Caries thresholds
Simple oral health surveys usually apply criteria from WHO guidelines, recording the signs of disease only at an advanced stage, on a dichotomous principle (yes or no);
that is, the surface can be recorded only as either sound or carious (caries is recorded as present when a lesion in a pit or fissure, or on a smooth tooth surface, has a detectably softened floor, undermined enamel, or a softened wall. A tooth with a provisional restoration should also be included in this category. On approximal surfaces, the examiner must be certain that the explorer has entered a lesion. Where any doubt exists, the surface is recorded as sound.
In surveys, some compromise is inevitable, and this factor is not inconsequential.
Because the criteria are established for practicality and convenience under specific conditions, a surface that does not fulfill the minimum criterion for a positive diagnosis is not necessarily sound. For convenience and practicality, a number of false-negative results will be accepted; along with the truly sound surfaces, surfaces with some degree of caries will also be denoted as sound. In other words, the true number of carious teeth and surfaces is considerably underestimated in epidemiologic studies conducted according to WHO criteria.
Pitts (1997) has pointed out that noncavitated enamel lesions (D1 and D2) are about three times more common than are lesions in dentin (D3 and D4), particularly those with cavitation into the dentin. The precision of caries diagnosis is illustrated as an iceberg in Fig 231; the level at which the iceberg “floats” will depend on the selected threshold. In this figure, the water level is at the threshold used in classic dental epidemiologic studies; caries into dentin (D3); ie, the examiner ignores all signs of lesions less severe than clinically detectable lesions in dentin and records such
surfaces as “caries free.” The iceberg has been stratified into discrete levels, or diagnostic thresholds, from the most severe D4 (lesions extending into the pulp chamber) to subclinical lesions, less advanced than even clinically detectable D1 lesions (enamel lesions with apparently intact surfaces).
The D1 to D4 terminology has formerly been widely used by the WHO. The two most commonly selected thresholds are D3 (dentin caries, comprising D3 and D4 lesions only) and D1 (enamel caries, comprising lesions at D1 + D2 + D3 + D4). Figure 231 shows clearly that examinations in clinical practice will detect more lesions than will examinations using the same methods at a different threshold in a survey. Similarly, the use of diagnostic aids will also result in the detection of more lesions. For example, in contrast to national epidemiologic surveys according to WHO
criteria, surveys in Sweden routinely record approximal caries on the basis of bitewing radiographs, and enamel lesions (D1, D2) as well as noncavitated and cavitated lesions in dentin are detected. Compared to other national surveys, epidemiologic data from Sweden, which include noncavitated approximal lesions in dentin, are therefore overestimated. Figure 232 from the county of Jonkoping, Sweden, illustrates the proportion of approximal enamel (D1, D2) and dentin lesions detected on bitewing radiographs in 3-, 5-, 10-, 15-, and 20-year-old children and
young adults in 1973, 1983, and 1993 (Hugoson et al, 1999).
There are, moreover, additional problems when caries is measured as a dichotomous variable, because mineral loss from the surface, leading to cavitation, represents a continuum of changes as a result of the carious process. To dichotomize this continuous variable inevitably results in some loss of information (just as it would, for example, if the same were done when the height of a growing child was measured).
Unfortunately, there are no methods of measuring the lesion as a continuous variable
(like height or weight). There is, however, no a priori reason to classify lesions only
as either one of two categories (present or absent). Alternatives have been proposed,
eg, the Norwegian five-point scores for occlusal, approximal, and secondary caries
(Espelid and Tveit, 1986; Espelid et al, 1994; Tveit et al, 1994; see chapter 5).
With all methods of measuring caries, two additional descriptive dichotomous
categories are always included: filled (presumably because there had, at some time,
been a carious lesion), and missing (for teeth extracted because of caries). An
additional problem arises in epidemiologic studies of dental caries. The unit of attack
of a lesion is usually the surface of a given tooth, eg, occlusal, mesial, buccal, distal,
or lingual. Depending on the purpose of the study, these surfaces may constitute the
unit of diagnosis: The worst lesion present on the surface determines the
classification. It may sometimes be necessary to classify the surfaces differently: for
example, when information on caries affecting different morphologic types (eg, pits,
fissures, and smooth surfaces) is required. The unit of diagnosis is not fixed. For rapid
surveys, it may be appropriate to classify each tooth, rather than the surface, based on
the worst condition on any surface. Epidemiologic studies in children deal almost
exclusively with primary coronal caries. In adults, however, coronal and root caries
are usually considered separately, and from a treatment needs aspect, secondary caries
is also included. Each method has its own strengths and limitations, and some
compromise may have to be made regarding what loss of information is tolerable for
the specific purpose of the study.
