HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (18) Citing Articles via Google Scholar Google Scholar Articles by Nyvad, B. Articles by Baelum, V. Search for Related Content PubMed PubMed Citation Articles by Nyvad, B. Articles by Baelum, V.

Construct and Predictive Validity of Clinical Caries Diagnostic Criteria Assessing Lesion Activity

¹ Royal Dental College, Faculty of Health Sciences, University of Aarhus, Vennelyst Boulevard 9, DK-8000 Aarhus C, Denmark;
² Clinic of Oral Pathology, Operative Dentistry and Endodontics, Kaunas University of Medicine, Eiveniu 2, LT-3007 Kaunas, Lithuania;

*corresponding author, nyvad{at}odont.au.dk

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

 
Even though there is no "gold standard" for determining caries lesion activity, it is nonetheless possible to evaluate the validity of such diagnostic measures. The aim of this study was to estimate the construct and predictive validity of caries lesion activity assessments by means of their ability to reflect known effects of fluoride on caries. A three-year trial of the effect of daily supervised brushing with fluoride toothpaste was carried out among 273 12-year-old children. All children were examined clinically according to diagnostic criteria for activity assessment. The relative risk (fluoride vs. control) for caries lesion transitions among diagnostic categories was calculated. Fluoride inhibited progression of caries at all stages of lesion formation while at the same time enhancing lesion regression. The effects were most pronounced for active non-cavitated lesions. It is concluded that the clinical diagnostic criteria have construct and predictive validity for the assessment of caries lesion activity.

KEY WORDS: caries diagnosis • clinical trial • fluoride toothpaste • validity

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

 
Evaluation of caries diagnostic criteria includes assessment of their validity. Several forms of validity can be distinguished, including construct validity (measurement reflects theoretical concepts regarding the phenomenon); content validity (measurement reflects the domain of the phenomenon); and criterion validity, which has two aspects, concurrent (measurement and criterion are concurrent) and predictive validity (measurement precedes criterion). Assessment of criterion validity requires the existence of external criteria ("gold standards") for the phenomenon (Last, 1995). So far, most caries diagnostic criteria have focused on depth of lesion penetration, and it has therefore been possible for their concurrent validity to be validated with the use of the histological appearance of the lesions as the "gold standard" (Downer, 1989). However, some aspects of caries have no histological "gold standard", and forms of validity other than the concurrent validity must be considered.

Recently, we introduced a set of clinical caries diagnostic criteria that includes assessment of the activity of lesions (Nyvad et al., 1999). These criteria are based on the observation that caries lesion activity is reflected in surface reflection and texture—chalky and rough lesions being active, and smooth, shiny, and hard lesions being inactive/arrested (for review, see Thylstrup et al., 1994). These criteria were adopted to reflect the clinical observation that non-cavitated caries lesions do not always convert into a cavity but often become arrested or regress (Black, 1914; Backer Dirks, 1966).

In the absence of a concurrent "gold standard" for caries lesion activity, two other validity aspects can be used to address the validity of these new caries diagnostic criteria. Hence, the criteria may be said to have construct validity if they can reflect the known caries-controlling effect of fluoride that is brought about by inhibition of demineralization and enhancement of "remineralization" of the dental hard tissues (for review, see ten Cate and Featherstone, 1996). Similarly, the criteria have predictive validity if different diagnostic categories can predict different outcomes of the caries process.

Based on these considerations, validity of the new caries diagnostic criteria would be reflected in confirmation of the following predictions: (1) The chance that an active non-cavitated (ANC) caries lesion becomes inactive or sound is higher among surfaces exposed to fluoride than among non-exposed surfaces; (2) the risk that a sound surface or an inactive non-cavitated (INC) caries lesion progresses to an ANC lesion or worse is lower among surfaces exposed to fluoride than among non-exposed surfaces; and (3) the risk that an ANC lesion progresses to more severe conditions is higher than for INC lesions and sound surfaces. The aim of the present study was to evaluate these predictions by means of a re-analysis of data from a three-year clinical trial of the effects, on caries increments, of mouthrinsing following supervised brushing with fluoride toothpaste.

