Association and comparison between visual inspection and bitewing radiography for the detection of recurrent dental caries under restorations

Abstract

Purpose: The aim of the present study was to investigate, in posterior teeth, the association between the characteristics of the margins of a restoration visually inspected and the presence, under restorations, of recurrent or residual dental caries detected by radiographic examination. Furthermore, the agreement between visual inspection and radiographs to detect dental caries was assessed. Methods: Eighty-five permanent molars and premolars with resin restorations on the interproximal and/or occlusal faces, from 18 patients, were submitted for visual inspection and radiographic examination. The visual inspection involved the criteria of the International Caries Detection and Assessment System (ICDAS). Bitewing radiographs were used for the radiographic examination. Logistic regression was used to analyse the association between the characteristics of the margins of a restoration assessed by visual inspection (absence of dental caries, or early, established, inactive and active lesions) and the presence of recurrent caries detected by radiographs. Kappa coefficients were calculated for determining agreement between the two methods. Results: The Kappa coefficient for agreement between visual inspection and radiographic examination was 0.19. Established lesions [odds ratio (OR) = 9.89; 95% confidence interval (95% CI): 2.94–33.25; P < 0.05] and lesion activity (OR = 2.57; 95% CI: 0.91–7.27; P < 0.05) detected by visual inspection, were associated with recurrent or residual dental caries detected by radiographs. Restorations with established and active lesions at the margins had a greater chance of exhibiting recurrent or residual lesions in the radiographic examination. Clinical significance: The present findings demonstrate that restorations with established and active lesions at the margins when visually inspected often require removal and retreatment.

INTRODUCTION

The development of recurrent dental caries is one of the main reasons for having to replace a dental restoration1., 2.. The early detection of such lesions is fundamental in preventing further tooth decay and establishing a favourable prognosis3., 4..

In vitro studies have compared radiographic examinations with other imaging techniques and with visual inspection using the criteria of the International Caries Detection and Assessment System (ICDAS) for the evaluation of dental caries. One study reported that the ICDAS is more reliable and accurate than radiography for detecting and estimating the depth of lesions in both primary and permanent teeth when proximal lesions are directly inspected⁵. Another study also reported the better performance of a visual inspection in comparison with laser fluorescence, a fluorescence-based camera and bitewing radiographs for the detection and treatment decision of occlusal caries⁶. Similar results have been reported for occlusal dental caries in the comparison of a visual inspection with radiographic examination⁷.

Despite the reliability of a visual inspection for the detection of dental caries, dentists generally request a complementary examination for confirmation8., 9., 10.. This combination of detection methods is particularly important for the evaluation of recurrent caries, as the restorative material can hinder the visual inspection. A radiographic examination with photographic film or a digital detector is often used for the detection of recurrent dental caries. A recent in vitro study compared intra-oral images and two cone beam computed tomography (CBCT) systems for the detection of artificial buccal recurrent caries under restorations¹¹. Although the authors found that CBCT outperformed intra-oral radiography, expensive imaging examinations may not be widely available in many populations.

The aim of the present study was to investigate, in posterior teeth, the association between the characteristics of margins of a restoration visually inspected and the presence of recurrent or residual dental caries under restorations detected by radiographic examination. Furthermore, agreement was assessed between visual inspection and radiographs for the detection of dental caries.

MATERIAL AND METHODS

This study received approval from the Human Research Ethics Committee of Universidade Vale do Rio Verde (Brazil) under process number 347.980 and was conducted in full accordance with the Declaration of Helsinki of the World Medical Association. All participants received dental treatment at the School of Dentistry of the University and provided written informed consent.

Sample

Eighteen individuals undergoing dental treatment at the dental clinic of the Universidade Vale do Rio Verde (Brazil) were included in the study, providing a total of 85 filled teeth. Based on a pilot study involving 16 teeth, it was determined that a sample of 73 teeth would be sufficient to detect 95% of recurrent dental caries through visual inspection, considering a 5% precision level and 95% confidence interval (95% CI).

Eligibility criteria

Only premolars and permanent molars with resin restorations were included in the study. Teeth with severe dental caries, restorations on the buccal or lingual faces and those with orthodontic bands were excluded from the analysis.

