the control group, in which restorations were lost to recall; seven were replaced with crowns, bridges, or veneers; eight were debonded or missing; 15 were replaced inadvertently in the predoctoral clinic; and six were replaced due to caries. For the treatment group, 25 restorations were lost to recall; five were replaced with crowns, bridges, or veneers; 10 were debonded; five were replaced inadvertently; and five were restored due to caries. These were reported at various yearly intervals as patients appeared for recall appointments.

Table 4 Microscopic Contour of the Restoration Margin at Each Recall Period (Number of Restorations)^a

View Fullscreen

Clinical Evaluation

Clinical evaluations were made for each restoration using the modified Ryge criteria at baseline, posttreatment, and at each yearly recall.²⁶ The four major characteristics evaluated were color match, margin discoloration, margin adaptation, and recurrent caries. The data for color match did not reveal any noticeable change in color over the five years, and the resealed areas were not readily visible from the original restoration by color change.

For margin discoloration at baseline (Table 5), approximately 77% to 82% in each group showed penetrating stain. About one-third of these restorations showed penetrating stain along greater than 50% of the visible margin on the exposed surface. Since discoloration was a criterion for inclusion, there were no restorations without stain, and 23% showed only surface stain at the interface. At posttreatment (TxP), 98% of restorations were rated without discoloration or with a slight surface translucency. For the restorations evaluated after one year, 116 restorations (71%) in the treatment group showed no evidence of discoloration, as opposed to 12 restorations (7%) in the control group. At each recall period, there was still a significantly greater number without discoloration in the treatment group: 57% vs 8% at three years and 49% vs 10% after five years. Figure 2 shows the contrast between the rates at each yearly recall and that the rate of recurrence increases steadily with time. At some future time, it can be expected that the margin discoloration in both groups will be similar if the trends continue. Figure 3 shows a sequence of photographs with recurring margin discoloration after five years.

Table 5 Clinical Evaluation of Margin Discoloration at Each Recall Period (Number of Restorations) ^a

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

For margin adaptation (Table 6), the distribution of defects was also similar for the treatment and control groups at baseline. Approximately 48% of each group had a one-way margin catch, and 41% to 47 % had a detectable crevice. Less than 11% of these margins had crevices that extended along more than 50% of the exposed margin. At posttreatment examination (TxP), 160 (90%) of the resealed margins were rated as barely detectable, 17 (9%) as readily detectable, and three were dislodged during the treatment. There were no margin crevices after retreatment. After one year, 80 resealed restorations (49%) were barely detectable; 72 (44%) had a definite one-way catch and only 11 (7%) had a localized margin crevice. In comparison with the control group, the effective reduction in margin crevices with resealing was 26% after one year. Five years after retreatment, 30 margins (27%) were barely detectable; 68 (62%) were readily detectable and 12 (11%) showed crevicing. In contrast, the untreated restorations had 68% with readily detectable margins and 21% with crevices.

Table 6 Clinical Evaluation of Margin Adaptation at Each Recall Period (Number of Restorations) ^a

View Fullscreen

The primary outcome of the study was to determine the failure rate of repaired restorations due to recurrent caries. The results in Table 7 show that the cumulative incidence of recurrent caries over five years was only 4.4% without retreatment and 3.7% with retreatment. In both groups, only 7% to 8% of the restorations were totally debonded or lost. Using a chi-square test at p<0.05, these losses were not significantly different with or without the resealing treatment. Figure 4 shows a baseline and five-year photograph of a typical lesion that was observed over the five yearly recalls without treatment.

Table 7 Clinical Outcomes After Five Years (Number of Restorations) ^a

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

In reviewing the literature at the time the study was proposed, the anticipated incidence of recurrent caries around the discolored margin of a composite restoration was about 30%.^28,29 In recent years, there has been increasing support to repair rather than to replace composite restorations with defective margins, assuming the restoration is otherwise in satisfactory condition.^21,22 In developing a restorative treatment plan for a patient with composite restorations that have discolored margins, it is a difficult decision whether to suggest continued observation, repair, resealing, or full replacement. Many risk factors also enter into this decision, even if the diagnosis of active recurrent caries is not justified, such as age of the restoration, oral hygiene of the patient, dental restoration history, fluoride application history, and operative access to the discolored margin. Given these variables, this fully randomized clinical trial was initiated to gain evidence to support two of the treatment suggestions.