	MATERIALS & METHODS

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

 
The data originate from 273 10- to 12-year-old Lithuanian schoolchildren (193 test and 80 control individuals) who completed a three-year clinical trial on the effect, on caries increments, of mouthrinsing following daily supervised brushing with a 1500-ppm-fluoride toothpaste (Machiulskiene et al., 2002). Briefly, the trial involved children from three schools, one of which served as the control group. Children in the other two schools performed daily supervised toothbrushing with or without a subsequent mouthrinse with water, respectively. No interventions were given in the control school. Schools, and not children, were randomized to the interventions, primarily for logistic reasons related to the organization of daily supervision of toothbrushing sessions, but also to avoid mixing of interventions. The groups were balanced at baseline with respect to age, gender, and radiographic DMFS count. The clinical baseline DMFS count was slightly lower in one of the intervention schools than in the control school. However, the slight difference in total baseline DMFS counts between two of the schools is unlikely to have had any impact on the results presented here, since the analyses carried out are conditional on the baseline caries scores. Children participated after their parents had provided written informed consent to the protocol that was reviewed and approved by the Independent Board of Ethics of Kaunas Medical Academy, Kaunas, Lithuania, and by the Ethical Committee of Sciences, Aarhus, Denmark.

All children were examined clinically by the same examiner (VM), at baseline and after 3 yrs, using caries-diagnostic criteria assessing the activity (active, inactive) and surface integrity (intact, surface discontinuity, cavity) of the lesions (Nyvad et al., 1999). The inter- and intra-examiner reliability of the recordings (kappa) ranged between 0.74 and 0.85 (Nyvad et al., 1999). All permanent teeth present were recorded at the surface level. Missing surfaces were recorded as either unerupted or extracted. Sixteen surfaces were excluded from analysis, since they could not be recorded due to the presence of orthodontic appliances. The examiner was blinded with respect to the baseline caries recordings at the three-year follow-up, whereas it was not possible to keep the examiner blind with respect to the presence or absence of daily supervised toothbrushing. The pronounced improvement of the oral hygiene condition in the two toothbrushing groups made it clear to the examiner that interventions were carried out.

For the purpose of the present analysis, the original caries recordings were re-coded as follows: Score 0 = Sound (original score 0), Score 1 = ANC lesion (original scores 1 and 2), Score 2 = INC lesion (original scores 4 and 5), and Score 3 = Cavity (cavity, filling, or extracted; containing the original scores 3, 6, 7, 8, 9, and X). Subsequently, for all possible caries score transitions between baseline and follow-up (e.g., from Score 2 to Score 1), the transition risk was calculated as the proportion of surfaces at risk that underwent the transition in question. Predictions 1 and 2 listed above were evaluated by calculation of the ratio of the transition risks (RR) among the fluoride-exposed and the controls, stratified by type of surface (occlusal, approximal, or smooth). Prediction 3 was evaluated by calculation of the ratio of the risks (RR) of caries transitions to Score 3 for surfaces that were active (Score 1) vs. inactive (Score 2) or sound (Score 0) at baseline. These risk ratios were calculated separately for the fluoride-exposed and the controls, and for each type of surface (occlusal, approximal, smooth).

