Negligible therapeutic impact, false-positives, overdiagnosis and lead-time are the reasons why radiographs bring more harm than benefits in the caries diagnosis of preschool children

Abstract

Background: To evaluate the clinical course and interventions required during two years of follow-up of dental surfaces of deciduous molars diagnosed, and consequently treated, by two different strategies: diagnosis made by clinical examination alone or associated with radiographs.

Methods: This is a secondary analysis of a two-arm randomized clinical trial with parallel groups related to the diagnostic strategy for caries detection in preschool children. 216 children (3-6 years old) were followed-up for two years. All dental surfaces were diagnosed by visual inspection and later, through radiographic assessment. Baseline treatment was made in accordance with the results obtained by visual inspection performed alone or combined with radiographic method, considering the allocated group. Dental surfaces with no restoration needs, or those restored at the beginning of the study were followed-up for two years. The treatment decision was made according to the allocated group. The outcome was the occurrence of failure (a new caries lesion or a restoration replacement) during the follow-up.

Results: 4383 proximal and occlusal surfaces of deciduous molars in 216 preschool children were diagnosed and treated according to the abovementioned diagnostic strategies and followed-up for 24 months. The assessment of radiographs made change the initial decision reached by visual inspection in about 30% of the surfaces when all types of interventions were considered. However, most disagreements occurred for initial lesions, where radiographs tended to underestimate them. Discordances between methods occurred in less than 5% of all surfaces when considered lesions requiring operative treatment. For discrepancy cases, the placed interventions guided by following the radiographic results did not present less failures against those made following only visual inspection. As a matter of fact, the use of radiographs in the diagnostic strategy for caries detection in children brought more harms than benefits due to the occurrence of false-positives, overdiagnosis and lead-time bias.

Conclusions: Simultaneous association of visual inspection and radiographic assessment for caries detection in preschool children causes more harms than benefits, and therefore, visual inspection should be conducted alone in the regular clinical practice. Trial registration Clinicaltrials.gov platform: NCT02078453, registered on 5th March 2014.

Keywords: Children; Clinical trial; Dental caries; Diagnosis; Primary teeth; Radiography.

Fig. 1

Conceptual framework built to perform the multilevel logistic analysis to evaluate the influence of the results obtained with different diagnostic strategies on the occurrence of new operative treatments during the follow-up

Fig. 2

Decision tree (general view) considering caries lesions detection and subsequent treatment (no treatment, non-operative and operative treatment) performed first with visual inspection and then with the adjunct radiographic method in occlusal and proximal surfaces of primary molars. Values indicate the absolute number of surfaces

Fig. 3

Decision tree focused on the dental surfaces with the indication of no treatment according to the visual inspection (Fig. 2, pathway a). Italicized numbers indicate the number of surfaces. Unformatted numbers indicated the frequency in relation to the total number of surfaces (n = 4383), and figures in parenthesis are the 95% confidence intervals adjusted by the cluster

Fig. 4

Decision tree focused on the dental surfaces with the indication of non-operative treatment according to the visual inspection (Fig. 2, pathway b). Italicized numbers indicate the number of surfaces. Unformatted numbers indicated the frequency in relation to the total number of surfaces (n = 4383), and figures in parenthesis are the 95% confidence intervals adjusted by the cluster

Fig. 5

Decision tree focused on the dental surfaces with indication of operative treatment according to the visual inspection (Fig. 2, pathway c). Italicized numbers indicate the number of surfaces. Unformatted numbers indicated the frequency in relation to the total number of surfaces (n = 4383), and figures in parenthesis are the 95% confidence intervals adjusted by the cluster

Fig. 6

Decision tree for the decision of operative treatment for occlusal and proximal surfaces of primary molars decided first by visual inspection and then with adjunct radiographic method. Italicized numbers indicate the number of surfaces. Unformatted numbers indicated the frequency in relation to the total number of surfaces, and figures in parenthesis are the 95% confidence intervals adjusted by the cluster

Fig. 7

Mediation analysis to evaluate if performing a restoration mediates the occurrence of new operative interventions in the surfaces where caries lesions were missed by visual inspection but detected by radiographs (a: direct effect; b: mediated effect). Numbers represent the multilevel logistic regression coefficients (standard errors) adjusted by type of teeth, dental surface, caries experience and child’s age. P value was derived through Sobel test

Fig. 8

Time-to-event analysis for the surfaces with caries lesions detected only by radiographic method considering their clinical course (n = 121). Each arrow indicates a dental surface with a negative result through visual inspection, but positive result with radiographs. When N is indicated, there were more than one surfaces following the same clinical course