Data collection – visual inspection

After prophylaxis, the clinical oral examination was conducted at a dental office using artificial light, compressed air drying, a mouth mirror and a probe. Teeth with resin restorations on the proximal (mesial and distal) and/or occlusal faces were inspected using the ICDAS criteria¹², by which the stages of dental caries are coded. ICDAS code 1 corresponds to the first visual change in the enamel; code 2 is a distinct visual change in enamel; code 3 is localised enamel breakdown; code 4 is underlying dentin shadow; code 5 is distinct cavity within visible dentin; and code 6 is extensive cavity within visible dentin. The absence (code 0) or presence of dental caries was determined in the tissue adjacent to the restoration. The presence of dental caries was categorised into early decay (codes 1 and 2), established decay (codes 3 and 4) and severe decay (codes 5 and 6). However, severe decay was excluded from the analysis in the present study. Moreover, activity of the carious lesion was determined based on visual appearance, local susceptibility to plaque build-up and surface texture¹³. Lesions not linked to a restoration were not considered in the analysis. The examinations were performed by a dentist who had undergone a training and calibration exercise and had achieved Kappa coefficients of 0.70 for interexaminer agreement (comparison with a gold-standard expert in cariology) and 0.85 for intra-examiner agreement. This exercise was performed on a sample of 28 extracted teeth with resin restorations.

Data collection – radiographic examination

After the visual inspection, digital bitewing radiographs were taken using a 0.4-second exposure time. The images were processed using a digital system (Sensor XDR Radiology, Olympic Blvd, Los Angeles, CA), printed on films (DRYSTAR, Agfa Helthcare, Mortsel, Belgium). The teeth were evaluated by an expert in dental radiology who had no awareness of the results of the visual inspection. The presence/absence of a radiolucent image underlying the restoration and compatible with dental caries was considered in this examination. Before the examination, the dentist underwent an intra-examiner calibration process and achieved a Kappa coefficient of > 0.90. In this process, a member of the research team selected 32 bitewing radiographs and coded them in envelopes. The radiographs were images of permanent premolars or molars restored with resin on the proximal, distal or occlusal face. Some teeth had carious lesions under the restoration and others did not. During the calibration exercise, a member of the research team removed the radiograph so that the examiner was unaware of the code. The examiner’s diagnoses at the two evaluations, separated by a 2-week interval, were sealed in the envelope together with the radiograph.

Data analysis

Statistical analysis was performed using the SPSS 20.0 program for Windows (SPSS Inc., Chicago, IL, USA). Initially, the descriptive analysis (frequency) and Kappa agreement were performed. A logistic regression model was used to assess the association between the presence of recurrent dental caries in the radiographic examination (dependent variable) and the status of the region adjacent to the restoration during the visual inspection. The absence of dental caries (sound tissue), stage of dental caries (early or established) and caries activity (inactive or active) were the independent variables.

RESULTS

During the visual inspection, 53 (62.3%) teeth exhibited carious lesions associated with the restorations, and the distribution of early and established lesions was almost equal (n = 26 and n = 27, respectively). Most teeth with active caries exhibited enamel breakdown and/or an underlying dark shadow from the dentin (established lesions). On radiographic examination, 25 (29.4%) teeth had dental caries under the restorations.

A large number of teeth with carious lesions detected during the visual inspection demonstrated no radiolucency in the radiographic examination. Moreover, 15.6% of the teeth with radiolucency exhibited no signs of dental caries during the visual inspection. Therefore, the Kappa value was low (Table 1).

Table 1.

Agreement between visual inspection and radiographic examination for detection of dental caries

Visual inspection	Radiographic examination		Total n (%)	Kappa value
	Caries free n (%)	With caries n (%)
Caries free	27 (84.4)	5 (15.6)	32 (100.0)
With caries	33 (62.3)	20 (37.7)	53 (100.0)	0.19

The presence of established lesions [odds ratio (OR) = 9.89; 95% CI: 2.94–33.25; P < 0.05] and lesion activity (OR = 2.57; 95% CI: 0.91–7.27; P < 0.05), detected by visual inspection, were associated with recurrent dental caries detected by radiographs (Table 2).

Table 2.