The risk ratios were calculated by means of epitab/cs in Stata 7.0 (Stata Statistical Software, 2001). The clustering of tooth surfaces in individuals may lead to an underestimation of the variance of the risk ratio estimates due to interdependence of surfaces in the same mouth, but this may be countered by the use of robust estimates. Unfortunately, robust estimators are not immediately available for risk ratios. We therefore used an indirect method to obtain robust variance estimates for the risk ratios. Hence, a "robust" variance correction factor may be calculated as the ratio of the robust and the conventional variance estimates when the analyses are based on odds ratio estimates rather than risk ratio estimates. We thus performed a series of logistic regression analyses of the relevant associations (e.g., between fluoride exposure [yes/no] and follow-up caries score, conditional on the baseline caries score), depending on the particular caries score transition in question. These logistic regression analyses were carried out by means of procedure logistic in Stata 7.0 (Stata Statistical Software, 2001), with options robust and cluster to provide robust variance estimates, and without these options to provide uncorrected variance estimates. The ratio of these estimates (robust/uncorrected) was subsequently multiplied by the uncorrected risk ratio variance estimate to provide a robust risk ratio variance estimate. This method is useful because the odds ratio, OR, is a good estimator of the risk ratio, RR, particularly when the "rare disease" assumption applies (Greenland and Rothman, 1998). We evaluated this assumption by comparing the RRs with the estimated ORs and found that, for most estimates, the discrepancy between the RRs and the ORs was minute. Finally, "robust" confidence intervals were calculated for the RRs by means of the "robust" risk ratio variance estimates thus derived.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

 
Most of the surfaces considered in this study had a baseline caries diagnosis of sound, and almost 90% of these surfaces remained sound after 3 yrs (Table 1). Even so, a considerable number of ANC and INC lesions were observed at baseline and thus at risk of transitions. Overall, sound fluoride-exposed surfaces had a lower risk of progression to an ANC lesion (RR_All = 0.72) or to a cavity (RR_All = 0.77) than had sound surfaces in the control group, whereas the risk of progression from sound to INC was similar among fluoride-exposed and non-exposed surfaces (RR_All = 0.99) (Table 2). Fluoride-exposed INC surfaces had a lower risk of progression to an ANC lesion (RR_All = 0.35) or to a cavity (RR_All = 0.72) than had non-exposed surfaces (Table 2). Fluoride-exposed ANC lesions had a higher chance of becoming sound (RR_All = 1.60) or INC (RR_All = 1.28) and a lower risk of becoming cavitated (RR_All = 0.60) than did non-exposed lesions. Surfaces erupting during the study had a slightly greater chance of remaining sound (RR_All = 1.08) and a lower risk of developing an ANC lesion (RR_All = 0.47) or progressing to a cavity (RR_All = 0.51) when exposed to fluoride (Table 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

Our second prediction—that fluoride-exposed sound surfaces and INC lesions have a lower risk of progressing to ANC lesions or worse than surfaces not exposed to fluoride—may seem peculiar in view of the fact that fluoride exerts its effect only when an active caries process is taking place. However, the fact that a surface has been recorded as sound or having an INC lesion at the beginning of a three-year clinical trial in a high-caries population (Machiulskiene et al., 2002) cannot be taken to indicate that it would stay that way. Hence, several sound surfaces and surfaces with INC lesions may experience an active caries process at some later point during the trial, and the anticipated effect of fluoride will be elicited. Our second prediction is therefore also relevant to illustrate the anticipated caries-controlling effect of fluoride. However, all other things being equal, the risk of lesion progression would be expected to be higher among surfaces with ANC lesions than among surfaces with INC lesions or sound surfaces at baseline, particularly since the population in question experienced high three-year caries increments (Machiulskiene et al., 2002).

Our observations have confirmed the first two predictions regarding the anticipated effect of fluoride on caries lesion dynamics. In fact, the entire transition matrix presented in Table 2 shows that the presence of fluoride in the oral environment inhibits the progression to active or cavitated stages for all baseline stages of lesion formation, although the effect appears more pronounced for unerupted surfaces and ANC lesions. Similarly, fluoride enhanced the regression of both ANC and INC lesions to sound, the effect being most pronounced for the ANC lesions. Hence, the activity assessment criteria for recording caries lesions in the present study proved capable of reflecting our hypothesis that fluoride exposure reduces the risk of lesion progression among ANC as well as INC lesions or sound surfaces, while at the same time increasing the chance of lesion regression among ANC and INC lesions, compared with the non-intervention group. We conclude, therefore, that the present caries-diagnostic criteria have construct validity for the assessment of caries lesion activity.

