The evolution of tooth wear indices

The earliest indices shared common, arbitrary criteria, relying on descriptive terms such as slight, mild, moderate, severe and extensive. Restarski et al. [26] developed a six-point grading system to evaluate the severity of erosive destruction observed on the lingual surfaces of rat and puppy molars, but concerns were raised with regards to reproducibility. With vague criteria definitions, variability in recording is expected. Each animal was allocated a total score, calculated by summing the mean molar quadrant scores. Whilst producing simple data for analysis, it is acknowledged that averaging scores in this manner leads to the loss of much data. If the number of teeth severely affected is small, the erosion score will be low; but this could mask a significant, localised clinical problem [27].

Eccles [9] originally classified lesions broadly as early, small and advanced, with no strict criteria definitions, thus allowing wide interpretation. Later, the index was refined and expanded, with greater emphasis on the descriptive criteria [10]. It was presented as a comprehensive qualitative index, grading both severity and site of erosion due to non-industrial causes, and is considered as one of the cardinal indices from which others have evolved. In essence, it breaks down into three classes of erosion, denoting the type of lesion, assigned to four surfaces, representing the surface where erosion was detected (Table 1).

Greater accuracy was introduced by Xhonga and Valdmanis [33] who divided erosions into four levels by measurement with a periodontal probe: none, minor (less than 2 mm), moderate (up to 3 mm) and severe (greater than 3 mm). They further differentiated types of erosion by morphological descriptions, such as wedge, saucer, groove and atypical. They did not address the problem of inter- or intra-examiner variability.

It is perhaps significant that the earliest index documented by Broca [4] and used as a foundation for the development of further indices graded horizontal or oblique patterns of occlusal wear without presupposing the aetiology. Smith and Knight [30] introduced the more general concept of measuring tooth wear per se, irrespective of the cause, and since then more recent indices have been developed or modified from Smith and Knight that do not rely on a prior diagnosis and are more clinically relevant. Most of these stress the importance of user training sessions and calibration exercises.

Smith and Knight [30] took Eccles’ ideas a stage further, producing the tooth wear index (TWI), a comprehensive system whereby all four visible surfaces (buccal, cervical, lingual and occlusal–incisal) of all teeth present are scored for wear, irrespective of how it occurred (Table 2). This avoids the confusion associated with terminology and translation or differences in opinion for diagnosis of aetiology based on clinical findings. Guidelines for using the criteria were produced in a booklet by the authors to aid training and standardisation with other investigators; in cases of doubt, the lowest score is given. Complete enamel loss (score 4) may, however, be misleading, as there is almost always a rim of enamel at the worn surface margins—the colloquial “enamel halo.”

This index was the first one designed to measure and monitor multifactorial tooth wear; a further pioneering feature was the ability to distinguish acceptable and pathological levels of wear, by comparison with threshold normal values for the age groups studied. Tooth wear was defined as pathological if the teeth became so worn that they do not function effectively or seriously mar the appearance—before they are lost through other causes—or the patient dies. Results from inter- and intra-reproducibility exercises were within a range regarded as acceptable for epidemiological purposes, and the index appears simple to use clinically—intra-orally or from models and photographs. However, some problems have been identified with the TWI, including the time necessary to apply to a whole dentition, amount of data generated and the comparisons with threshold levels for each age group; the thresholds proposed were high, erring towards understatement rather than exaggerations of pathological wear. Full use of the index as a research tool is not feasible without computer assistance.

A sign of professional confidence in this index is its adopted use by a number of different investigators in the UK looking at tooth wear prevalence and severity [3, 20, 25, 27], aetiology and risk [1, 21]. Other researchers have used it with modifications pertaining to the particular age group being studied. Millward et al. [18] made adjustments to study erosion in the primary and secondary dentitions, excluding cervical surfaces. Subjects were then grouped together to produce three classifications: no or mild erosion—no score >1, moderate—at least one tooth in dentition score 2, severe—at least one tooth in dentition score 3 or 4.