Results of logistic regression and percentage agreement between visual inspection and radiographic examination according to lesion stage and activity

Visual inspection	Radiograph examination		OR (95% CI)
	Caries free n (%)	With caries n (%)
Dental caries stage
Early decay	22 (84.6)	4 (15.4)	1.81 (0.42–7.71)
Established decay	11 (40.7)	16 (59.3)	9.89 (2.94–33.25)*
Dental caries activity
Inactive	26 (83.9)	5 (16.1)	0.98 (0.23–4.21)
Active	7 (31.8)	15 (68.2)	2.57 (0.91–7.27)*

^*P < 0.05.

DISCUSSION

Most dental caries detected during the visual inspection did not present a radiolucent image under the restoration that was suggestive of recurrent or residual caries. Thus, considering the current emphasis on less invasive treatment, dental caries restricted to the margins of restorations could be repaired without replacement of the existing restorative material. Nonetheless, studies indicate that the presence of caries at the margin is one of the main reasons for replacing restorations2., 14..

A low level of agreement was found between visual inspection and the radiographic examination in the present study. This finding stems from the high number of early and inactive lesions at the margins of the restorations that were detected during the visual examination, which did not present carious lesions under the restorations during the radiographic examination. On the other hand, most teeth with a radiolucent image under the restorations had dental caries (early and established, active or inactive lesions) at the margin of the restorative material that was detected during the visual examination. However, the presence of an inactive and early lesion at the margin of the restoration was not associated with a radiolucent image under the restoration. These results are in agreement with a study which found that only a frankly carious lesion at the margin of the filling constituted a reliable diagnosis of secondary caries¹⁵. These findings underscore the importance of a complementary radiographic examination for the detection of recurrent dental caries, especially in the case of established and active lesions. Although a conventional radiographic examination is inferior to CBCT for the detection of artificial recurrent caries¹¹, it remains the complementary imaging examination employed most often for the diagnosis.

In the present study, it was not possible to state that the radiolucency under a restoration was caused by caries. This may be because of the presence of demineralised dentin when the restoration was performed. However, a recent in vitro study found no significant differences between teeth with or without demineralised dentin beneath restorations regarding their cervical marginal integrity, susceptibility to caries or microleakage of the placed restorations¹⁶. Thus, both in vitro and longitudinal studies are needed to determine whether these are only recurrent lesions in progression that are manifested in the region adjacent to the restoration. This is important because the presence of demineralised dentin does not indicate the need for intervention.

A histological examination is the gold standard for determining the extent to which tissue is affected by dental caries¹⁷. However, this method is not available in clinical practice. Fluorescence methods have also proven superior for the detection of residual dental caries during the preparation of the cavity18., 19.. In the present study, the decision was made to employ the most common types of examinations used by dentists for the detection of recurrent or residual caries under dental restorations. However, the limitations of conventional bitewing radiographs lead to a subjective diagnosis. Indeed, patients who change dentists have a larger number of restorations replaced in comparison with those who remain with their original dentist²⁰. It is therefore important to stress the need for clinical follow up before deciding to replace a restoration because the fate of recurrent or residual caries may be dependent on the sealing of the restoration. A study involving the sealing of gross occlusal caries suggests that residual caries may be arrested following restoration²¹, indicating that the presence of recurrent or residual caries does not seem to be a criterion for the replacement of a restoration²².

Studies addressing recurrent or residual dental caries stress the importance of comparing different diagnostic tools. However, we found no previous studies that tested the association between dental caries at the margin of a restoration and radiographic evidence of recurring or residual dental caries under the restoration. Moreover, the present study was limited by the absence of a gold standard for confirming the radiographic diagnosis and by failure to collect information on other clinical variables, such as marginal defects in the restoration and individual oral hygiene practices. Further studies are needed for a more robust sample calculation, considering the different stages of dental caries.

In conclusion, the occurrence of dental caries, especially established and active lesions, detected through visual inspection, allows the prediction of recurrent or residual caries under dental restorations.

Acknowledgements

This study was supported by the National Council for Scientific and Technological Development (CNPq), the Ministry of Science and Technology and the State of Minas Gerais Research Foundation (FAPEMIG), Brazil.

Disclosure of conflicts of interest

The authors declare no conflicts of interest.