Our findings also support the conclusion that the caries-diagnostic criteria have predictive validity for the assessment of caries lesion activity. Hence, ANC lesions had a higher risk of progressing to a cavity or filling than did INC lesions. The observation that this effect was more marked in the control group than in the fluoride group should be seen in the light that the fluoride exposure was brought about by daily toothbrushing with a 1500-ppm-F toothpaste, and this intervention has additional effects. Hence, the supervised brushing led to a considerable improvement of the oral hygiene status (Machiulskiene et al., 2002), which is likely to have resulted in a reduction of the cariogenic challenge and made the active lesions less likely to progress (Neilson and Pitts, 1991). Moreover, the daily brushing sessions may have resulted in the removal of a part of the outer demineralized enamel on easily accessible smooth surfaces, thereby effectively converting ANC lesions into INC lesions or sound surfaces by means of an abrasive effect (Årtun and Thylstrup, 1986). Certainly, this would explain why we observed that ANC lesions in smooth surfaces in the fluoride group had a lower risk of progression to a cavity or filling than had INC lesions.

In three of 25 cases, the direction of the RR estimates for approximal surfaces deviated from those of the other surfaces (Tables 2, 3). In all three cases, INC lesions were involved: Two cases involved sound surfaces, and one case involved ANC lesions, indicating that the distinction between ANC and INC lesions and between non-cavitated lesions and sound surfaces may be particularly difficult for approximal surfaces. This is not surprising, since more than 80% of the misclassifications with the current diagnostic system may involve non-cavitated diagnoses (Nyvad et al., 1999).

A particularly intriguing observation of the study was that the inhibitory effect of fluoride appeared to be more pronounced for unerupted surfaces and ANC lesions, than for sound surfaces or surfaces with INC lesions. It has thus been suggested that post-eruptive enamel maturation is driven by processes that can be regarded as sub-clinical caries (Richards et al., 1977; Fejerskov et al., 1994). As a result, it would be expected that post-eruptive enamel maturation and active ongoing caries processes share features of lesion behavior, and our observations lend support to this expectation.

In this study of the validity of a set of new diagnostic criteria for caries lesion activity assessment, we have circumvented the problem incurred by the absence of a concurrent "gold standard" for lesion activity by evaluating the construct and predictive validity of the criteria. We thus conclude that the criteria have construct as well as predictive validity, just as we have previously demonstrated that the criteria can be used reliably (Nyvad et al., 1999). Caries lesion activity assessment is important for the determination of the need for treatment, since active lesions require intervention (operative or non-operative), while inactive lesions may demand no further attention, except for daily use of fluoride toothpaste (Nyvad and Fejerskov, 1997). The results of the present study, combined with those of previous reports (Nyvad et al., 1999), demonstrate that valid and reliable diagnostic criteria are available for caries lesion activity assessment which may be used to improve treatment decisions.

	ACKNOWLEDGMENTS

 
The study was supported by Aarhus University (Denmark), and by the Aarhus University Research Foundation (grant no. 95.2825.2874). Additional financial support was received from Colgate Palmolive (USA) and Dandy A/S (Denmark).

Received April 17, 2002; Last revision September 26, 2002; Accepted November 1, 2002

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS & METHODS RESULTS DISCUSSION REFERENCES

 
Årtun J, Thylstrup A (1986). Clinical and scanning electron microscopic study of surface changes of incipient enamel caries lesions after debonding. Scand J Dent Res 94:193–210.[ISI][Medline]

Backer Dirks O (1966). Post eruptive changes in dental enamel. J Dent Res 45:503–511.[Abstract/Free Full Text]

Black GV (1914). Operative dentistry. Vol. 1. 2nd ed. Chicago: Medico-Dental Publishing Co., pp. 188-190.