One interesting parameter which might give a suggestion as to which type of discoloration would be most likely to develop a margin lesion, given a characteristic defect, is the location of the stain along the interface. During the baseline evaluation, a surgical microscope was used to evaluate the location of the stain and the margin configuration at the point of staining. Randomization of patients as they were recruited proved to equally distribute the characteristics between the treatment and control groups for all of the criteria studied (Tables 3-6). In looking at the baseline distribution, it seems likely that stain in the composite material and stain from accumulated material lodged in the interface would be less likely to develop into recurrent caries than stain that has affected adjacent tooth structure. Such stain may be an indication that a wall lesion is beginning to form.^18,30-32 The sample size in this group was small (10% to 14%); however, in evaluating the 18 restorations that showed color in the tooth side of the interface, 13 were sound at five years, three had been replaced for noncarious reasons after being noncarious at two years, one did not attend any recalls, and only one became active caries at the three-year recall. Therefore, this factor was not as good a predictor for recurrent caries development as was anticipated.

Another factor that has been indicated as a contributing cause for recurrent caries is crevice formation at the interface.^29,33 It is assumed that a margin crevice creates a hidden area for bacteria to accumulate and incubate to the point of causing demineralization. This may be true, but it was a very slow process in this population. Of the 11 restorations that were diagnosed with active recurrent caries through the five years, five were diagnosed at two years, two at three years, two at four years, and two at five years. Eight of these restorations (73%) had crevices microscopically at baseline; six were in the control group and two were in the treatment group. However, that represents only 6% of the total restorations in the control group that had microscopic crevices at baseline and later developed caries during observation. Therefore, the development of a margin crevice also seems to be a relatively weak predictor of future recurrent caries.

From Table 5 and Figures 2 and 3, margin discoloration did reappear in repaired restorations over the five years of this study. The important factor is the time period curve and the extent of this recurrence interval. After five years, recurrence of the staining was still only at 46% as opposed to a consistent 80% for the untreated control restorations, an effective reduction of 35%. The incremental recurrence was 18% the first year, 9% additional at the second year, 9% at the third year, 7% at the fourth year, and 3% at the fifth-year recall. The yearly incremental recurrence may have decreased with time because of the patient characteristics of those returning for recall, the effect of increased awareness on oral hygiene, the effectiveness of the resin bond at the margin with a flowable material, or the location of the stained margin on the restoration. There was no comparison made to see if the recurrent stain was in the same location on the restoration margin or within the margin as the original one that was repaired. An assumption can also be made that the rate of recurrence diminishes over time in a population like this.

When evaluating the quality of restorations returning for recall over a long period of time, there also has to be consideration of the restorations that were lost in returning patients and those that did not return at all for recall. In the retreatment group, there were four restorations (3%) lost and 10 (6%) not available at one year, 13 (9%) lost and 27 (15%) not available at three years, and 25 (19%) lost and 42 (24%) not available after five years. It is impossible to say whether the incidence of returning stain for the recalled group is representative of those unevaluated restorations or whether their incidence would have been greater.

Another variable that should be considered is the study population. No attempt was made to formally evaluate each patient's individual risk, but every patient had a restorative history with existing restorations to qualify for the study. The estimated caries risk for the group was low to moderate in most cases, applying criteria published by Fontana and Zero.³⁴ The patient pool at the dental clinic tends to be from a somewhat lower socioeconomic status. The water in the area is fluoridated, and the patients recruited were attending the clinic for recall maintenance or routine restorative treatment. All were educated in oral hygiene procedures and were concerned about their oral health. Had the study been conducted in a population with high caries risk, the recurrence of margin staining and the incidence of recurrent caries might have been greater. Then the resealing technique would probably have been less effective.

The incidence of recurrent caries was less than 5% for both groups over the five years, and there was not a significant difference between the two groups. This indicates that there was no improvement in restoration longevity by resealing the margins of stained restorations as opposed to continued observation over five years. This conclusion is in agreement with previous studies, which have shown a positive result from restoration repair,^33,35 but there is no documentation for a group undergoing continued observation at yearly recalls without any intervention. If resealing stained restoration margins does not improve longevity, certainly full restoration removal and replacement are contraindicated, except to improve esthetics. Based on these data, it is reasonable to conclude that staining of a composite restoration margin does not lead to the development of recurrent caries in adjacent tooth structure.

CONCLUSIONS

Under the conditions stated for this study of discolored composite restoration margins over five years, the following conclusions can be made:

1. The incidence of recurrent caries was very low and similar for restorations with discolored margins that were either resealed or just observed over five years.

2. Neither the depth of discoloration nor the location of the stain within a restoration margin were valid predictors of subsequent recurrent caries over the same period.

3. Penetrating margin stain was removed by resealing but returned in a significant number of restorations over the five years.

4. Noncarious discolored or defective composite margins are not a definitive indication for restorative intervention.