In a survey of elderly people, Steele et al. [31] combined low wear scores representing small losses of enamel and used the worst surface score per tooth as an overall tooth score. Donachie and Walls [7, 8] outlined various flaws in the tooth wear index as an epidemiological tool in the ageing population and suggested a need to increase the sensitivity of TWI at extremes of tooth wear, to take account of the capacity of the elderly to have adequate function in cases of significant wear. They suggested altering threshold values, amplifying scoring criteria and creating a sixth point to distinguish between exposure of secondary dentine and frank pulp exposure.

Many other indices have been proposed for measuring erosive tooth wear [2, 14, 15, 17, 23] which have their roots in the indices of Eccles [10] and Smith and Knight [30]. Linkosalo and Markkanen [15] utilised a qualitative index with listed diagnostic criteria to confirm lesions as erosive and a four-scale grading of severity, relating to involvement of dentine. Their scoring system was modified by Lussi et al. [17] to create an erosive index that has been used widely by European workers to score the facial, lingual and occlusal surfaces of all teeth except the third molars (Table 3).

O’Sullivan [23] proposed a new index for the measurement of erosion specifically in children. The index was qualitative with a broad attempt at quantification noting whether less or more than half of the surface was affected. Every tooth was examined and assigned a three-digit score relating to the site of erosion, severity (grade 0–5) and area of surface affected. O’Brien [22] reported the use of a partial recording system for measuring erosion in children in UK children’s dental health surveys, where only the facial and lingual surfaces of the primary and permanent maxillary incisor teeth were scored for erosion. Again, the criteria were mostly qualitative and descriptive, with a broad attempt made to quantify the area involved.

Bardsley et al. [2] pioneered a new, simplified version of TWI [30] when carrying out epidemiological studies on large numbers of adolescents in North West England (Table 4). Tooth wear scoring was essentially dichotomised into the presence or absence of dentine, with even cupping of dentine scoring one. A partial recording system was used, collecting data from 40 surfaces including occlusal surfaces of the four first molar teeth and the labial, incisal and lingual–palatal surfaces of the six upper and lower anterior teeth.

However, despite calibration and training, difficulties were experienced diagnosing dentine exposure in the epidemiological field and there is some debate as to the significance of dentinal cupping when exposed dentine does not relate to significant amounts of tissue loss [12]. In a recent study of Ganss et al. [12], teeth were visually and histologically examined for presence of exposed dentine, and the correlation in accuracy between the two examinations was poor. The diagnosis of exposed dentine is obviously important but may not be accurate from visual examination alone.

Oilo et al. [24] concentrated on a different type of scoring system, with criteria based on treatment need. They criticised the use of indices that used a nonlinear scoring method, claiming calculated mean wear scores can be misleading. Their index was based on Ryge and Snyder’s [28] system for evaluating the quality of restorations and had five categories divided into two broad camps; Romeo, Sierra and Mike were satisfactory, whilst Tango and Victor indicated unacceptable levels of wear requiring treatment. All groups except Romeo were subdivided according to degree of dentine exposed and clinical findings such as pain, sensitivity and fracture of restorations, giving the impression of a cumbersome system. Dahl et al. [6] modified it with the introduction of even more categories, with an aim to establish subjective dental criteria for present and future evaluations of tooth wear and the need for treatment. In practice, these indices require experience for reliable use; individuals with differing clinical backgrounds will not get consistent, objective results.

Indices that attempt to visualise, measure and monitor the amount of worn enamel or exposed dentine by difficult direct clinical interpretation [13] and indirectly on accurate, serial study casts which must be retained [5, 11, 29, 32] have been described. Larsen et al. [14] recommended a new clinical index based on a combination of clinical examination, photographs and study casts, with complicated qualitative and quantitative criteria. Plaque-free teeth were clinically examined and photographed prior to taking silicone impressions for epoxy resin casts. They considered clinical and photographic data to be supplemental with final wear classification based on visual inspection of casts at ×10 magnification. Each tooth surface was scored, with six grades of erosion severity modelled from Smith and Knight [30]; the index and its criteria are complicated and time consuming.