Downer MC (1989). Validation of methods used in dental caries diagnosis. Int Dent J 39:241–246.[Medline]

Fejerskov O, Thylstrup A, Larsen MJ (1981). Rational use of fluorides in caries prevention. A concept based on possible cariostatic mechanisms. Acta Odontol Scand 39:241–249.[ISI][Medline]

Fejerskov O, Larsen MJ, Richards A, Baelum V (1994). Dental tissue effects of fluoride. Adv Dent Res 8:15–31.[Abstract]

Greenland S, Rothman KJ (1988). Measures of effect and measures of association. In: Modern epidemiology. 2nd ed. Rothman KJ, Greenland S, editors. Philadelphia: Lippincott-Raven, pp. 47-64.

Groeneveld A (1985). Longitudinal study of prevalence of enamel lesions in a fluoridated and non-fluoridated area. Community Dent Oral Epidemiol 13:159–163.[ISI][Medline]

Larsen MJ, von der Fehr FR, Birkeland JM (1976). Effect of fluoride on the saturation of an acetate buffer with respect to hydroxyapatite. Arch Oral Biol 21:723–728.[ISI][Medline]

Last JM (1995). A dictionary of epidemiology. 3rd ed. Oxford: Oxford University Press.

Machiulskiene V, Richards A, Nyvad B, Baelum V (2002). Prospective study of the effect of post-brushing rinsing behavior on dental caries. Caries Res 36:301–307.[ISI][Medline]

Neilson A, Pitts NB (1991). The clinical behaviour of free smooth surface carious lesions monitored over 2 years in a group of Scottish children. Br Dent J 171:313–318.[ISI][Medline]

Nyvad B, Fejerskov O (1997). Assessing the stage of caries lesion activity on the basis of clinical and microbiological examination. Community Dent Oral Epidemiol 25:69–75.[ISI][Medline]

Nyvad B, Machiulskiene V, Baelum V (1999). Reliability of a new caries diagnostic system differentiating between active and inactive caries lesions. Caries Res 33:252–260.[ISI][Medline]

Pot TJ, Groeneveld A, Purdell-Lewis DJ (1977). The origin and behaviour of white spot enamel lesions. Ned Tijdsch Tandheelkd 85:6–18.

Richards A, Larsen MJ, Fejerskov O, Thylstrup A (1977). Fluoride content of buccal surface enamel and its relation to dental caries in children. Arch Oral Biol 22:425–428.[ISI][Medline]

Stata Statistical Software (2001). Release 7.0. College Station, TX: Stata Corporation.

ten Cate JM, Duijsters PPE (1983). Influence of fluoride in solution on tooth demineralization. I. Clinical data. Caries Res 17:193–199.[ISI][Medline]

ten Cate JM, Featherstone JDB (1996). Physicochemical aspects of fluoride-enamel interactions. In: Fluoride in dentistry. Fejerskov O, Ekstrand J, Burt B, editors. Copenhagen: Munksgaard, pp. 252-272.

Thylstrup A, Bruun C, Holmen L (1994). In vivo caries models—mechanisms for caries initiation and arrestment. Adv Dent Res 8:144–157.[Abstract]

This article has been cited by other articles:

				A. F. Zandona and D. T. Zero Diagnostic tools for early caries detection J Am Dent Assoc, December 1, 2006; 137(12): 1675 - 1684. [Abstract] [Full Text] [PDF]

				J. C. Hamilton and G. Stookey Should a dental explorer be used to probe suspected carious lesions? J Am Dent Assoc, November 1, 2005; 136(11): 1526 - 1532. [Full Text] [PDF]

				K.R. Ekstrand Improving Clinical Visual Detection--Potential for Caries Clinical Trials J. Dent. Res., July 1, 2004; 83(suppl_1): C67 - C71. [Abstract] [Full Text] [PDF]

				J. J. ten Bosch, M.-C. D.N.J.M. Huysmans, B. Nyvad, V. Machiulskiene, and V. Baelum Construct and predictive validity of clinical caries diagnostic criteria assessing lesion activity. J. Dent. Res., November 1, 2003; 82(11): 862 - 862. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (18) Citing Articles via Google Scholar Google Scholar Articles by Nyvad, B. Articles by Baelum, V. Search for Related Content PubMed PubMed Citation Articles by Nyvad, B. Articles by Baelum, V.